Inscriptions without boundaries: how action at a distance is enabled on social media

This paper reports on a grassroots initiative, RefugeeAid ^[1], aimed at supporting refugees arriving at Lesbos during the autumn of 2015. The initiators raised SEK 9,000,000 (EUR∼900,000) in one week and delivered thousands of boxes of clothes, blankets and medical supplies to refugees during the subsequent months. These achievements were gradually accounted for on the initiative's Facebook page. We discuss how the initiative is acted upon by a crowd of stakeholders ^[2] and how this process is shaped by the inscriptions provided on the Facebook page.

Social media differ distinctly from earlier digital phenomena in organisations (e.g. ERP systems, knowledge management tools, intranets, Internet-based reporting) in that they combine engagement of the masses, interaction between multiple actors, user-generated content and large amounts of data – in real time (Kaplan and Haenlain, 2010). Their consequences for accounting practice are being increasingly explored, including, for instance, the blurring, accentuating or redrawing of the boundaries between occupational groups (Agostino and Sidorova, 2017; Al-Htaybat and Von Alberti-Al-Htaybat, 2017; Arnaboldi et al., 2017; Brivot et al., 2017), between professional and lay auditors (Jeacle and Carter, 2011) and between centres of calculation and distant peripheries (cf. Latour, 1987; Agostino and Sidorova, 2017; Scott and Orlikowski, 2012, see also Unerman and Bennett, 2004). The latter has in part been explored in relation to so-called evaluative infrastructures (Kornberger et al., 2017), for example, eBay and TripAdvisor, where individuals review the conduct of others. As these reviews feed into numerical scales, providers of goods and services are increasingly acted upon, by both clients and their employers (Jeacle and Carter, 2011; Scott and Orlikowski, 2012). Also, organisations' presence in their own social media has been studied (Bellucci and Manetti, 2017; Manetti and Bellucci, 2016; Manetti et al., 2017 see also Blankespoor, 2018 for a review). Here, social media may blur the distinction between centres of calculation and distant peripheries as the social medium in itself becomes a platform where both the organisation and its external stakeholders act upon each other in a shared space (e.g. Agostino and Sidorova, 2017; Bellucci and Manetti, 2017). In essence, social media seem to provide ample opportunity for action at a distance, across both time and space (cf. Latour, 1987).

An important basis for action at a distance is mobile, stable and combinable inscriptions, which allow actors to accumulate knowledge about, i.e. dominate, a distant periphery (Latour, 1987). This has been extensively discussed in the accounting literature (e.g. Robson, 1992; Cuganesan and Dumay, 2009; Dambrin and Robson, 2011; Preston, 2006; Qu and Cooper, 2011). However, the role of such inscriptions has not gained much attention in the literature on Web 2.0 and accounting, partly because of a larger focus on the very actions taken and how these influence the relationship between centre and periphery (e.g. Agostino and Sidorova, 2017; Jeacle and Carter, 2011; Scott and Orlikowski, 2012). Another reason why the role of inscriptions in enabling action at a distance on social media remains largely unexplored is the quantitative nature of many prior studies (e.g. Bellucci and Manetti, 2017; Blankespoor et al., 2014; Manetti and Bellucci, 2016; Manetti et al., 2017; Yang and Liu, 2017). This naturally limits a deeper understanding of how inscriptions are interpreted and used to act upon a distant periphery. Social media have a number of characteristics that could influence how inscriptions emerge and shape action at a distance. In general, social media seem to adhere to a certain logic, featuring, among other things, the presence of multiple voices (Arnaboldi et al., 2017), a high degree of informality (Blankespoor, 2018), a mix of text and other visuals (ibid.), emotional intensity (Gerbaudo, 2016) and tendencies towards populism and individualised charismatic authority (Gustafsson and Weinryb, 2020). Emojis have even been ascribed calculative powers, as they have been shown to feed into decisions on sustainable investments (Arjaliès and Bansal, 2018). This logic is interesting, as it runs counter to much of what we know about inscriptions, for example, the importance of periodic and aggregated depictions for enabling action at a distance (Robson, 1992), or the idea that they are intended to be conveyed in one direction – from periphery to centre of calculation (Latour, 1987). The aim of this paper is, therefore, to explore the role of inscriptions on social media in enabling action at a distance. The research question guiding our investigation is:

1. How and by what mechanisms can inscriptions on social media shape action at a distance?

A setting where the relationship between inscriptions and action at a distance is assumed to be particularly pronounced is organisations' external reporting. Increasingly, organisations use social media for reporting to external stakeholders: Firms tweet their earnings announcements (Blankespoor, 2018), philanthropist organisations describe their use of money on Facebook (Bellucci and Manetti, 2017) and government agencies use Facebook to inform the public about their services and to respond to complaints (Manetti et al., 2017). Hence, organisations publish inscriptions online that enable their stakeholders to act upon them. As these studies are quantitative, we mainly know that such inscriptions exist, and that they may trigger action, for example, likes, shares and investor decisions (e.g. Bellucci and Manetti, 2017; Blankespoor, 2018). Therefore, we argue that the setting of external reporting on social media constitutes an interesting domain for empirical inquiry of the relationship between inscriptions and action, and that such inquiry would benefit from a qualitative approach (cf. Bellucci and Manetti; 2017; Manetti et al., 2017).

We draw on a qualitative analysis of the Facebook page of RefugeeAid where the founders of the initiative interact with stakeholders over one year. We identify two mechanisms by which inscriptions on social media can shape action at a distance: the flow of micro-level inscriptions and the joint stabilisation of inscriptions. Inscriptions typically come in the format of a flow rather than a batch, characterised by “obscure tales” (Latour, 1987, p. 220). By ensuring a steady flow of such micro-level inscriptions, RefugeeAid in part avoid being acted upon by their stakeholders, by signalling that good things happen and will happen and by creating a sense of closeness. Instead, action becomes a more mutual exercise where inscriptions are made sense of and “stabilised” (Latour, 1987; Robson, 1992) jointly between the centres of calculation and the distant periphery. This joint stabilisation also shapes action at a distance, by creating “powerful explanations” (Latour, 1987) that nurture, or disrupt, action. In essence, the boundaryless nature of inscriptions on social media seems important for how action unfolds. We contribute to the literature on Web 2.0 and accounting by shifting focus from the relationship between centres of calculation and distant peripheries to the relationship between inscriptions and action at a distance. This allows us to understand in more detail the conditions under which action at a distance unfolds – and may alter relationships – on social media.

In the next section we review literature on Web 2.0 and accounting from the perspective of mobile, stable and combinable inscriptions enabling action at a distance, in order to clarify what we know about inscriptions on social media and what remains unknown. We also address the idea of distance itself – and of closeness – and how these have been discussed in the literature on Web 2.0 and accounting. Thereafter, we describe our methodology. We then present the findings of our study: how RefugeeAid provided inscriptions to their stakeholders, and how these inscriptions shaped action at a distance. This is followed by a discussion, some concluding remarks and suggestions for future research.

Below we outline the ideas of mobile, stable and combinable inscriptions that enable action at a distance, including a discussion of what distance – and its counterpart closeness – may imply. In particular, we discuss whether and how these ideas have been touched upon in the literature on Web 2.0 and accounting. This outline will serve to highlight both what we know and what we do not know about the role of inscriptions on social media in enabling action at a distance.

