A Controversial Digitalization Strategy for the Police’s Crime Prevention in Denmark

Abstract

Data-driven strategies that leverage digital technologies for task improvement are increasingly being adopted across organizations. However, theoretical and empirical insights into how such strategies are implemented—and the organizational tensions they may generate—remain scarce. This study explores how a digital, data-driven strategy is interpreted and enacted within a complex organizational setting. We examine in a qualitative case study the Danish National Police’s digitalization strategy for a shift from reactive crime response to proactive crime prevention. Theoretically, the study is based on institutional theory, in particular, institutional logics (such as the state, corporation, and profession) that may underlie such new strategies. A qualitative case study was conducted drawing on document analysis, a review of key empirical studies, and additional interviews and meetings during strategy implementation (2013–2022). The findings reveal that the implementation process was shaped by divergent interpretations of the new data-driven strategy, rooted in institutional logics. The different interpretations surfaced underlying tensions about organizational priorities and practices. The theoretical novelty of the study is that we contribute to the concept of intra-institutional complexity, showing how conflicting interpretations of a single institutional logic—rather than clashes between distinct logics—can generate significant organizational friction. We identify three forms of such complexity: mission dilemmas, resource allocation challenges, and identity pressures. These findings advance the understanding of how digital strategies may unfold in practice and highlight the interpretive flexibility—and potential contestation—of institutional change within organizations.

1. Introduction

The concept of a data-driven organization is prevalent in many organizations that aim to strengthen their strategic planning and enhance governance procedures using digital technologies for data collection (Brayne, 2017; Hardyns & Rummens, 2018; Zeng & Glaister, 2018). Risk assessment is an important component of data-driven strategies. For example, in banking, services are based on an analysis of customer group credibility, and in public health, prevention efforts are directed at citizen groups who, according to risk assessments, may develop lifestyle-related diseases. However, the meaning of new strategic concepts has not been provided (Pallas et al., 2016; Zilber, 2002). Meanings can reflect different values in an organization’s social context and can be used for various local interpretations (Van Grinsven & Heusinkveld, 2023; Vossen & Van Gestel, 2019; Waldorff & Madsen, 2023). Thus, when an organization implements a new approach, actors’ interpretations may evoke different and potentially conflicting interpretations.

Although the implementation of strategies for creating data-driven organizations, often with advanced digital technologies providing big data, is one of the most profound transformations in today’s organizational practice, empirical research on this shift and how actors’ interpretations influence implementation is limited (Hardyns & Rummens, 2018; Marciniak, 2021). This study aims to fill this gap in the literature. Our empirical case focuses on the Danish National Police, which has pursued a data-driven approach since 2013, officially introduced as the “National Strategic Analysis”. The policing field is particularly subject to transformation due to emerging threats (e.g., cybercrime), increasing public demands for safety (e.g., at large events and demonstrations), and a growing need for proactive crime prevention (e.g., terrorism and neighborhood-related issues). Police organizations are receiving growing attention in both political and academic discourse, particularly considering ongoing concerns related to security, migration, climate change, poverty, drug policy, and economic instability. Moreover, police forces themselves are subject to the social, demographic, and value-based transformations occurring within their own institutions (Schröter et al., 2023). The organization’s commitment to data-driven analysis reflects a broader international trend, rendering this case especially timely and relevant.

Internationally, data-driven decision-making in police organizations dates back to the beginning of the twentieth century (Fest et al., 2023). Since then, the police have become increasingly dependent on a wide variety of information systems, which, among other things, play a role in planning police activities (Meijer et al., 2021), providing officers on the street with information about citizens and locations, and registering police actions (Manning, 2008; Fest et al., 2023). The USA and England have developed approaches, models, and methods for police work, and data-driven management has been introduced (Hestehave, 2013; Marciniak, 2021), although data-driven decision-making in the USA has only been systematically incorporated into law enforcement practices in recent decades (Brayne, 2017). The practical use of data-driven technologies in Europe has recently evolved through several pilot programs (Hardyns & Rummens, 2018; Jansen, 2018).

