Technology and Dis/Organization: Digital data infrastructures as partial connections

Introduction

The prominent role of digital technology in shaping contemporary society is undeniable (Kolb, Dery, Huysman, & Metiu, 2020; Wajcman, 2015; Zuboff, 2019). Since Bloomfield (1995) claimed that ‘there is no technology without organization and no organization without technology’ (p. 495), multiple studies have demonstrated how technology inflects organizing processes. Some studies unpack how the development of technology is itself an organized activity, yet also has organizing capacities in its configuration of organization, users and the very boundary between organization and technology (Ratner, Andersen, & Madsen, 2019; Suchman, 2012). Others focus on digital infrastructures, exploring the relationship between digitalized business information and organizational (in)visibilities (Hansen & Flyverbom, 2015; Leonardi & Treem, 2020) or how digital technology enacts new kinds of organizational expertise (Knox, O’Doherty, Vurdubakis, & Westrup, 2007). In recent years, a growing number of studies have scrutinized the omnipresence of new digital data infrastructures, revealing how the datafication and computation of organization produce new distributed and socio-technical modes of power (Flyverbom, 2019; Leeker, Schipper, & Beyes, 2017; Williamson, 2017). These studies highlight how pervasive and powerful digital technologies organize our worlds and ways of living together.

In contributing to these debates, we argue that digital data infrastructures are not only powerful in how they organize, but also in how they disorganize. Approaching digital data infrastructures as ‘technologies that store, transmit and process data and are based on networked computation’ (Beverungen, Beyes, & Conrad, 2019, p. 622), we wish to expand the one-sided focus on the ordering capacities of digital infrastructures by also attuning to their disorganization – not because we underestimate such performative powers – but because we believe that examining the multiplex ways they work may help us better understand them. Our contribution to organization studies is thus conceptualizing the largely overlooked dis/organization of technology, inspired by Cooper’s (1986) pioneering work demonstrating the co-constitution of organization and disorganization. Building on Cooper and others (Munro, 2001; Ratner, 2020; Spoelstra, 2005), we examine how the organizing capacities of digital infrastructures also entail disorganization.

We theorize the dis/organization of technology as a mutually constitutive dynamic that works by partially connecting and transforming relationships. To analyse this empirically, we turn to social anthropologist Marilyn Strathern’s (2004) analytical concept of ‘partial connections’, which allows us to examine how digital data infrastructures work through connections, disconnections and cuts (cf. Munro, 1997). As opposed to assuming that they work through an all-encompassing connectivity, approaching digital data infrastructures as partially connecting is helpful in empirically understanding their dis/organization. Viewing digital data infrastructures from Strathern’s relational perspective allows us to explore how their outwards-connective capacities entail disconnections and how these reassemble organizational realities. Acknowledging their open-endedness, the concept of partial connections allows us to study how digital data infrastructures may reverberate and produce unanticipated gaps or interstices that dis/organize (Harvey, Jensen, & Morita, 2017; Ratner & Gad, 2019). Thus, this concept addresses both the connecting and disconnecting capacities of digital data infrastructures, crisscrossing the planned and unplanned, the organized and disorganized. This viewpoint, we suggest, offers novel insights into how dis/organization is integral to the emerging powers of digital data infrastructures.

We explore the potential of this argument in a qualitative case study of two data-ethical controversies that occurred in response to the introduction of a mandatory, digitalized survey of students’ well-being in primary and lower secondary public schools in Denmark. This survey is a powerful technology as it digitally indexes well-being as an object of governance, organizing students, schools and local governments as commensurable and comparable entities. At the same time, however, the well-being survey also disorganizes relations in data-ethical controversies occurring in response to the survey. This requires the Danish Ministry of Education to continuously renegotiate the survey’s ordering capacities. Analysing the digital data infrastructures of the well-being survey through the notion of partial connections, we find, provides insights into its resilience and power, thus enabling us to understand dis/organization as integral to the forces of technology. The case study is based on qualitative methods, compiling documents and media coverage over a four-year period (2015–2019) during which we followed the survey’s implementation and two data-ethical controversies.