In accounting parlance, “the term inscription refers to the material and graphical representations that constitute the accounting report: writing, numbers, lists, tables” (Robson, 1992, p. 685); it is “a material translation of any setting that is to be acted upon” (Robson, 1992, p. 691). Inscriptions are assumed to be accumulated in a centre of calculation (Latour, 1987), where action may then be taken upon a distant periphery (ibid.). Such “action at a distance” is typically understood as the efforts of those in the centre of calculation to dominate or influence the periphery. Although Robson generally stresses the power and legitimacy associated with numerical inscriptions, non-numerical inscriptions have also earned attention (e.g. Busco and Quattrone, 2015; Preston, 2006; Qu and Cooper, 2011). In our case, a grassroots initiative, RefugeeAid, post inscriptions on their Facebook page (e.g. numbers, narratives and photos to communicate how the collection of money and goods is proceeding) to keep their geographically dispersed stakeholders (e.g. donors and potential donors) informed about the project at a distance. RefugeeAid constitute the distant periphery that is acted upon by the stakeholders, those in the centre of calculation. Also, the Facebook page per se can be seen as a centre of calculation (cf. Agostino and Sidorova, 2017), being the venue where stakeholders gather to act upon RefugeeAid and, as shall be illustrated, where RefugeeAid also act upon their stakeholders. Latour (1987) elaborates the idea of inscriptions in the context of science, the context of collecting traces and “obscure tales” from the world and bringing them back to the laboratory to organise and make sense of them. In order for inscriptions to transcend time and space, to be accumulated in centres of calculation, and, ultimately, to enable action at a distance, three interrelated characteristics are particularly helpful: inscriptions should be mobile, stable and combinable (Latour, 1987; Robson, 1992).

A first premise is that inscriptions can be transferred from the distant periphery to the centre of calculation, mobilised across both time and space (Latour, 1987). This can be achieved, for instance, by means of printed reports (e.g. Preston, 2006) or handwritten notes (e.g. Qu and Cooper, 2011). Inscriptions on social media are typically depicted as highly mobile, seen in how content reaches – and engages – many people at different points in time. Examples include stakeholder likes, comments or retweets (Agostino and Sidorova, 2017; Bellucci and Manetti, 2017; Manetti and Bellucci, 2016), investor perceptions (Blankespoor et al., 2014; Cade, 2018; Kipp et al., 2019), and – in the case of evaluative infrastructures – consumer choices and reviews (Kornberger et al., 2017; Jeacle and Carter, 2011; Scott and Orlikowski, 2012). The fact that organisations are often eager to manage such action (Agostino and Sidorova, 2017; Arnaboldi et al., 2017; Brivot et al., 2017; Cade, 2018; Scott and Orlikowski, 2012) further underlines the importance ascribed to mobility in the literature. In the case of Facebook, the mobility of inscriptions is conditioned by the algorithm that impacts what content appears in each individual's feed, based on, for example, the individual's behaviour on Facebook and his/her connections (see Bucher, 2012). In a similar vein, the most popular tweets are displayed in a larger font size than other tweets, something which seems to generate higher engagement among investors on social media (Yang and Liu, 2017).

Overall, however, the mobility afforded by social media is often illustrated as decisive for action at a distance – across both time and space – as inscriptions can be made visible in real time to masses of people, whose (re)actions may have a significant impact on an organisation.

In order to survive the transcending of time and space, inscriptions need to be stable with regard to their relationship to the phenomenon they are to represent (Latour, 1987). One dimension of mobility concerns whether inscriptions can withstand physical corruption. As illustrated in Qu and Cooper's (2011) study of a Balanced Scorecard project, handwritten notes taken during a workshop were not easily understood by co-workers who had been absent, partly because some of the handwriting itself was difficult to decipher and partly because additions and changes had been made to the notes as the workshop proceeded. In the context of social media, inscriptions can therefore be seen as stable because they may remain online forever (which is related to mobility, discussed above). On the other hand, they can be potentially unstable, given that content on social media can be modified (cf. Treem and Leonardi, 2012). This type of instability, however, has not earned much attention in the literature on Web 2.0 and accounting.

Another dimension of stability concerns how recognisable inscriptions are to the user, relating to labels, definitions and classifications (Latour, 1987; Robson, 1992). Stability thus involves an element of agreement or negotiation and may necessitate strong proponents or other mobilising forces. Then, “powerful explanations” (Latour, 1987) can emerge. This has been illustrated, for example, in how a comparative cost accounting system in Australian healthcare was allowed to emerge through the mobilisation of healthcare professionals, government officials and consultants (Chua, 1995), and in how performance measures for drug reps were accepted in spite of their unclear correspondence with underlying work, in that the drug reps' identities were built more on medicine than on sales (Dambrin and Robson, 2011). Although the latter study points to the contrary, stability may require the successful translating back, from inscriptions to local traces, i.e. rendering visible the process underlying a specific inscription: “In case of a dispute, other tallies, code words, indicators, metres and counters will allow dissenters to go back from the nth final inscription to the questionnaires kept in the archives and, from it, to the people in the land” (Latour, 1987, p. 234). In addition to social processes, the technology per se can play a role in stabilising inscriptions. Themsen and Skærbæk (2019) show how the use of a digital risk management system in a megaproject – building on particular categories and boxes – contributes to constructing managers' views of what constitutes a risk and what does not. Also, and importantly, inscriptions may lack stability in terms of representation of the underlying reality yet still enable action at a distance, because they serve to construct ideas of how things can be interpreted and done (Busco and Quattrone, 2015). Such processes of making sense and acting hence need not be uniform among actors; rather, inscriptions outline “the space of the possible” that invites ordering and interrogation (ibid., p. 1253). In a similar vein, rather than representing reality, inscriptions may be used to legitimise reality (Preston, 2006).

Inscriptions on social media typically do not adhere to particularly strict formats, labels and definitions, thus potentially challenging stability. One explanation for this lies in the less standardising nature of social media, compared to, for example, more traditional digital accounting technologies that may streamline inscriptions (cf. Themsen and Skærbæk, 2019). There are variations, however. Companies that publish financial inscriptions on their social media typically do so using established terms such as “revenue growth”, specifying the time periods concerned. As such, these inscriptions can be interpreted as stable. These inscriptions enable investors to act upon these companies, as seen in how financial tweets generate likes, shares and market activity (Blankespoor, 2018; Blankespoor et al., 2014; Yang and Liu, 2017). At the other end of the spectrum are non-numerical inscriptions, which can be regarded as considerably less stable. Studies of organisations' use of Facebook show how narratives, photos and the occasional number, including a mix of information, entertainment and calls for stakeholder engagement, are published without labels clarifying the purpose of the content (Bellucci and Manetti, 2017; Manetti and Bellucci, 2016). Nevertheless, these inscriptions generate action from the stakeholders' side in the form of likes, shares and comments. Numerical or non-numerical inscriptions notwithstanding, these studies are quantitative and therefore reveal little about how inscriptions on social media shape action at a distance and what the role of stability is therein. Studies of evaluative infrastructures such as TripAdvisor have mainly focused on the numerical inscriptions enabled by the preset scales of the platforms, which in turn are translated into scores and rankings (Jeacle and Carter, 2011; Scott and Orlikowski, 2012). These inscriptions can be considered more stable than non-numerical inscriptions on a Facebook page, as described above, yet less stable than financial numbers published by established organisations. The tracing back from the scores and rankings to the underlying evaluations is far from clear, partly because it is performed by the secret algorithm (Jeacle and Carter, 2011) and partly because people's reviews are highly individual and context-specific (Scott and Orlikowski, 2012). In spite of the predefined aspects (e.g. tidiness, location) that are to be graded, which provide a hint of why an overall grade is positive or negative, these aspects lack definitions and are graded based on the personal experiences and expectations of the traveller. In that sense, the inscriptions lack stability. Still, and importantly, both consumers and employees seem to put their trust in these numbers – precisely because they are numbers (Jeacle and Carter, 2011) – thus indicating some amount of stability and, furthermore, ascribing certain importance to such stability in enabling action at a distance.