The Danish National Police’s initiative for a data-driven organization is in line with common concepts in other countries, such as “smart policing”, “intelligence-led policing”, or “predictive policing” (Ratcliffe, 2008; Ferguson, 2017). Like other European countries, the aim is “to collect data in real time, analyze heterogeneous databases, and support the police in their investigation or predict crime” (Jansen, 2018, p. 10). The main advantage of data-driven analysis in policing compared to more traditional methods is the use of time and spatial dimensions to predict and prevent crime (Hardyns & Rummens, 2018). Therefore, a data-driven strategy should reduce crime rates and prevent them rather than merely cure them (Memmert, 2016). Despite these promises, there are also concerns about the premise that data-driven technologies have been developed, namely that the police can monitor everyone in public spaces and act based on data (Brayne, 2017). Questions have been raised about privacy, shortcomings in data practices, and a lack of clarity about the criteria and processes for including individuals in databases (Hardyns & Rummens, 2018; Jansen, 2018). It has also been suggested that digital technologies can make work more efficient but also slow it down (Marciniak, 2021). Moreover, the rationale behind adopting new technologies in policing may be a response to other reform forces, such as budget cuts, problematic policing behavior, and systemic biases, but they do not necessarily solve these problems (Ferguson, 2017).

Therefore, it is necessary to investigate how data-driven policing programs are applied and their organizational consequences. In doing so, it is important to examine cases considering the wider institutional context of the country, organization, and specific profession (Jansen, 2018; Marciniak, 2021). As mentioned above, our analytical focus in this article is the Danish police’s implementation of a national data-driven digital approach. The National Police of Denmark is a traditional bureaucracy with well-established procedures and routines. The police’s ambition with the data-driven approach was to enable the organization to enhance safety and security for all citizens when crime, such as terrorism, cybercrime, trafficking, and economic crime, has become more complex and crosses national borders (D23 Visionsudvalget, 2005). Ideally, data-driven organizations should identify current threats and crime patterns using large amounts of data, followed by recommendations on preventing crime rather than just solving crimes after they have been committed. Following our aim to explore how a digital, data-driven governance concept is interpreted and enacted within a complex organizational setting, the research question is as follows: how is a new data-driven strategy interpreted by internal actors in the police, drawing differently on institutional logics, and what possible tensions do these interpretations lead to?

The theoretical background of this study is institutional theory. Institutional scholars have long investigated the implementation of strategic concepts in organizations. Studies of field-level dynamics have shown how concepts are diffused and implemented by organizations to legitimize responses to external institutional pressures (DiMaggio & Powell, 1983; J. W. Meyer & Rowan, 1977). In this study, we focus on the literature on institutional logics that helps investigate the organizational principles of organizations within a field (Friedland & Alford, 1991; Reay & Hinings, 2009). Our study applies a theoretical framework with three logics (related to the state, corporation, and profession), showing how, viewed from each of these logics, the ambitions of a data-driven police organization are interpreted differently, creating latent tensions about how police should work in practice. We contribute to the discussion of “institutional complexity” (Greenwood et al., 2011) by theorizing three types of intra-institutional complexity—mission dilemmas, resource allocation, and identity pressure—that can arise not only between logics but also as initiated by each logic, depending on different interpretations in the field.

In the next section, we present our theoretical framework. Then, we further describe the case of the Danish national police and our methods. The findings reveal that the implementation process of the new data-driven strategy was shaped by divergent interpretations, showing underlying tensions within the police system about organizational priorities and practices around the organization’s mission, resource allocation, and identity. Theoretical and practical contributions of this study are discussed at the end of the paper.

2. Theoretical Framework

The theoretical framework draws upon the literature on “institutional logics” (Thornton et al., 2012). Institutional logic is a set of principles for understanding organizational reality and guidelines for appropriate and legitimate behavior in a given context (Greenwood et al., 2011). Friedland and Alford (1991) introduced the concept of institutional logic in organizational studies, arguing that a wider social context must be included in the analysis to understand and explain individual and organizational behaviors. Following Friedland and Alford (1991), Thornton et al. (2012) theorized which institutional logics are particularly prominent in the Western world, discerning seven types of logic: state logic, corporation logic, profession logic, market logic, community logic, religion logic, and family logic.