The paper proceeds as follows. We begin by reviewing the literature exploring the organizing powers of digital data infrastructures. We then outline our conceptual framework, theorizing digital data infrastructures as partial connections that dis/organize. Next, we introduce the case study’s methodology and present the two data-ethical controversies to be analysed. The first concerns a local government’s illicit use of well-being data for preventive casework among students deemed ‘at-risk’. The second regards the State’s storage of non-anonymized, sensitive data about students. In analysing dis/organization, we show how the controversies emerge from the partial connectivity of the well-being survey’s digital data infrastructures. Partially connective processes of data production, storage and sharing, we argue, dis/organize relations amongst involved actors and generate normative struggles with various affective engagements. In conclusion, we discuss the paper’s contributions to organization studies of digital data infrastructures, arguing that dis/organization is integral, rather than external, to their powers.

The Organizing Powers of Digital Data Infrastructures

The field of organization studies has sustained a longstanding interest in understanding the entwinement of organization and technology (Bloomfield, 1995; Knox et al., 2007). A contemporary context of pervasive digital data infrastructures is recasting human attention and interactions as machine-readable and manipulative data entries (Leonardi & Treem, 2020; Newlands, 2021). This, in turn, has spawned a renewed interest in the relationship between technology and organization, focusing on digital data infrastructures.

For many, digital data infrastructures signify the emergence of new, powerful forces that govern and shape organizations. Digital information technologies have, for example, been depicted as ‘constitut[ing rather than merely] mediat[ing]’ our lives (Burrows in Beer, 2009, p. 987), and software is increasingly characterized as a governing force that ‘has come to intervene in nearly all aspects of everyday life’ (Thrift, 2005, p. 153). A number of these critical discussions concern the rise of algorithms (Ziewitz, 2016). Gillespie (2014), for instance, has proposed a critical investigation of the powers of algorithms with regard to their role in sorting information and governing our digital interactions. Pasquale (2015), in turn, has argued that hidden and powerful algorithms in a big data economy constitute a ‘black box society’. In these ways, scholars ask us to scrutinize the ways in which invisible and hidden algorithms govern digitally augmented interactions, shaping our ways of living together in a world increasingly saturated by digital data infrastructures (Beer & Burrows, 2013).

Looking into these questions, other studies explore how the numerical binary and abstract logic of computation reduces, changes or twists representations of reality. Kallinikos (2009), for instance, questions the ‘constitutive power of numeracy (. . .), i.e. a reality of frequencies, ratios, indices and other abstract relationships’ (p. 197). Commenting on the ‘mediatedness of organization,’ Beverungen et al. (2019) also discuss how ‘reality is indexed by a digital machine, [and] the resultant data are processed and feed back into the world as a set of observations or instructions’ (p. 627). These datafied feedback processes are not neutral. Noting how digital technology ‘has reached a state of being “ideologically invisible”,’ Selwyn (2015, p. 70) proposes that it reconfigures organizations as sites of neoliberal practices. Totaro and Ninno (2014) further suggest that algorithms represent the most advanced form of modern rationality.