Overall, the stability of inscriptions on social media has earned less attention in the literature than their mobility. Judging from the literature, inscriptions on social media could potentially vary greatly in stability primarily because they sometimes lack definitions, labels and traceability backwards. However, the literature on Web 2.0 and accounting reveals little about how stability, or instability, shapes action at a distance. Stability has been discussed primarily in relation to evaluative infrastructures, where it is seen as something that is constructed by people's beliefs in the numbers and that, consequently, nurtures action.

Related to mobility and stability, inscriptions are expected to be combinable (i.e. possible to summarise or compare). For example, combinable inscriptions would allow stakeholders to compare the cost efficiency of different products and hospitals (Chua, 1995). Without neglecting their limits, Robson (1992) argues that accounting inscriptions are particularly powerful as they are often quantitative, expressed in the same metric and built on established definitions. However, different things may be combined without accounting; apples and oranges may be aggregated by relabelling them as “fruit” (ibid.). This is what Latour (1987) refers to as “additional work”: the process of aggregating the different inscriptions from the seemingly different realms of reality. Inscriptions thus need to be translated into higher-order categories, to “say the most […] by saying the least” (Cuganesan and Dumay, 2009, p. 1181). Otherwise those in the centres of calculation will be overwhelmed and in no better position to understand the underlying reality (Latour, 1987). Balance must be sought, however, between rendering inscriptions displayable and retaining their features. On the one hand, a mean is more graspable than a large number of quantities. On the other hand, it is less informative than if the variance is also included (ibid.). Hence, inscriptions of different orders seem to play different roles in enabling action at a distance: abstract inscriptions support the communication of the relative importance of different elements and their relationships, whereas narratives convey a complex story “without reduction” (Cuganesan and Dumay, 2009, p. 1182).

In the context of social media, examples of combinability can be seen in evaluative infrastructures where reviews accumulate into scores, and rankings, thereby promoting comparison between providers of goods and services (Jeacle and Carter, 2011; Kornberger et al., 2017; Scott and Orlikowski, 2012). However, related to what has been said above about the difficulty of tracing the underlying opinions behind rankings, such comparisons have been problematised because of their inherent and somewhat disguised subjectivity (Scott and Orlikowski, 2012). Nevertheless, inscriptions in these studies are combined and serve as a basis for action at a distance for both consumers and managers who employ these scores in evaluating the staff. As was discussed in relation to stability, the literature suggests that the numerical character of these inscriptions is decisive for action at a distance to occur (Jeacle and Carter, 2011), and it appears that such inscriptions – in spite of a certain amount of instability – are interpreted as both stable and combinable. Other than that, examples of combinability on social media are scarce in the literature. Financial inscriptions on social media, given their relative stability in terms of definitions and labels, as was discussed above, could also be considered combinable. In particular, combinability could potentially be more pronounced with recent efforts to tag financial data on companies' social media (see Lowe et al., 2012). Whether comparisons occur, however – over time or between companies – in turn generating investor action, has not been discussed in the literature. Furthermore, financial inscriptions are not always as combinable as they might appear: prior studies show that financial tweets are sometimes accompanied by more informal and subjective elements, such as positive narratives (Yang and Liu, 2017). Studies of non-numerical inscriptions, such as organisations' accounts on Facebook (Bellucci and Manetti, 2017; Manetti et al., 2017), also reveal little about whether and how stakeholders combine different accounts. While these inscriptions are typically not aggregated to comparable higher-order categories, and thus appear to be uncombinable, we do not know how this lack of combinability is interpreted and acted upon by stakeholders.

Overall, the combinability of inscriptions on social media is rarely discussed in the literature. With the exception of studies on evaluative infrastructures, which have drawn our attention to the seeming combinability of quantitative inscriptions, there are few examples of whether and how inscriptions on social media can be seen as combinable and of the role of combinability in enabling action at a distance.

Having outlined the idea of mobile, stable and combinable inscriptions that enable action at a distance, we will conclude the theoretical background by addressing the distance itself, and its counterpart, closeness. Although distance has earned less attention in the accounting literature than action and inscriptions, some underline its importance for understanding the role of inscriptions (e.g. Corvellec et al., 2016; Sundström, 2011).

Taking their point of departure in Latour's (1987) idea that distance is produced within networks and pertains to knowledge generation rather than to spatial coordinates, Corvellec et al. (2016) illustrate how distances are shaped by inscriptions and actions. They study a municipality acting upon its residents by means of invoices for food waste collection. The invoices present side-by-side inscriptions on weights and costs of waste, which bring residents closer to their waste, literally, as they are encouraged to sort it more thoroughly. The inscriptions furthermore reduce the distance between residents and the political goals of the municipality. The shaping of distance is conditioned by the trust in the municipality's waste management system and the financial literacy of the residents. Also, when invoices do not reach the individual residents, but are instead directed to the owners of residences, distance increases and behavioural change is not realised (ibid.). Sundström (2011) also elaborates on the idea that distance is not a given, nor is it a question of “either or”. Studying the use of performance measures in a theatre, aimed at different groups of stakeholders, Sundström suggests that distance has to do with the level of understanding of the context that is being acted upon. Contextual understanding reduces distance and allows users to relate inscriptions to other knowledge and, consequently, to refrain from solely relying on the inscriptions when acting upon the organisation (ibid.).

The idea that distance between centres of calculation and distant peripheries can be reduced resembles the idea of closeness (Rawls, 1972), in the way it has been discussed in the context of non-profit accounting. Operationalised as, for example, a personalised writing style and the use of photos where people look straight at the camera so that they seem to be making eye contact with the reader (Taylor et al., 2014), closeness has been suggested to downplay the importance of formal accounting reports (ibid.; Gray et al., 2006; see also Costa et al., 2019). As Gray et al. (2006) suggest: “If one can, for example, drop into an NGO and ask it about its activities then a formal specific ‘account’ is not required” (p. 335). Although various interpretations of both distance and closeness can be discerned in the literatures, they can both be seen as something that changes with the knowledge, understanding and trust that may follow as centres of calculation and distant peripheries interact.

In the context of social media, distance is mainly discussed in relation to the space for mutual action at a distance that social media provide. Most notably, Agostino and Sidorova (2017) – underlining the socially constructed dimension of distance – suggest that distance is reduced as customers start acting upon a company on a social platform. They furthermore argue that the distance is not a one-way path of knowledge accumulation but rather a space where the centre of calculation and the distant periphery gain similar levels of power and relevance. What the role of the very inscriptions may be in such reversed and distributed action at a distance is not discussed, however. In line with the idea of closeness in the non-profit accounting literature as something that can emerge with the use of visuals and a personalised tone, studies on financial reporting via social media have shown that the use of narratives or informal language in relation to financial numbers may generate higher engagement among investors (e.g. Rennekamp and Witz, 2020; Yang and Liu, 2017). While not discussing either distance or closeness, these studies suggest that the type of inscription could potentially matter for how action at a distance unfolds. This, however, is not further explored in these studies.

In sum, distance and closeness on social media have been discussed mainly in terms of the common spaces for mutual action that social media allow, yet the role of inscriptions in shaping distances remains unexplored in the emergent literature.