Not many studies employ institutional logic to explore the police (for a recent exception, see a study of a local police department in Italy by Sorrentino et al., 2023). A few scholars have investigated related organizations, such as Purdy and Gray (2009), who explored how a bureaucratic and democratic logic enabled multiple practices to become institutionalized in state offices of dispute resolution, and McPherson and Sauder (2013), who examined how professionals use the logic of criminal punishment, rehabilitation, community accountability, and efficiency to negotiate decisions in a drug court. For our analysis of the data-driven approach of the Danish police, we identified three ideal-type logics (Reay & Jones, 2016)—state logic, corporate logic, and profession logic—as particularly relevant for the organization. Each logic emphasizes a specific core principle, a different basis for legitimacy and authority, and a particular focus on strategy. State logic promotes the ideal of democratic governance as the proper mechanism for allocating and safeguarding resources in society, which is legitimized and upheld through citizens’ democratic participation, such as voting, with the overall purpose of increasing community goods or all citizens’ welfare. Corporate logic promotes hierarchy as the appropriate form of organization, where the focus is on the organization’s performance and efficiency, to increase the organization’s (marked) position. Profession logic promotes the importance of professionals’ relational networks, as these strengthen their expertise and ensure the quality of work.

Table 1. Three ideal types of institutional logics.

Logic Characteristics	State Logic	Corporation Logic	Profession Logic
Core principle	State as a redistribution system	Corporation as a hierarchy	Profession as a relational network
Basis of legitimacy	Democratic participation	Organizational position	Personal expertise
Basis of authority	Bureaucratic domination	Top management	Professional association
Basis of strategy	Enhance the common good and safeguard social and economic stability	Improve organizational performance and increase efficiency	Increase professional reputation and ensure the quality of the work

Source: Derived and adapted from Thornton et al. (2012), Table 3.2, p. 73.

lists the main features of each logic type used.

Over the past three decades, many studies have drawn on institutional logic to understand organizational (field) change. First, studies have focused on explaining the shift from one dominant logic to another in the organizational field, from professional logic to state logic. Gradually, more studies discussed multiple, co-existing logics in an organizational field, leading to “institutional complexity” (Greenwood et al., 2011). Recently, the study of “intra-logics complexity” has been proposed, pointing to the idea that each logic can have different meanings when interpreted in practice (R. E. Meyer & Höllerer, 2014, 2016). Below, we elaborate on these three phases in the institutional logic literature as the basis for our empirical analysis.

2.1. Shift from One Dominant Logic to Another

Most empirical research on institutional logic has assumed a shared meaning in organizational fields, that is, one institutional logic or a shift from one logic to another (e.g., Marquis & Lounsbury, 2007; Purdy & Gray, 2009; Scott et al., 2000).

For example, Thornton and Ocasio (1999) studied the shift in logic in the higher-education publishing industry from the institutional logic perspective. Historically, the dominant editorial (professional) logic shifted to market logic, profoundly impacting internal relationships and the purpose of executive succession. Van Gestel and Hillebrand found ongoing shifts in dominant logics over time in the Dutch national employment services field, indicating that “outcomes may also be characterized by ongoing change rather than stable fields with one dominant or multiple co-existing logics. In other words, while one dominant logic may emerge, it does so only temporarily, and one change is followed by another” (Van Gestel & Hillebrand, 2011, p. 233).

2.2. Multiple Co-Existing Logics

Gradually, institutional scholars have considered the multiplicity of coexisting institutional logics and theorized how organizations respond to conflicting demands in their institutional environment. While previous literature has focused on substituting one dominant logic for another, an alternative conceptualization was developed, assuming that actors can develop collaborative strategies supporting competing logics’ co-existence. For example, Reay and Hinings (2009) found the co-existence of professional and market logic in the healthcare field of Alberta, Canada, pointing to collaborative strategies by management and medical experts to manage the rivalry between the two logics. Waldorff et al. (2013) examined healthcare reforms in two countries and theorized how different “constellations of logics” impact the design and accomplishment of national initiatives. Logic can simultaneously constrain and enable actions. Lattu and Cai (2023) found competing institutional logics in Open Science practices of university-industry research collaboration, with tensions between the market logic, the state logic, and the profession logic.

Recognizing multiple logics being similarly active also led to new concepts, such as “hybrid organizations” (Battilana & Dorado, 2010; Denis et al., 2015), “institutional bricolage” (Christiansen & Lounsbury, 2013) and “institutional complexity” (Greenwood et al., 2011; Raynard, 2016) that occurs when “the prescriptions and proscriptions of different logics are incompatible, or at least appear to be so, they inevitably generate challenges and tensions for organizations exposed to them” (Greenwood et al., 2011, p. 18). However, institutional logic is not given but is “filtered” by the organization—particularly its position within a field, structure, ownership, governance, and identity (Raynard, 2016). Studies of multiple, coexisting logics emphasize a dynamic understanding in which organizational actors can move between and draw on different logics, although they work together in the same organization (Berg Johansen & Waldorff, 2017).