These scholars’ concerns tie in with broader questions of how ‘datafication’ shifts power balances. The digital production, calculation and circulation of data renders organizations comparable and commensurable, enabling them to be governed ‘from a distance’ through the numerical constitution and standardization of organizational quality (Dahler-Larsen, 2014; Espeland & Sauder, 2007). This entails a shifting of trust and legitimacy from professionally informed judgements to new ‘centres of calculation’ (Hoeyer & Wadmann, 2020; Miller & Rose, 1990). Within education studies, to which our work also speaks, scholars have interrogated the powers of large-scale actors such as national governments, the EU and the OECD, analysing emerging digital data infrastructures as new ‘policy spaces’ (Brøgger, 2018). Some scholars have also scrutinized the emergence of software vendors that access and shape public education through standardized digital infrastructures (Williamson, 2017). Their inclusion in educational infrastructures gives rise to new ‘digitalized arrangements of governmentality’ (Hartong, 2016, p. 533), as well as private-sector ownership of student data (Sellar, 2017). These studies testify to shifting power balances towards the transnational as well as the increasingly influential role of private technology companies, convincingly demonstrating how these shifts are enabled by digital data infrastructures. Thus, they speak to broader organizational concerns about a changing and deregulated landscape where tech companies capitalize on transactional data, a ‘behaviorial surplus’ from digitally infrastructured activities and interactions. These changes enable a ‘surveillance capitalism’ (Zuboff, 2019), saturating multiple organizational and private spheres and requiring us to rethink democracy, governance, economy and theory (Bigo, Isin, & Ruppert, 2019; Mumby & Plotnikof, 2019).

In these ways, a varied body of knowledge within and beyond organization studies reveals the powers of invisible, yet ubiquitous digital data infrastructures. In order to further understand such powers, however, we argue there is a need to explore more than their organizing effects. Indeed, we risk undervaluing the complex and unanticipated ways in which power works through digital data infrastructures if we fail to also consider dis/organization as integral to their ordering capacities.

Conceptualizing the Dis/organization of Technology as Partial Connections

This paper develops a conceptualization of technology as dis/organizing. We do so by combining Cooper’s (1986) notion of dis/organization with Strathern’s (2004) notion of ‘partial connection’. This, we suggest, allows us to bypass the connectivity trope, which is rather prevalent in extant studies of digital data infrastructures. Instead, we approach digital data infrastructures as partial connections that dis/organize, with the implication that power and normative orders are continually reconfigured. Given the breadth of Cooper’s and Strathern’s work, our inspirations unfolded below only present minor parts of their extensive scholarships.

In ‘Organization/Disorganization’ (1986), Cooper challenges the functionalist assumption that order and rationality are the primary mode of organization. Instead, he argues for a relational understanding of dis/organization, making organization ‘the appropriation of order out of disorder’ (p. 328). Inspired by Saussure, Bateson and Derrida, he argues that organization is ambiguously structured in ongoing formation processes that connect and thereby order relationships, which nevertheless encompass ‘undecidable’ differences. This process both creates and transforms these relationships, often causing struggles over meaning and power. Hence, the ongoing organization of relationships does not take place at the expense of their disorganization, but in direct relation to it. This makes dis/organization mutually constitutive as an ongoing, changing relationality: ‘the work of organization is focused upon transforming an intrinsically ambiguous condition into one that is ordered so that organization as a process is constantly bound up with its contrary state of disorganization’ (Cooper, 1986, p. 305).

Inspired by this, an emerging stream of disorganization studies stresses the disorderly and messy aspects of organizational life (Hassard, Kelemen, & Cox, 2008; Plotnikof & Pedersen, 2019; Ratner, 2020; Vásquez & Kuhn, 2019). In foregrounding disorganization, they explore the power-infused struggles and tensions between order-disorder, inclusion-exclusion and control-resistance as constitutive of what is recognized as ‘normal’ organization. In particular, the early work of Munro (2001, 2003) argues that

conflict and power has the additional merit of establishing the central problem of order: the reversibility of any order. What may begin as a ‘supplement’ to the everyday disorder of living and breathing becomes, in turn, privileged and then is given priority as the organization. (Munro, 2001, p. 397)

Following Munro’s (2003) on understanding of this ongoing dis/organization as unnoticed, yet power-infused processes of unmanaging, Spoelstra (2005) suggests to understand dis/organization as emerging transformations of relationships and normative orders. Dis/organization has since been examined as, for example, gendered undoing (Pullen & Knights, 2007), spatial becoming (Knox, O’Doherty, Vurdubakis, & Westrup, 2015; Ratner, 2020) and communicative disorganization (Vásquez & Kuhn, 2019). A shared concern here is not only to forefront disorder, but to unpack the relationality of dis/organization.