Summarising our theoretical points of departure, the idea of mobile, stable and combinable inscriptions that enable those in a centre of calculation to act upon a distant periphery has earned little in-depth attention in the literature on Web 2.0 and accounting. In this literature, it is often the mobility of inscriptions that is highlighted as enabling action at a distance – given the potentially vast and rapid spread of inscriptions in social media that in turn enables people to act upon organisations. The roles of stability and combinability, on the other hand, have been only briefly and rather implicitly discussed. Judging from the literature, stability and combinability can be interpreted as varying with the type of social media use, for example, evaluative infrastructures versus financial reporting versus non-financial reporting. For instance, despite underlying instabilities of certain numerical inscriptions, these may still appear to be both stable and combinable. Overall, however, how stability and combinability of inscriptions on social media shape action at a distance remains unexplored, in part due to the quantitative nature of many prior studies. As a complement to the discussion of the role of inscriptions on social media in enabling action at a distance, we have highlighted the idea that distance – and closeness – can change as a centre of calculation and a distant periphery interact. Such change could occur as a result of the shared space for action at a distance that social media provide, as has been suggested in the literature on Web 2.0 and accounting, but also – possibly – through the knowledge and understanding that emerge with the interaction. The idea, however, warrants further study.

Since we were interested in exploring a nascent phenomenon, qualitative methods appeared particularly suitable (Edmondson and McManus, 2007). In order to generate knowledge on how inscriptions on social media enable action at a distance, we designed a study based on the chronological development of inscriptions and actions on Facebook, i.e. web archiving (Lomborg, 2012). According to Lomborg (2012), “social media archives are highly useful data corpuses for the systematic, fine-grained study of naturally occurring, textualised interactions on social media” (p. 221). In particular, Facebook is known to generate more interaction compared to Twitter and Instagram (Manetti et al., 2017). Therefore, we opted for a qualitative analysis of a specific Facebook page. In line with our exploratory approach, our choice of case was not based on representation but rather on the extensive opportunities to understand the phenomenon that the case seemed to offer (cf. Stake, 1995). First, the case involved frequent publishing of inscriptions and extensive interaction between the organisation and their stakeholders in relation to these inscriptions. Second, many of the inscriptions and interactions revolved around the collection of money and the subsequent performance of activities related to the money, thus offering insight into the accumulation of knowledge in a centre of calculation and how it shaped action on a distant periphery (Latour, 1987; Robson, 1992).

The data span one year, starting in August 2015. This covers the inception of RefugeeAid as well as its trajectory from an ad hoc initiative to an organisation with a more long-term purpose. By the spring of 2016, the interaction between RefugeeAid and their stakeholders had already decreased sharply. This gradually rendered the material less suitable for understanding the role of inscriptions in enabling action at a distance, hence our choice to analyse only the first year. The data comprise around 300 pages (with some overlapping, due to technical constraints) in PDF format, detailing conversations between RefugeeAid and their stakeholders. All posts and comments were manually clicked open before the entire feed was converted to a PDF in October 2016. The Facebook page of RefugeeAid is public and thus open, so we have only used publicly available data (cf. Janta et al., 2012). In order to preserve anonymity, the organisation has been assigned a fictional name, and neither the names nor the exact dates related to the quotes are revealed (ibid.; Ekpo et al., 2014). As all quotes are originally in Swedish, the translation to English also contributes to the anonymity as verbatim formulations remain concealed.

Using NVivo, the material was coded and analysed in several rounds. Following the principles of qualitative coding (Saldaña, 2009), we first approached the data in an inductive manner. In keeping with our broader interest in knowledge accumulation as a means to act upon distant others, we identified instances of account giving as often seen in the accounting literature, i.e. inscriptions concerning the money, how it was used – or not used – and explanations and justifications related to this. Simultaneously, and in line with the perspective that inscriptions may evolve in a process (Latour, 1987), we identified situations where stakeholder comments revolved around the evaluation of performance, for instance, posing questions or expressing satisfaction or dissatisfaction in relation to how operations proceeded and how the money was used. Overall, our focus was on understanding both stakeholders' and RefugeeAid's views of what constituted relevant knowledge in accounting for their operations. Of less interest to us were data pertaining to logistical concerns, such as where and how supplies were to be packed.

The relevant data were then categorised as proactive or reactive, and as pertaining to the past or to the future. Still in an inductive vein, these categories were deemed relevant for understanding the amount of correspondence between inscriptions, underlying operations and stakeholder demands. Now turning to the literature in more detail, Latour's (1987) ideas of “obscure tales”, “additional work”, and mobility, stability and combinability of inscriptions helped us to define and express our observations in terms of two mechanisms: the flow of micro-level inscriptions and the joint stabilisation of inscriptions. These two mechanisms – and how they shape action at a distance – are illustrated throughout the empirical section and are further discussed thereafter. The empirical section is structured around a number of situations, typically occurring chronologically yet with some overlap in time. These situations were identified in a later round of analysis when we searched for the prevalent ideas, or “powerful explanations” (Latour, 1987), that emerged in the interactions in different phases.

RefugeeAid was founded by a group of Swedish friends who had no prior experience in humanitarian aid. The initiative was a response to the increasing numbers of refugees arriving at the shores of Lesbos. In August 2015 the group created a Facebook page on which they made a public call for money (to be used for food, water and transportation, for example) and for supplies (e.g. clothes and hygiene articles) that would be sent by truck. The Facebook page was subsequently used for posting inscriptions about how the work was proceeding and for interacting with their stakeholders (e.g. donors and potential donors).

We will look more closely at a number of situations where inscriptions and actions revolve around the influx of money and/or supplies and how these were being used. Each situation is characterised by one or more ideas, “powerful explanations” (Latour, 1987), shaped in the interaction between RefugeeAid and their stakeholders.

As soon as the money started coming in, which was on the first day of the initiative, RefugeeAid began to post inscriptions, on a daily basis, about the growing amount of funds. The inscriptions did not merely reflect a mathematical exercise but also reflected the feelings of the founders, for example, “We are over SEK 230,000 (∼EUR 23,000) in less than 24 h. Goosebumps and tears of joy! Thank you all! Now let's fight.”; “Five minutes ago we exceeded SEK 5,000,000 (∼EUR 500,000)! SEK five million! You are MAGIC!” The amount of money reported at any point in time was indeed framed as extraordinarily large, exceeding all expectations. Such inscriptions were followed by numerous likes (up to around 5000), shares (up to around 500), emojis and encouraging comments from stakeholders, such as: “Fantastic and wonderful!”, “Thank you for a great initiative!” and “Unbelievable! What a great job you're doing!” The idea of an extraordinary amount of money was hence strengthened by the actions of a huge number of allies, or indeed, inscriptors (Chua, 1995); actions which were largely enabled by features typical for this type of medium – buttons for sharing and for expressing likes and emotions. Moreover, the fact that the steadily growing amount was updated on a daily basis provided many occasions for both RefugeeAid and their allies to repeat the message that the extraordinary amount of money had grown, an amount that would relieve suffering and pain on Lesbos. The tone of the Facebook group was without doubt euphoric.

Little concern was voiced regarding the private phone number and bank account used for collecting funds. There were numerous enthusiastic posts stating “I just swished!” (bank transfer through mobile phone), indicating that people trusted the initiative (and that they were eager to account for their generosity). Not only did stakeholders seem to trust that the money was safeguarded, but they also seemed to trust that good things would happen, or were already happening. Many of the comments reflected a belief that the initiative would relieve suffering and effect real change, and thus that the money they just had swished was well spent. “Fantastic to have an initiative where the money is actually used for its purpose,” a stakeholder commented just a few days into the initiative, before RefugeeAid had time to use any of the money, and even less to provide any inscriptions about the use of the money received. Around the same time, another stakeholder said: “Never has it felt so good to donate money. You do it so well.” In addition to providing stakeholders with the idea of a better future, the mere existence of the inscriptions seemed to appeal to them. “The help feels so close when we get to follow you. I believe in you!” said a stakeholder one week into the initiative, also before any activities had been carried out. Furthermore, stakeholders frequently thanked RefugeeAid for their “reports”, or gave credit to them for answering all questions so promptly. RefugeeAid indeed responded to most individual questions, typically within a few hours, and also to many of the encouraging comments, thanking the specific stakeholders for their support. The frequency of the inscriptions, and the opportunities for interaction in relation to them, thus appeared to create a comforting point of contact – a centre of calculation where stakeholders could gather to accumulate knowledge about a distant periphery, RefugeeAid, and draw conclusions about the state of the operations (cf. Latour, 1987).