2.3. Intra-Institutional Complexity

For a long time, most studies on institutional logics assumed that these should be understood as one-dimensional. It has been suggested that institutional logics, such as state, corporate, or professional logic, each have a clear message or substance. More recently, the unequivocal meaning of institutional logic has been questioned, and attention has been drawn to potentially different interpretations (R. E. Meyer & Höllerer, 2014). With a closer look, the logic itself may be interpreted in many ways and induces institutional complexity. Thus, moving beyond the inter-institutional dimension, a call has been made for analyzing the “intra-institutional” dimension to add depth to institutional studies (R. E. Meyer & Höllerer, 2014, 2016).

Elaborating on the idea that logic may be interpreted differently, scholars have explored the empirical practice of logic in concrete organizational settings. These studies highlight the mediating role of professional actors in understanding organizational change (Bévort & Suddaby, 2016; Pallas et al., 2016). Pallas et al. studied media activities in a governmental agency and identified three local, profession-based value systems that shaped a new media logic’s translation and organizational embeddedness. Other studies point to social positions and identity as crucial in understanding how interpretations of logics affect organizational change (Currie & Spyridonidis, 2016; Høiland & Klemsdal, 2022; Reay et al., 2017). For example, Høiland and Klemsdal (2022) explored how the complexity of contemporary professional work and services was addressed differently by managers at the strategic level and professionals at the operational level. They demonstrate how significant conflicts stem from differences within organizations in how logics are handled. Some studies have examined the role of everyday practices and social interactions (McPherson & Sauder, 2013; Smets & Jarzabkowski, 2013) in capturing the impact of institutional logic from a bottom-up, interactive perspective. Vossen and Van Gestel (2019), for example, found that legislation in line with a managerial logic for rehabilitation was interpreted differently within organizations. Ill workers received more support for their return to work when employers perceived positive social interaction. Furthermore, Voronov et al. (2013) underscore that logics might provide multiple cues that actors use for making sense of their reality and determining what actions are appropriate: ‘Actors must be selective not only in which logic to adhere to, but also, possibly, in how to adhere to a particular logic’ (Voronov et al., 2013, p. 1564). These studies show that institutional logics are not reified cognitive structures but interpreted and given meaning by actors within specific contexts (Zilber, 2002; Smets & Jarzabkowski, 2013).

Despite growing attention on how logic is translated in practice, different interpretations and strategies to deal with complexity are attributed to factors other than the logic characteristics, such as the identity and role of the involved actors. When reviewing the literature, it is surprising that little research has focused on how a logic itself is internally connected (Lounsbury et al., 2021) and has an unambiguous meaning. As indicated in Table 1, Thornton et al. (2012) emphasized how logic consists of several core elements related to sources such as legitimacy, authority, and the basis for strategy. However, a detailed understanding of how the different logic elements are interrelated is still missing (Friedland, 2012; Lounsbury et al., 2021), and how different logic interpretations may impact the logic’s core elements. This study contributes to the theoretical challenge of dissecting a single logic. To this end, we analyze the possible variation in interpreting the three logics in an empirical study and conclude the discussion on intra-institutional complexity.

Before turning to our findings, we introduce our research setting and methods.

3. Methods

3.1. Research Setting

As an organization, the National Police in Denmark has existed for more than 300 years (Degnegaard & Waldorff, 2013). Throughout its long history, the organization has undergone minor changes, but in 2007, the so-called police reform came into force, implying a major centralization (D17 Ministry of Justice, 2006). The former 54 police districts were reduced to 12, and all local commissioners were referred to the National Commissioner. The aim of this reform was to strengthen control and coordination and monitor performance and best practices. The reform was a response to growing expectations from politicians and citizens that the police should carry out more new types of functions, improve the level of service, and be more cost-effective and cost-reducing. In 2011, an internal initiative, namely, the Monitoring and Analysis project (D3 Danish National Police, 2011), concluded that it was important to develop analytically based police work or other preventative methods (Memmert, 2016). Following this initiative, major changes have occurred since 2013 in recruiting strategic analysts, establishing central and local analytical units, providing courses for police officers in intelligence analysis, introducing new standards and guidelines for producing data-driven analysis locally, and introducing new IT tools.