Building on this, we argue that dis/organization also applies to technology, in our case, digital data infrastructures. Following Cooper (1986, 2005), Munro (2001) and Spoelstra (2005), we view the dis/organization dynamic as ongoing processes of connection and differentiation, which transform relationships and generate normative power struggles. That is, dis/organization emerges from ‘continually changing, continually informing’ which differences and parts can be connected and recognized as the normal order of the day (as well as the opposite) (Spoelstra, 2005, p. 114). In the case of technology, it becomes central to explore how ongoing efforts to order relationships through digital infrastructures (partially) connect and transform the involved actors – and how this inevitably dis/organizes those relations, often generating normative struggles. In our examples, these struggles play out as data-ethical controversies.

To analyse such dis/organizing dynamics in digital data infrastructures, we turn to Strathern’s concept of ‘partial connection’ (2004) and its use in organization studies and beyond (Ratner & Gad, 2019; Munro, 2005; Candea, 2010; Latimer, 2013). Enabling us to analytically unpack how connections may take the form of (disconnecting) cuts, we propose that ‘partial connection’ is helpful in empirically unpacking the dis/organization of relations. The concept of ‘relation’ has a central position in Strathern’s theorizing, ranging from her studies of Melanesia over her exploration of how reproductive technologies reconfigure kinship to her characterizations of the concept as an ‘epistemological device’ for social anthropology (Strathern, 1995). As opposed to a dominant ‘fetish of connectivity’ in contemporary social theory (Pedersen, 2014), Strathern invites us to imagine relations ‘as having capacity to at once join and separate’ (Strathern, 2000, p. 3). As Munro (2005) notes, Strathern invites us to ‘rethink relations as forms of connection’ that may also disconnect (p. 247).

‘Partial connections’ views ‘social relationships as connections that are both a part of and not part of each other’ (2004, p. 40). Strathern develops this idea as a non-representative argument for ’post-pluralism’ in a commentary on social anthropology. Rather than a plurality of analytical perspectives, intended to produce more complete knowledge of an object, she argues that complexity is the property of analysis: ‘the relativizing effect of knowing other perspectives exist gives the observer a constant sense that any one approach is only ever partial, that phenomena could be infinitely multiplied’ (Strathern, 2004, p. xiv). This critique of holism speaks to Cooper’s deconstruction of organization: ‘the primary whole is always a lost whole’ (Cooper in Spoelstra, 2005, p. 110), thereby allowing us to link Cooper’s and Strathern’s theorizing. According to Munro (2005), Strathern’s post-pluralism implies leaving behind the very idea of organization as a static object of study (p. 259), given its fragmentation in an infinite plurality of perspectives.

It follows that a complete perspective of any given phenomenon will never be achieved, vis-a-vis Cooper’s argument that we will never have a complete order of organization. The concept of ‘partial connections’ is thus an argument against wholeness or completeness: There can never be a complete analysis, and parts can only be made to connect partially to ‘wholes’. In this view, partial connections extend and transform one’s capacities – without encompassing them and without merging the partially connected entities into a unity or sameness. In our case, while digital data infrastructures may render Danish students’ well-being machine-readable and governable by mediating them as data points to be connected in a database, integral to this digitally mediated ‘whole’, we will argue, there are points of resistance aiming to disconnect and withdraw from that database. Thus, the holism of a database is partial, with this partiality constituting a ‘paranodal space’, i.e. ‘paranodes’ that have been left out of the database network, and to which the digital data infrastructure ‘reacts and rearranges’ itself when the Danish authorities respond to the data-ethical controversies (Mejias, 2013, p. 154). Partial connection, in this way, entails that disconnection is invariably part of the connectivity of digital data infrastructures, at once organizing and disorganizing relationships. Such an understanding of relations is particularly rewarding when studying a government-introduced digital data infrastructure as it reproduces a ‘Western conception of levels of “wholes”’ (Munro, 2005, p. 257) through its capacity to aggregate data that enact the social across the level of the individual (student), the organization (school) and society (the State).