Overall, early inscriptions of RefugeeAid were frequently reported and seemingly uniformly – and overtly – interpreted by many stakeholders, and by RefugeeAid themselves. This contributed to many similar actions being inscribed on the Facebook page, jointly creating the idea that good deeds were taking place and would continue to do so.

After the initial enthusiasm, some stakeholders started to voice concerns and, consequently, other types of action at a distance were seen. Some of these concerns were related to issues common for established fundraising organisations, such as efficiency (e.g. what per cent of revenues was being used for the stated purpose) and control of the money (e.g. whether there was an auditor or similar role). In response to such questions, RefugeeAid stated that they were different, that help was needed quickly and that there was therefore no time to go through administrative procedures to become a registered fundraising organisation.

Most concerns, however, were tied to the amount, or rather amounts, per se, as reported in the preceding days. “I really want accounts, especially considering the amount: SEK 6,000,000 (∼EUR 600,000)”; “SEK 8,000,000 (∼EUR 800,000) comes with a lot of responsibility”; “I donated money because you are different – now do not get carried away with this amount.” Some stakeholders compared the amount of money with that raised by established help organisations in response to the refugee situation, claiming that, as RefugeeAid had raised more money, they should donate some of it to established organisations. Hence, the fact that the amount exceeded that raised by other organisations was reason enough for some stakeholders to act upon RefugeeAid in this more demanding sense. This was despite the fact that RefugeeAid had – from the outset – framed their purpose as providing aid in a place where established organisations were surprisingly absent. These arguments were countered by other stakeholders, arguing that RefugeeAid should be left to manage the money on their own without being questioned by others. “Why so suspicious when finally someone is doing something?” asked one stakeholder, before any activities had taken place. “Let them work in peace and quiet instead of spending time on responding to suspicious comments. They keep us updated on social media,” said another stakeholder, already convinced that the updates would continue.

One of the most recurring arguments from the more critical stakeholders was that with so much money RefugeeAid should be able to support other causes, such as other camps on Lesbos, places outside Lesbos – or even outside Greece.

This type of argument was put forth by both the more positive and the more negative stakeholders, yet, as stated above, the arguments shared the fact that they built on earlier inscriptions of the amounts of money. The buzz surrounding the earlier inscriptions seemed to have evolved into an idea of an excessive amount of money. Judging from the Facebook page, this idea was not based on a combination of inscriptions providing a calculation of how far the money would go. The arguments were rather based on an emergent consensus among some stakeholders that the amount of money was large, and the amount of suffering too. In support of such claims, some stakeholders mobilised inscriptions from other parts of the internet, such as blog posts and news links, of the misery in other parts of Europe.

While acknowledging the miserable situation in other places, RefugeeAid's response to these claims rested upon very practical grounds; because of their small group, 15 people with daytime jobs, and because of extant contacts on Lesbos, they could not expand operations. These responses were later reposted or referred to by several stakeholders in their attempts to defend RefugeeAid, for example, “If you read what they have written, the reason is that they have established contacts with a local organisation.” Another stakeholder, also defending RefugeeAid against critics, did so from a completely different perspective.

By combining inscriptions about the amount of money with inscriptions about the refugee situation (based on a news link) that had also appeared previously on the page, this stakeholder was the first to provide a view of the money as a scarce resource and did so by outlining the different inscriptions and assumptions involved in the calculation. Just a few days later, a similar argument appeared in an inscription from RefugeeAid themselves, where they accounted for the amount of money they had at the moment, in combination with information about the daily influx of refugees.

In addition to discussing the amount of money, stakeholders articulated a certain degree of surprise about the status of the operations. Despite several earlier inscriptions where RefugeeAid had declared when and how the collection of supplies would take place, some stakeholders acted upon RefugeeAid before the collection had taken place. “Have they [the trucks] left, how did it go?”; “Could you not fly in everything? So that it goes faster. It feels like all help is needed NOW.” Repeatedly, RefugeeAid reminded their stakeholders that the collection events were yet to take place, and that this was their top priority at the moment. This evoked a comment from yet another stakeholder: “What have you done up until now, except for collecting money?”

Overall, the inscriptions of the collected money generated a wider array of interpretations and actions as the amount reached certain levels, justified in part by mobilising inscriptions from other places on the internet and in part by mobilising earlier inscriptions from RefugeeAid or from other stakeholders. Inscriptions often combined different levels of abstraction, pertaining to different realms of reality (cf. Latour, 1987). The inscriptions furthermore continued to shape the idea that things were happening – even when they were not. We will now proceed to the inscriptions published to illustrate how the money was used.

A few weeks into the initiative, the provision of inscriptions became even more pronounced. Inscriptions of the amount of money seemed not only to have lost some of their relevance, they also seemed to evoke different types of action at a distance than they did initially. As if operations were not advancing swiftly enough, RefugeeAid started posting inscriptions that referred to future inscriptions.

Words such as “all the way” and “continuous” furthermore reflected an idea of inscriptions as depicting – rather than summarising – a process and, also, an ideal of inclusion. As was illustrated earlier, gaining gradual insights rather than just an aggregated inscription of the major results appeared to speak to many of the stakeholders. “You may update with photos from your everyday life”; “Tell us what's going on with the money. It's been a while… preferably with photos!”; “It would be nice to have photo updates too, where you get to see the loaded trucks, updates from the journey and finally when you arrive and hand over everything so that one can see that the mission is accomplished.”

In Latourian (1987) parlance, stakeholders were interested in “obscure tales” – and these provided a basis for them to act upon RefugeeAid. The first inscription of more concrete activity was a post consisting of three photos of large groups of people sorting and packing clothes and a text saying “We have quite a lot to do now, as you may understand. ☺ Thanks for all the gifts and an extra warm thanks to the volunteers who are helping us sort and pack. You are fantastic!” It was followed by numerous likes, shares and comments – comments from stakeholders cheering on the good work, and comments from volunteer stakeholders attesting to a “joyful experience”, “incredible logistical arrangements” and “good structures that enabled efficient work from the beginning”. A large number of allies thus agreed that the inscription represented important and efficient operations of RefugeeAid.

As RefugeeAid began their operations on Lesbos, they started to post inscriptions from the island. The inscriptions always consisted of photos together with text, yet they varied, with some depicting the general conditions in the area and others showing the personnel engaged in concrete activities.

Such inscriptions received numerous likes and comments, some expressing dismay at the conditions on the island, and others offering positive messages such as “Fantastic work!” and “Thanks for being there!” Hence, in spite of sometimes containing little or no information about activities undertaken by RefugeeAid, but rather describing the situation at large, the inscriptions were inscribed by stakeholders as evidence of good deeds and effective performance; seeing the mere presence of RefugeeAid on Lesbos was sufficient for the stakeholders.

As inscriptions from Lesbos started to appear, other inscriptions evoked different actions from the ones taken before. A post containing a textual and visual portrait of a packing volunteer not only evoked fewer likes and comments compared to posts from Lesbos around the same time and compared to earlier posts about packing, but also a number of complaints: “Since the focus of this group was collection for the refugees on Lesbos, an update about just Lesbos feels important right now, and X's [name of volunteer] interview showed clearly how urgent the situation is there,” said one stakeholder. The comment combined the current inscriptions of RefugeeAid with previous inscriptions about the purpose of the initiative and about the experiences of their volunteers on Lesbos. (The interview was one in a magazine where a medical volunteer recruited by RefugeeAid talked about experiences from Lesbos, an interview that RefugeeAid had linked to on the Facebook page.) RefugeeAid argued that the packing of boxes was a prerequisite for them to provide help on Lesbos, yet the argument generated little understanding and support. Similarly, an inscription around the same time stating how many donations had come in thus far only evoked questioning comments. Stakeholders wondered why RefugeeAid were not on Lesbos at that moment and why they did not send more healthcare personnel there. Just as in an earlier example, some commented with reference to earlier inscriptions by RefugeeAid. Others did so by making comparisons with inscriptions from other initiatives or with inscriptions depicting the miserable situation.