In this broader context, the first “National Strategic Analysis” of the National Police was published in 2015 (D7 Danish National Police, 2015c). This document outlined several objectives for 2016–2020, including the fact that police work should be more strongly based on knowledge and analysis to prevent and reduce problems. Local strategic analyses were conducted in all police districts to supplement the national report. The new data-driven strategy led to significant staff growth: in 2015, the Danish police organization had approximately 14,000 full-time employees, and in 2020, this number was almost 17,000 (D12 Danish National Police, 2023). Traditionally, the largest group consisted of police officers, but in line with the police’s ambition for a data-driven strategy focused on crime prevention, the police increasingly employ other professionals, such as economists, psychologists, IT experts, communications consultants, sociologists, and anthropologists.

In 2017, the data-driven strategy of the Danish National Police was followed by the subsequent National Strategic Analysis, covering 2017–2022 (D11 Danish National Police, 2017).

However, in 2020, we noticed a significant change in political support for analytical academic work by the police. Prime Minister Mette Frederiksen stated in her opening speech to the Danish Parliament that the National Police had to clean up the so-called pseudo-work. She argued that too many people were employed centrally instead of close to the citizens. Shortly after, a broad multi-year agreement for the police came into place (D20 Ministry of Justice, 2020). The political agreement affected the number of academics in the National Police. Contrasting with the earlier data-driven strategy that led to a wide range of (academic) specialists at the central level, there were significant transfers of person-years from the National Police to local police districts (D15 Ejlersen, 2023). Thus, the National Strategic Analysis 2017 (D11), covering the period of 2017–2022, was the last centrally developed strategic analysis. It then became up to the local police districts to develop local strategies for data-driven and intelligence-led policing (D2 Bjørnholdt, 2023, p. 1). The research setting and timeline in this subsection serve as a background for our empirical analysis.

3.2. Data Collection

This study employs a document analysis approach, drawing primarily on official texts related to the implementation of strategic analysis in the Danish National Police. Moreover, three key empirical studies were used as additional analytical input, and interviews and group meetings were used to inform and contextualize the design of the study and the collection of documents.

3.2.1. Document Sample

Document analysis, as a qualitative method, involves a systematic procedure for reviewing and evaluating printed and digital materials (Bowen, 2009; Prior, 2002). Our selection of documents was purposive and theory-driven, guided by the research aim to understand how strategic analysis was developed, communicated, and interpreted within the police organization. We identified documents that were (i) either produced by the Danish National Police or closely affiliated governmental bodies, and (ii) explicitly addressed strategic analysis, crime prevention, or the organization of police work. Documents were collected from official websites (e.g., the Danish National Police and the Ministry of Justice) and internal repositories (when accessible). Selection criteria followed principles recommended by Bowen (2009), including authenticity, credibility, representativeness, and meaning. To ensure analytical depth and coherence, we prioritized materials that were frequently cited, widely distributed within the organization, or referenced in official policy and planning documents. Key documents were the annual “National Strategic Analysis”, signaling the prioritized tasks of the police. These annual reports examined current threats and developments for crime prevention and reduction and pointed to future challenges. Each report focused on the next 5–6 years. For example, the report from 2017 assessed the years 2017–2022. Next to the annual reports, the National Police published in 2015 a document to make transparent which methods were used for the National Strategic Analysis reports, but also data uncertainties and organizational challenges in producing comprehensive data were described. In addition to these key documents, a wider collection of national and local official documents was studied concerning the organizing of police work related to the strategic analysis, for example: the National Police’s mission statements, central and local strategic plans, annual reports, performance contracts and evaluations; Finally, webpages and media texts engaging with the national strategic analysis were collected. The total sample of 23 core documents, totaling about 600 pages as the main basis for our analysis, is listed below. It contains police reports (8), police websites (2), local strategic plans for the police in North Jutland, Midt and West Jutland, and Funen (4), media sources (3), and Reports from the Ministry of Justice (5) and the National Audit Office (1):

Police websites

D1	(Bjørnholdt, 2013)
D2	(Bjørnholdt, 2023)