Therefore, we adopt the concept of partial connections to analyse the dis/organization of technology. This means that we do not view the organization, for example, of a digital infrastructure as a hybrid or ‘totalizing form of “being-with”’ (Latimer, 2013, p. 77). Instead, relationality can take the form of ‘being alongside’: ‘it is this attaching and detaching to different others, partially connecting and partially disconnecting’ (Latimer 2013, p. 81). Rather than seeing connection and disconnection as opposites, partial connection thus implies an understanding of (dis)connectivity as an inherently relational phenomenon (Candea, 2010).

As studies of users quitting Facebook have also demonstrated, disconnecting from Facebook is not a matter of disengagement from technology but rather ‘using Facebook’s own infrastructural logics’ to separate from one’s online identity (Karppi in Hesselberth, 2018, pp. 2004–2005). This aligns with Munro’s discussion of the relationship between connection-disconnection as coexisting and reinforcing normative forms of organizational control (Munro, 1997). Following this, we use partial connection to analyse how the connectivity of digital data infrastructures involve disconnection, in various forms, and how these are dis/organizing. Indeed, as we will argue in the analytical parts, such partial connectivity makes it rather difficult to demarcate where the survey’s digital data infrastructure begins and ends, which itself is dis/organizing.

Eliciting an ambiguous relationality that entails both connection and disconnection, we propose that the concept of partial connection is particularly suited to analytically capture the dis/organization of digital data infrastructures. This entails exploring how such technology materializes new dis/connections, forging new relationships between previously unconnected actors and organizations while also rendering these relationships fraught with normative struggles, controversies and ‘frictions’ (Edwards, 2010).

Research Context and Methods

We draw on a qualitative case study of a government-initiated digital well-being survey implemented in Danish primary and lower secondary public schools in 2015–16, focusing on two subsequent data-ethical controversies in 2017–18 that played out in the press and broadcast media as well as on social media. Coincidentally, both authors conducted fieldwork concerning broader questions of educational governance, which was affected by the survey. Thus, we became interested in the survey’s dis/organizing capacities and decided to study this systematically, by mapping the controversies as they were presented in various media. As public controversies allow us to explore the ‘“attachments” that people mobilize (and that mobilize people) in the performance of their concern with public affairs’ (Marres, 2007, p. 759), they may shed light on how the well-being survey’s digital data infrastructure forges new partial connections between the Danish Ministry of Education, local governments, parents and companies, including efforts to disconnect. Thus, by focusing on media coverage of public controversies, we were able to follow the (dis)organization of the digital data infrastructure beyond one site or group of actors. We examined how the survey and its subsequent controversies produced and transformed relationships, enabling both the ordering and also unexpected and disordering trajectories of well-being data.

The survey was introduced in the wake of a recent national school reform (2013) that set objectives for school governance with regard to improving students’ well-being. This called for a metrics of well-being, including a digital data infrastructure for conducting the survey and for storing, circulating and calculating well-being data in a database format. The survey was developed through the involvement of researchers in ‘an expert group’, established by the Danish Ministry of Education. The survey encompasses 40 questions (20 for students in grades 0 to 2), with closed answer categories, such as ‘Q3: Do you feel lonely at school? Yes, often_ Yes, sometimes, No_,’ or ‘Q4: Do you enjoy recess at school? No_, Yes, a little_, Yes, a lot_.’ The questions target four well-being indicators, namely social relations, school competencies, motivational factors and physical orderliness. These indicators were developed through a statistical validation of responses after the first round of measurements. Since 2015, it has been mandatory for all public schools to conduct the digital well-being survey annually.