If earlier almost any inscription was interpreted as a sign of good deeds and “fantastic work”, or at least good reporting practice, now inscriptions depicting RefugeeAid's and others' presence on Lesbos had come to take precedence as signs that good things were happening – even though they did not necessarily show any of the activities undertaken.

Overall, “obscure tales” dominated the inscriptions during the autumn of 2015 and were rarely summarised by RefugeeAid through “additional work” (Latour, 1987). Instead, stakeholders were free to mobilise and combine these inscriptions – and inscriptions from other parts of the internet – to act upon RefugeeAid. These first generated action in terms of positive evaluations of performance, and later, action in terms of criticism of insufficient performance – even when inscriptions were rather similar. Additional work was, however, undertaken after a few months. This is elaborated below.

After three months, inscriptions of expenditures associated with undertaken activities were posted on a regular basis. Inscriptions were hence more aggregated and also increasingly retrospective. Sufficient time seemed to have passed for RefugeeAid to spend more funds and to do the additional work of summarising their activities and finances. There were, however, still inscriptions that described particular events and experiences from volunteers on Lesbos and inscriptions describing plans that were yet to be realised. Around this time, RefugeeAid also announced that they had become an official fundraising foundation, with a broadened purpose of helping refugees beyond a particular place and time.

Somewhat in parallel to this shift, stakeholders started acting upon RefugeeAid differently than before. Explicit stakeholder demands for updates decreased, indicating that stakeholders no longer experienced an absence of inscriptions. Instead, inscriptions received fewer likes and comments, and inscriptions describing intentions and plans were rarely interpreted as actual deeds and performances. For example, an inscription from early 2016 reporting that RefugeeAid planned to distribute winter jackets to asylum accommodations in Sweden only generated stakeholder comments concerning practical matters, such as what type of jackets could be donated – not commendations such as “fantastic work”. An inscription in April 2016 about the miserable conditions in the refugee camps on the Greek mainland, stating that RefugeeAid intended to start supporting those, evoked very little response. This is surprising considering the stakeholders' earlier urging to redirect some of the money to other places in Greece – with reference to such miserable conditions. The inscriptions that mostly prompted stakeholders to act upon RefugeeAid during 2016 were those describing activities that had already taken place, such as the distribution of slippers, cool water and food packages in camps. Action mainly included comments about “heroes”, “nice work” and “fantastic performance”, with the occasional question about whether certain expenses were eligible for a fundraising organisation. An inscription about the establishment of a mobile medical clinic on Lesbos and associated expenditures stood out compared to other similar posts at this time.

Starting with “REFUGEEAID'S OWN CLINIC! Finally we are able to tell you about the most powerful thing you have helped us to achieve here on Lesbos – an own mobile medical clinic!”, it received more than twice as many likes as similar reports posted at the time. It was also followed by over 30 comments (which must be considered extensive for this point in RefugeeAid's history) applauding the work, including comments such as “money well spent” and “such effective use of money!” The inscription also contained information about the equipment (e.g. oxygen bottles, electrocardiogram machines, blood-pressure gauges) and the staffing (one doctor, one nurse and two interpreters, whose salaries would be covered by RefugeeAid for the next six months). Our data do not reveal why this particular inscription was so clearly defined by stakeholders as not only fantastic but also “efficient work”. It is noteworthy, however, that RefugeeAid evaluated their own achievement to the extent that it was described – and inscribed – as “the most powerful thing”.

Overall, as inscriptions became increasingly retrospective and aggregated, summarising actual achievements, action at a distance became less extensive, less demanding and – also – less appreciative. However, compared to the “obscure tales” published during the same time, the more aggregated inscriptions generated the most action, including interpretations of “fantastic work”.

Above we have described how the Facebook page of RefugeeAid emerges as a centre of calculation. Whereas evaluative infrastructures (e.g. Jeacle and Carter, 2011; Scott and Orlikowski, 2012) demonstrate a more clear-cut relationship between organisations and groups of external stakeholders, our case reveals a more fluid relationship. On the one hand, RefugeeAid's inscriptions enable stakeholders to act on the organisation – for instance, to decide whether or not to donate money, to choose to recommend that others donate money, or to praise or criticise achievements (cf. other studies of organisations' use of Facebook or Twitter, e.g. Bellucci and Manetti, 2017; Blankespoor, 2018). On the other hand, when publishing inscriptions, RefugeeAid simultaneously act upon their stakeholders in the sense that they influence stakeholder behaviour, both by encouraging certain actions and by postponing other actions. Returning to our research question, we will discuss what this case tells us about how and by what mechanisms inscriptions on social media can shape action at a distance. The question will be addressed in terms of two interrelated mechanisms that come forth as important in our study: the flow of micro-level inscriptions and the joint stabilising of inscriptions.

We discern three characteristics of the inscriptions of RefugeeAid that jointly render them distinct from inscriptions seen in many prior studies of Web 2.0 and accounting. First, the inscriptions are often “obscure tales” rather than “accumulated knowledge” (in contrast to, e.g. Blankespoor, 2018; Jeacle and Carter, 2011). Second, the inscriptions do not adhere to any explicit or established labels or categories (e.g. relating to a certain time period or to a certain type of event or performance), as stable and combinable inscriptions do (cf. Blankespoor, 2018; Scott and Orlikowski, 2012). As such, the inscriptions largely lack the additional work that would render them more well-established representations of “reality”. Third, the inscriptions are not grouped in batches, as periodic accounting reports on social media (cf. Blankespoor, 2018), and hence have no clear end. Therefore, we argue, inscriptions in this case come in the format of a flow of micro-level inscriptions. How this flow can shape action at a distance is elaborated below.

In the absence of inscriptions, stakeholders act upon RefugeeAid on the basis of this very absence. Absence may be short yet prompt demands for inscriptions or criticism of the lack thereof. Since such action is openly inscribed on the Facebook page, RefugeeAid quickly address such comments. Those who act upon RefugeeAid may not have donated anything, but that is of no importance (cf. who is the most important customer, Agostino and Sidorova, 2017). The presence of inscriptions also invites action from stakeholders – often in terms of praise. In providing a flow of micro-level inscriptions, RefugeeAid hence dominate their stakeholders by influencing the foci of their actions (cf. Agostino and Sidorova, 2017). A fast-paced flow of inscriptions seemingly signals measures and achievement, shaping stakeholders' interpretations of what has happened. Seemingly unstable inscriptions that are published one by one, rather than in a batch, furthermore appear to influence interpretations of what will happen. The bottom line has not been drawn, so there is reason for stakeholders to hope that something else is waiting around the corner, yet another step on the path towards an envisaged result. Moreover, a flow of inscriptions may create a sense of closeness that relaxes demands for more formalised accounting (cf. Gray et al., 2006) and consequently relaxes how RefugeeAid is acted upon. Hence, both the mere extent of inscriptions and their unstable nature seem to be at play here. Social media per se have been discussed as shared spaces where numerous stakeholders act alongside the organisation (Agostino and Sidorova, 2017; Bellucci and Manetti, 2017; Jeacle and Carter, 2011; Scott and Orlikowski, 2012). Our case suggests that social media, by means of encouraging obscure tales, also enable closeness in terms of tone and content (cf. Taylor et al., 2014). With the exception of the possible role of narratives and informal language in enhancing investors' engagement and positive perceptions in relation to financial tweets (e.g. Rennekamp and Witz, 2020; Yang and Liu, 2017), this type of closeness has not been discussed in literature on Web 2.0 and accounting.