Police Reports

D3	(Danish National Police, 2011)
D5	(Danish National Police, 2015b)
D7	(Danish National Police, 2015c)
D8	(Danish National Police, 2015a)
D9	(Danish National Police, 2016a)
D10	(Danish National Police, 2016b)
D11	(Danish National Police, 2017)
D12	(Danish National Police, 2023)

Local Strategic Plans

D4	(Danish National Police, 2013)
D6	(Danish National Police, 2015d)
D16	(Midt- and West Jutland Police, 2017)
D22	(Tryg Aalborg, 2014)

Media sources

D13	(EDRi, 2014)
D14	(EDRi, 2017)
D15	(Ejlersen, 2023)

Ministry of Justice

D17	(Ministry of Justice, 2006)
D18	(Ministry of Justice, 2015a)
D19	(Ministry of Justice, 2015b)
D20	(Ministry of Justice, 2020)
D23	(Visionsudvalget, 2005)

National Audit Office

D21	(Rigsrevisionen, 2021)

3.2.2. Empirical Data

We reviewed existing empirical literature on the Danish National Police and selected three key empirical studies purposively to build on in our study, based on their relevance to our focus on strategic analysis in contemporary policing and their empirical richness (Hartmann, 2014; Hestehave, 2013; Memmert, 2016). All three studies are peer-reviewed and grounded in qualitative methodologies, such as ethnography, interviews, or document analysis, offering in-depth accounts of how strategic work is enacted and made meaningful within policing contexts. The empirical studies served as an additional analytical input for this study, next to our document analysis. More specifically, we included the studies’ findings and empirical observations as illustrative examples of how strategic analysis is enacted in the Danish police.

The first author also conducted two individual interviews and two group meetings, involving a total of six participants. The conversations were open-ended but addressed selected themes such as the aims of the Police’s data-driven strategy, the design of the implementation process, and the organizational challenges emerging in the change process. The individual interviews each took 45 min, and the group meetings lasted approximately two hours. Their purpose was to provide valuable insights that helped refine our analytical focus and inform the selection of additional documents for review. They were not intended to constitute part of the formal dataset subjected to systematic analysis, but each interview and meeting contributed specific perspectives relevant to distinct aspects of the study.

The first interview (2013) with the National Police Commissioner served to obtain a detailed understanding of the strategy: its aims, context, and motivations of the National Police. In 2015, focusing on the implementation of this strategy, we held two group meetings with four key employees of the National Police. These officers were selected for being closely involved in developing and implementing the data-driven approach throughout the country. They were key actors being tasked with developing procedures for implementation and criteria for local delivery. The group meetings served to provide in-depth background information not only about how the new strategy should be enacted but also about how it was (differently) interpreted in practice. The group meetings also pointed to further relevant documents to understand the impact of the new strategy on the police organization. After the data-driven strategy was severely criticized by the new prime minister (2020) and shifted to decentralization and local decision-making, the first author conducted an interview (2021) with a former strategic analyst to reflect on the strategy process since 2013 and its changes in hindsight.

To address the ethical considerations embedded in the research design, interviewees were informed of the research purpose and questions in advance. The first author assured conversational partners confidentiality and anonymity and asked for consent before they participated in the study.

3.3. Data Analysis

Building on our theoretical framework (Thornton et al., 2012), we used the concept of institutional logics as an interpretive lens to analyze how organizational change is represented, justified, and contested in official texts. Our analytical approach combined inductive coding with pattern matching (Reay & Jones, 2016), enabling us to trace how different logics were embedded and interpreted in strategic and administrative documents. This approach involves applying ideal-type logic characteristics to capture and distinguish each institutional logic (see Table 1). The benefit of this approach is exploring and distilling how institutionalized meanings are enacted and interpreted in empirical practice.

First, we conducted a close and iterative reading of the selected documents, allowing patterns to emerge, looking for various meanings associated with implementing a data-driven approach in the police organization. This initial coding process involved identifying recurring discursive elements, such as rationales for change, conceptions of professionalism, and justifications for performance measurement. These first-order codes were grounded in the language and structure of the documents. Following the principles of document analysis (Bowen, 2009; Prior, 2002), we treated texts not as neutral reflections of practice, but as socially constructed artifacts that communicate institutional values, power structures, and modes of governance.