The survey is used in national and local educational governance, e.g. in local governments’ and schools’ annual quality reports. Teachers are expected to use survey results to improve students’ well-being, although the responses are anonymized at class level. Survey data is not anonymous but protected by data legislation, including the assurance of full confidentiality. This means that central government can use personally identifiable student data for statistical and research purposes, while others (e.g. teachers) can only view data aggregated at class level. As such, the nationwide well-being survey is an example of a digital data infrastructure that connects and reorganizes relations across the government, local governments, schools, students and parents. However, in our research we found that these connections not only reorganize but also disorganize in unexpected ways, which we will unpack shortly.

A large part of the data-ethical controversies took place in public and social media, as part of the digital survey’s online ‘lifeworld’. Our case study is primarily based on the systematic compilation of online communication, information and interaction regarding the survey and related data-ethical concerns, in news media, social media, central and local governments, private companies using well-being data as well as stakeholder associations. In this case, public and social media platforms co-constituted the data-ethical controversies and its ‘data publics’ (Ruppert, 2015), such as parents concerned with the (mis)use of their children’s well-being data. Thus, we approach our empirical data from various media as performative rather than representative; they enact, and thus participate in the continued ‘maintenance’ of the well-being survey as data-ethical controversial.

We organized the online data collection into three search areas: official documents and press releases, news coverage/public debates, and social media debates about the survey. First, we collected background reports on a national level; documents on the development of indicators, methods, the expert group reports, legal texts, related policy and strategy documents, as well as press releases regarding the survey from 2015–19. Sources include government websites outlining the survey’s purpose, methods and utility; a national website containing Danish legislation; and websites and press releases from public and private organizations conducting the survey. Second, we focused on national and professional news media covering the survey, e.g. media interviews with politicians, civil servants and parents. Many of these accounts raised data-ethical concerns about the survey. Third, we focused on social media interactions (e.g. Facebook and Twitter) and communications from stakeholder associations (e.g. the Danish Union of Teachers and School and Parents). All in all, this amounted to more than 150 documents, which we gathered in a database and ordered in an event timeline (Plotnikof & Pedersen, 2019).

First, through legislation, government websites and national media coverage, we mapped how digital data infrastructure(s) produced, circulated and stored well-being data, and who could access them. The analytical process also included a thematic data analysis involving both a data-centred coding process and a concept-driven coding process. First, we developed the timeline, ordering data chronologically to get an overview of how the data-ethical controversies developed over time. We then coded data inductively in collective analysis sessions, mapping how and when the survey was problematized as well as by whom. Through this, we developed a coding scheme concerning the role that the survey appeared to have for different actors (i.e. what it was explained to be; how it was articulated as controversial; and how new solutions were proposed). By bringing together these resulting patterns with our initial mapping of the digital data infrastructure, we could begin to analyse the relationship between the emerging digital data infrastructure and the data ethical controversies as partial connections.

Findings

Our analysis unfolds the dis/organization of the well-being survey. In the following two sections, we illustrate how this technology produces new partial connections between actors and organizations (e.g. parents and local governments) while also disorganizing these relationships by saturating them with normative struggles over how to use students’ well-being data.

The first part of our analysis explores how local governments’ freedom to choose a digital information system for conducting the survey gives them access to personally identifiable well-being data. This, however, runs counter to the principle of confidentiality in the legislation governing the survey’s data. We analyse this as a partial connection between different information systems, eliciting the Danish Ministry of Education’s ensuing efforts to disconnect from local governments’ illicit data use as dis/organizing. The second part of our analysis unpacks a data-ethical controversy regarding the central government storing of personally identifiable well-being data. Eventually, this leads the ministry to allow for anonymous participation in the survey and to have existing survey data deleted, although, in the end, they were unable to fulfil this promise. We analyse this as partial connections between different databases, those owned by the ministry and those owned by researchers. The two analytical parts emphasize the unintended shadow lives of digital data infrastructures as we see how they partially connect and dis/organize relationships between actors involved in the well-being survey.