The flow of micro-level inscriptions is more prevalent in the case of RefugeeAid than on social media of established organisations. Whereas established organisations provide inscriptions pertaining to a wider variety of situations, e.g. different philanthropic projects (Manetti and Bellucci, 2017), government agency services (Manetti et al., 2017) or financial results of different periods (Blankespoor, 2018), the inscriptions of RefugeeAid concern a more delimited situation and time period: aiding refugees on Lesbos “as soon as possible”. The inscriptions of RefugeeAid are therefore more coherently linked, forming precisely a flow of micro-level inscriptions, at least during the first months. The stakeholders also expect the inscriptions to be coherently linked so that the events of yesterday, today and tomorrow make sense in relation to each other. Earlier studies on organisations' reporting on Facebook show that stakeholders act rather differently on different categories of inscriptions from the same organisation (e.g. seen in the number of likes and comments in relation to awareness-raising content versus performance reports) (e.g. Bellucci and Manetti, 2017). In the case of RefugeeAid, stakeholder action – at least during the first months – is massive (often in the form of comments) regardless of the type of inscription, something which may indicate that the inscriptions are generally interpreted in terms of how they contribute to an overarching aim. Sometimes the flow of micro-level inscriptions fails to make sense to the stakeholders, potentially reshaping action at a distance again. The ability of inscriptions to make sense across contexts, i.e. their stability (cf. Latour, 1987), will be further discussed in the next section.

To summarise the discussion thus far, our case indicates a particular pace and format of inscriptions – a flow of micro-level inscriptions – which seems important in shaping action at a distance. Such a flow – by signalling present and future achievement and by creating a sense of closeness – encourages certain types of action from the stakeholders' side, while downplaying other types of action. Our study thereby adds to earlier studies on Web 2.0 and accounting by highlighting how the format and pace of inscriptions may shape stakeholder perceptions, in turn influencing how RefugeeAid is acted upon. Furthermore, while prior studies have highlighted the changing relationships between those who act and those who are acted upon (e.g. Agostino and Sidorova, 2017; Scott and Orlikowski, 2012), our study shows how different forms of action at a distance – not just action per se – may define such relationships. This fluid relationship – and the role of the flow of micro-level inscriptions – is also influenced by what we refer to as the joint stabilisation of inscriptions, discussed below.

Even though the inscriptions of RefugeeAid can in many ways be considered unstable, we see that inscriptions are to some extent made into “stable and combinable mobiles” (Robson, 1992) and that this process also shapes action at a distance. Quite irrefutably – and in line with the argument outlined in the theoretical section – the mobility of the inscriptions appears crucial. In particular, we argue, mobility across space can be seen as a prerequisite for the process of making inscriptions stable and combinable. Mobility across time, however, has a less clear-cut role in our particular case. Unlike inscriptions in expert systems (Jeacle and Carter, 2011), which are incorporated and perpetuated in scores and rankings, the inscriptions of RefugeeAid sometimes disappear in the flow and consequently, in the stakeholders' minds. Sometimes, however, past inscriptions are remobilised by stakeholders to act upon RefugeeAid, hence, we do not suggest that mobility across time is unimportant.

As stakeholders openly act on the basis of an inscription, a process of joint interpretation among stakeholders of what the inscription represents begins. RefugeeAid also contribute to this process by evaluating their own inscriptions. Elements of self-evaluation have been observed in relation to companies' financial tweets (Yang and Liu, 2017), and our case enriches this observation by illustrating a more collective and mutual process of evaluation that includes not only RefugeeAid but also its stakeholders. The more similar actions are inscribed in the centre of calculation, the more “powerful explanations” (Latour, 1987) emerge from the inscriptions, in turn being remobilised and perpetuating how action unfolds between RefugeeAid and the stakeholders. However, as the number of inscriptions increases in the flow, without RefugeeAid attempting to say “the most by saying the least” (Cuganesan and Dumay, 2009, p. 1181), stakeholders start doing some additional work themselves. With the whole flow available, stakeholders combine inscriptions at their own discretion to draw conclusions. In this process, some inscriptions are lost, some gain new meaning, and inscriptions from outside the centre of calculation are brought in. Hence, multiple explanations may emerge in the flow, influencing what type of action is taken. By allowing alternative explanations to be posted and by responding to them, RefugeeAid to some extent confirm the stability of these arguments (cf. Cade, 2018). However, and in contrast to previous studies of organisations' use of Facebook and Twitter (Bellucci and Manetti, 2017; Manetti et al., 2017), there is frequent interaction not only between RefugeeAid and their stakeholders but also among stakeholders – interactions where arguments are often developed or remobilised. Stakeholders, in our case, must therefore be considered to have a highly active role in the continuous stabilisation and destabilisation of inscriptions, and consequently, in shaping the basis for action. The grounds for defining, comparing and – ultimately – acting are unclear (cf. Scott and Orlikowski, 2012) and may change swiftly. However, compared to “expert systems”, where the process of stabilising inscriptions is hidden (Jeacle and Carter, 2011; Scott and Orlikowski, 2012), the process of stabilising the inscriptions of RefugeeAid is more overt as stakeholders often explicitly refer – and even link – to other inscriptions in their arguments. Again, the flow of micro-level inscriptions lays inscriptions open to whoever wants to act at a distance, rather than assigning to a particular person – or algorithm – the task of “mopping up” the traces and accumulating knowledge. That the role of inscriptions can be versatile in the sense that they may evoke different reactions at different points in time within the same setting, thereby not always complying with intended outcomes, is documented also in the literature on more traditional accounting (Busco and Quattrone, 2015; Christensen et al., 2019). Yet, our case also illustrates how the format of inscriptions can be versatile or, rather, boundaryless and that this boundaryless nature can be important in shaping action at a distance.

To sum up this part of the discussion, inscriptions on social media can shape action at a distance by becoming stabilised in a process between RefugeeAid and their stakeholders. As RefugeeAid and their stakeholders act upon each other, action is inscribed in the flow and, consequently, contributes to turning inscriptions into “powerful explanations” (Latour, 1987). These tend to perpetuate action – until competing explanations emerge. In the case of “expert systems”, it is the quantitative and seemingly stable nature of the inscriptions that is considered key for the masses to trust the system, i.e. to continue acting upon hoteliers, waiters and other service providers (Jeacle and Carter, 2011). In our case, we see a more joint and dynamic process of developing beliefs – turning inscriptions into powerful explanations – that nurtures, or disrupts, action at a distance. We have highlighted the role of both RefugeeAid and their stakeholders in this process, as they engage in various interactions, mobilising inscriptions also from outside the centre of calculation. Prior studies of organisations' use of Facebook and Twitter, being quantitative (e.g. Bellucci and Manetti, 2017; Yang and Liu, 2017; see also Blankespoor, 2018 for a review), have not discussed these processes.