Concluding discussion

With this paper, we have expanded discussions on technology and organization by theorizing and exploring digital data infrastructures as a powerful dis/organization of social relations. In doing so, we proposed Strathern’s notion of partial connections as an analytical approach to empirically unpack dis/organizing digital infrastructures. This, we suggest, allows to explore the forging and transformations of social relations, with ensuing normative struggles, which in our case played out as data-ethical controversies over questions regarding anonymity and authorities’ use of well-being data. This approach, we suggest, offers insights into the relational, multifaceted forces and reverberations of digital data infrastructures, thereby including their unintended effects and shadow lives without underestimating their power. Our contribution advances understandings of exactly how such dis/organizing dynamics reinforce the performative powers of digital infrastructures and datafication in their (un)expected transformations of social relations. Below, we explore this in further detail.

Empirically, we examined dis/organization as partial connections through a case study of two data-ethical controversies related to a mandatory digital well-being survey for Danish students. The first analysis explored how local governments’ right to choose a digital survey tool of their own enabled them to make illicit use of personally identifiable well-being data. We unpacked the partial connections between several different digital information systems, which dis/organized the relationship between the ministry and local governments, forming normative struggles about data (mis)use. As shown, this involved disconnection from the legislation pertaining to the well-being survey, enabling local governments to continue collecting and using personally identifiable data alongside, and partially connected to, the national digital well-being survey. The second analysis showed how the ministry’s flawed information produced expectations of anonymity among parents and students, hence making the ministry’s storage of personally identifiable data controversial. Here, we elucidated how partial connections between ministerial and researchers’ databases made a subsequent promise of ensuring data deletion impossible. Moreover, to disconnect a student’s identity from the digital data infrastructure, students had to complete the survey on paper, rendering them visible and non-anonymous for the class, spurring normative struggles about what constitutes anonymity.

These findings produce fine-grained insights into the dis/organizing capacities of digital data infrastructures. On the one hand, a government-initiated survey functions as an organizing instrument for the ministry and local governments, rendering well-being measurable, comparable and commensurable across space and time. Simultaneously, (some) parents experience this as disorganizing, in terms of their experience of a loss of control over their children’s data. This dis/organization involves counter-productive forces of connectivity and disconnectivity, evolving in relation to one another during the digital data infrastructure’s ongoing formation process, prompting new relations and positionings as well as cuts and gaps in various infrastructural arrangements surrounding the survey. By foregrounding such dis/organizing dynamics of digital data infrastructures – highlighting how their partial connectivity changes relationships and generates normative struggles – we extend the scope of Cooper’s seminal work (1986) to organizational studies of (digital) technology. In particular, we add to Cooper’s understanding of partiality, where ‘anything we conceive of as a bounded thing is in fact generated from these boundaries, transforming itself through interaction with other partialities’ (Cooper in Spoelstra, 2005, p. 110), by unfolding it with Strathern’s notion of partial connection. Our conceptualization of dis/organization as partially connecting digital data infrastructures stresses their inherent dis/organization, and unintended shadow modes of organizing as powerful. In spite of the normative struggles, rendering the well-being survey a matter of public controversy, the local governments’ collection of identifiable well-being data and the State’s requirement for non-anonymous data have continued.

This demonstrates the importance of ‘partial connections’ as a concept in analytically unpacking the mutual making of digital data infrastructures and dis/organization. Rather than a simple contrast between the planned and unplanned in the organization of technology, the concept of partial connections analytically sheds light on infrastructures’ ‘double relationality’, entailing both inwards multiplicity and outwards spiralling out of control (cf. Harvey et al., 2017). As such, we add nuance to dominant understandings of power as something achieved through connectivity (cf. Kolb et al., 2020). Indeed, the dis/organizing partial connectivity of digital data infrastructures may even reinforce the powerful data collection, in spite of the normative struggles that ensue. Analysing dis/organization with the help of partial connections, we thus argue, advances studies on digital technologies by emphasizing their outwards-extending and partial connectivity, which transforms social actors and orders. This analytics allows us to understand digital relationality as emergent and working through, rather than in spite of, their unintended reverberations (Hesselberth, 2018; Mejias, 2013).

In order to understand and further explore the resilience of digital data infrastructures, we argue for the necessity to examine how they are dis/organizing. Their powers work not only through their capacities to reorganize relations, in our case between government, organizations and citizens, but also through their continual co-evolvement in response to infrastructural or legal interstices and normative struggles. This shows that gaps and disconnections, such as those between the ministry’s and researchers’ databases or between different data regulations, are not necessarily a threat to, but part of, such resilience (Munro, 1997). Attending to the mutual constitution of dis/organization of technology, the paper thus adds a critical focus on dis/organization to extant studies of technology as a new mode of order in ‘an organized world’ (Beverungen et al., 2019; Leonardi & Treem, 2020). Rather than an all-consuming digital organization, we need to take into account how digital data infrastructures through their inherent partial connectivity generate mess (e.g. ministerial miscommunication) and shadow systems (e.g. H&B), thereby creating not only unintended controversies, but also persistence and extensions of digital data infrastructures in dis/organizing ways.

These insights also contribute to studies of ‘governance by numbers’ (Espeland & Sauder, 2007). While much has been written about the organizing power of knowing and tracking the population through numbers, thereby making them governable from a distance (Desrosiéres, 1998; Miller & Rose, 1990), we know less about the instability and partiality of digital infrastructural arrangements. The well-being survey and its indicators generated new relationships of governance between the ministry, local governments and schools through well-being data, while also embodying a series of data-ethical crises that elicited an emerging mistrust of the State among citizens. As such, we gain empirical knowledge about the dis/organizing functions of changing calculative practices of governance by unpacking the partially connected modes of organizing across citizens, media, governing bodies and private sector companies. This offers knowledge regarding the unexpected reverberations of governing through numbers.

With respect to dis/organization studies (Plotnikof & Pedersen, 2019; Vásquez & Kuhn, 2019), we offer a concept to study the relational multiplicity (i.e. partial connections) of dis/organizing technologies. Importantly, disconnection in our view is not a-relational, but indeed relational through the appropriation of cuts and alongsidedness. These, too, transform relationships as well as intensify the ongoing dynamics of dis/organization. This approach qualifies a study of how connections and disconnections interweave and imbricate each other, complementing relational perspectives on communicative dis/organization as more than meaning, by including materiality (Ratner, 2020; Vásquez & Kuhn, 2019). Adding to this literature on, for example, how digitally connected communications materialize paradoxes (Porter & Jackson, 2019) and (in)visibility (Albu, 2019), our perspective facilitates exploring this as a matter of partial connectivity, with dis/organizing manifestations. This renders partiality a powerful communicative and socio-material form of relationality and may further studies on how digitally mediated communication fuel organizational struggles and novel forms of resistance.

Finally, building on discussions of power in dis/organization studies (Munro, 2003, 2005; Spoelstra, 2005), we echo their stressing of the always-already disordered aspect of any ordering force. With the concept of partial connections, we found that exploring connectivity as well as cuts, gaps and disconnections produce critical insights into how a dis/organizing technology may differentiate relations and cause normative struggles (e.g. a data-ethical controversy) that are neither controllable nor equally distributed (Munro, 2001; Plotnikof & Pedersen, 2019). Our perspective underlines that digital data infrastructures are not just technicalities within which organizational politics unfold. Rather, organizational politics are inherent to digital data infrastructures and their partial connectivity, with divergent concerns intertwining in processes of dis/organization that reconfigure not only infrastructural arrangements but also social relations. The implications for organizational actors such as governments or industry include the need to critically reflect on how digital data infrastructures also rely on – and result in – unexpected disorganization, when chasing dreams of big data and data-driven organizing. Avenues for future research may well involve advancing theory of such dis/organizing forces to enhance knowledge of the changing, ambiguous modes of organizational governance and power that are increasingly bound up with the dis/organization of partially connecting digital data infrastructures.