Having discussed how the two mechanisms can shape action both at a far and a near distance, they appear to be more powerful in the short term than in the longer perspective. Our data only allow us to speculate as to why this seems to be the case. Both the broadened purpose of RefugeeAid and diminished media coverage of the refugee situation might contribute to a decreased sense of urgency among stakeholders, thereby rendering micro-level inscriptions less important in shaping action at a distance. Instead, more accumulated and retrospective knowledge takes precedence in shaping stakeholders' understanding of how operations proceed. As stated above, this may be because stakeholders become less interested with time. It may also be that stakeholders increasingly expect more stable and combinable inscriptions as RefugeeAid formalise their organisation. However, considering that even the more accumulated inscriptions published during 2016 contain photos and self-evaluative narratives, the latter explanation appears less plausible. As regards the joint stabilisation of inscriptions, this mechanism obviously does not occur to the same extent with the overall decline in stakeholder action. Neither do competing explanations emerge; inscriptions are no longer freely combined by stakeholders in support of a particular argument. What appears most powerful in shaping action in this later stage are the self-evaluations provided by RefugeeAid. Whereas the above reasoning rests upon the assumption that decreased stakeholder engagement has influenced how inscriptions are received, we should not entirely exclude the possibility that the changing nature of the inscriptions published by RefugeeAid has influenced stakeholders' engagement.

In this paper we have explored the role of inscriptions on social media in enabling action at a distance. We have identified two interrelated mechanisms by which inscriptions can shape action at a distance: the format of a flow of micro-level inscriptions and the process of joint stabilisation of inscriptions. A flow of micro-level inscriptions signals that good things are happening and will continue to happen, and it creates a sense of closeness. This enables a process where both those in the centres of calculation and those in the distant periphery act upon each other. As these actions are continuously inscribed in this flow, they enter a process of jointly stabilising the typically unstable and uncombinable inscriptions of the flow by adding elements of definition and evaluation. Action and stability are thus highly intertwined and contribute to the emergence of powerful explanations that nurture – or interrupt – continued action. Both of these mechanisms indicate that the boundaryless nature of inscriptions on social media is important in shaping action at a distance.

We contribute to the literature on Web 2.0 and accounting by shifting attention from the relationship between centres of calculation and distant peripheries to the relationship between inscriptions and action at a distance. Prior studies have discussed the characteristics of inscriptions only to a limited extent, and it is primarily their mobility that is emphasised as important for enabling action. In contrast to other studies on external reporting on social media (e.g. Manetti et al., 2017; Blankespoor, 2018), which are mainly quantitative, our qualitative analysis indicates a fluid relationship between centres of calculation and distant peripheries and some possibly important mechanisms at play as inscriptions shape action at a distance. Compared to studies of evaluative infrastructures (Jeacle and Carter, 2011; Scott and Orlikowski, 2012), which to some extent have touched upon the stability of inscriptions, our study provides insight into how seemingly unstable inscriptions outline what is possible and might influence how action at a distance unfolds. If earlier studies have focused on the medium per se as a space for acting upon each other (Agostino and Sidorova, 2017), this study shows that the inscriptions, alone and in combination, provide different opportunities for action at a distance – and for reducing the distance. It is hence not merely the mutual space that reduces the distance, but the inscriptions and their characteristics.

The characteristics and role of inscriptions on social media deserve further study in other settings. One could argue that our observations are mainly valid for the non-profit context, given the strong focus on the purpose of collected money (cf. Costa et al., 2019). Our case is furthermore loosely organised in comparison to the fundraising organisations in prior studies (e.g. Bellucci and Manetti, 2017), possibly also creating different conditions for inscriptions and actions to emerge. Other mechanisms may be at play in established organisations and, above all, in for-profit settings. However, issues in non-profits are often of broader concern (Hall and O'Dwyer, 2017). Similarly to RefugeeAid, crowdfunded business ventures start from scratch and report their progress via social platforms. There is also increased awareness among established organisations of the opportunities and threats associated with social media presence (Al-Htaybat and von Alberti-Al-Htaybat, 2017; Brivot et al., 2017). While established organisations need to adhere to mandatory forms of accounting, they may also be considering complementary forms, potentially catering to a wider variety of stakeholders (Andon et al., 2015).

Furthermore, our findings hint at the elusive concepts of transparency and accountability – and the intricate relationships between these – that have been recognised in the accounting literature (e.g. Catasús, 2008; Roberts, 2009). One interesting avenue for future research would be to disentangle different forms of transparency in relation to the boundaryless inscriptions of social media. Can a flow of micro-level inscriptions be understood to create transparency or, rather, opacity? On the one hand, such a flow – by providing more timely and specific updates that can be openly combined – could be seen as increasing transparency, compared to, for example, “black boxed” ratings and rankings (Jeacle and Carter, 2011). On the other hand, without systematic efforts to order and aggregate inscriptions, transparency could be jeopardised (cf. Latour, 1987). Also, transparency may decrease as the joint stabilisation of inscriptions has the potential to frame perceptions of what constitutes an achievement. We believe that the sense of closeness deserves further attention in such a discussion, for example, whether a “step closer” is also a step towards transparency or rather a step towards myopia, and whether closeness could be seen as a proxy for transparency or as a substitute. In line with the idea that stakeholder views of what constitutes useful accounting information differ (Andon et al., 2015), our study suggests that transparency may well be perceived quite differently across groups of audiences, hence calling for future research to empirically investigate and define transparency in different dimensions. While a rich body of literature has addressed different perspectives on transparency (for reviews, see Albu and Flyverbom, 2019; Christensen and Cheney, 2015), less attention has been directed at understanding how transparency is interpreted and constructed in practice (Heimstädt and Dobusch, 2020). Another promising direction for future research on transparency relates to the longstanding question of representation, i.e. the extent to which inscriptions can be considered adequate representations of an underlying “reality”. While interpretative researchers have certainly promoted a view of inscriptions as resulting from complex social processes rather than neutrally mirroring operations (e.g. Robson, 1992, for reviews see also Albu and Flyverbom, 2019; Christensen and Cheney, 2015), the context of social media evokes the question of whether and how such social processes interact with the process of generating likes, shares and comments. For example, the latter could potentially result in a shift in how achievements are defined and valued, in turn influencing internal negotiations over how to best represent operations. Yet another avenue for further research lies in addressing the relationship between transparency and accountability. One point of critique raised by accounting scholars is that an accountability that builds primarily on transparency risks being self-occupied (Roberts, 2009), possibly hampering responsibility (Catasús, 2008; Power, 1997). Roberts' calls for a more intelligent (2009), or socialising (1991), form of accountability would be interesting to explore in the context of social media. A more intelligent accountability, Roberts (2009) argues, involves “active inquiry – listening, asking questions, and talking” (p. 966); it “extends over time” (p. 966) and “it is typically a face-to-face encounter, rich with information, in which communication is less easily stage-managed and rhetoric can be constantly compared to actual practice” (p. 966). Social media seem to hold the potential for active inquiry (Agostino and Sidorova, 2017), but also for “organised hypocrite” (She and Michelon, 2019). As our case illustrates, inscriptions without boundaries enable action at a distance in more than one direction, whereby organisations may avoid being acted upon by being “transparent”. This would suggest that accountability may be replaced by transparency. However, our case also shows that this type of transparency can invite questions, demands and continuous dialogue. Whether a more intelligent accountability, or responsibility, is indeed enacted will depend on how far organisations go in addressing stakeholder inquiries – both in terms of providing inscriptions and in terms of actual behaviour. Such processes therefore deserve further empirical study.

Overall, accounting inscriptions on social media need further study in order to gradually untangle what is related to a social medium per se and what is related to the type of organisation. To what extent does the Facebook medium generate and perpetuate expectations on certain types of inscriptions and actions, and to what extent are such expectations intertwined with the purpose and operations of the organisation using the platform?

The authors thank participants at Nordic Workshop on Management Accounting 2018 and at seminars at Centre for Accounting Theory and Research at Innsbruck University and at the Department of Business Studies at Uppsala University for valuable comments. The authors also thank Johan Graaf for a careful reading and helpful comments. The authors are also thankful for the suggestions provided in the review process. The authors furthermore express our gratitude for the research grants provided by Handelsbankens Forskningsstiftelser and Riksbankens Jubileumsfond.