In the second stage, we moved from descriptive codes to analytical categories by clustering related codes that reflected underlying logics. Following our theoretical framework, we identified in our data coding three institutional logics impacting the change process: the state, corporation, and profession. We found that each institutional logic promoted certain core principles and bases for legitimacy, authority, and strategy. Codes referencing legal obligations, public interest, and formal procedures were grouped under the state logic; codes emphasizing productivity, quantifiable outcomes, and managerial control aligned with a corporate logic; while those highlighting expert knowledge, professional autonomy, and training standards were categorized as part of a profession logic.

In the final phase of analysis, we examined how the three identified logics intersected and diverged across different document types and organizational levels (see Table 2, Results section). We paid particular attention to tensions and contradictions, especially in how various actors interpreted the promise and risks of data-driven organizing in relation to established and new practices. These interpretive differences revealed how institutional logics coexisted and—even more interesting—that each logic gave rise to different and incompatible interpretations. In this last phase, we also searched for existing empirical studies of strategic analysis in the Danish police that could inform our work. After a close reading of three studies and reflection on their findings, we decided to include illustrative examples from their work. Finally, we structured our findings around each logic and the divergent interpretations related to (i) established and (ii) new practices.

Validity determines if the research findings are accurate (Yin, 2014; Creswell & Creswell, 2018). We realize that the ideal types of institutional logics are not reality but representations of reality. How a logic is interpreted and enacted always depends on the specific empirical context. But using the ideal types (see Table 1) to detect various interpretations can sharpen attention to (conflicting) institutionalized norms and values, which may otherwise remain less visible. To increase validity, the research uses multiple sources of evidence in the data collection, including interviews, group meetings, and secondary data. Key informants were involved in the research. External validity (generalisability) implies defining the domain to which a study’s findings can be generalized. For this matter, we use institutional theory and institutional logics perspectives in our single case study.

Reliability, demonstrating that the operations of a study, such as data collection, can be repeated, with the same results, is achieved by developing a case study database (Yin, 2014; Creswell & Creswell, 2018), see the Documents overview above. The coding reliability was safeguarded through a rigorous process in which the authors discussed the codes to ensure that the data were relevant for answering the research question and that the analyses and findings were transparent. To capture the complexity in the change process, we included various documents and interpretations by more than one person or group of actors, “allowing the story to unfold from the many-sided, complex and sometimes conflicting stories” (Flyvbjerg, 2006, p. 21). In terms of the wider relevance of the study, we believe our case was a critical one “having strategic importance in relation to the general problem” (Flyvbjerg, 2006, p. 14). This means that if this top-down controlled police organization facilitated multiple interpretations of its data-driven strategy, this would probably also be the case for other (hierarchical) organizations.

4. Results

The following three subsections present how the data-driven organization strategy is understood differently within the National Police organization. We thus perceive intra-institutional complexity (multiple interpretations of single logics) in our study. Interpreted in light of state logic, corporation logic, and professional logic, our findings highlight that each logic may have different meanings, as illustrated in Table 2 and elaborated on below.

Table 2. Ideal type institutional logics, challenged for multiple meanings.

	Case: The Data-Driven Strategy Of the Danish Police
Institutional logic	Meaning 1	Meaning 2
State logic	Strategic analysis will contribute to the police mission to prevent crime and threats. It is expected that the police work will change gradually in the future so that there—besides traditional case work—also will be increasing demands that the Police is capable of handling and minimizing risks. The police should not only respond to crimes that already have occurred. (D10 Danish National Police, 2016b, p. 79)	Strategic analysis will challenge the police mission to respond to acute crime and fears. The police have not drawn up a strategy or set goals to implement more analysis-based patrolling and thereby achieve a higher degree of targeted patrolling, even though more analysis-based patrolling was one of the goals in the police’s multi-year agreement for 2016–2019 (extended to 2020). (D21 Rigsrevisionen, 2021, p. 3)
Corporation logic	Strategic analysis will contribute to improving organizational performance and efficiency. We support the preventing and crime-reducing work by a targeted optimization of our crime analysis and threat assessments, so we can allocate our resources to those areas where the need is greatest. At the same time, we will strengthen both the cross-district and local investigation. (D9 Danish National Police, 2016a, p. 4.)	Strategic analysis will challenge organizational performance and efficiency. […] It is not that this data basis is without limitations. One reason is that the data are collected for administrative purposes and therefore it is not collected to answer the problems that we may want to examine in the present [strategic] analysis. (D5 Danish National Police, 2015b, p. 6).
Profession logic	Strategic analysis will strengthen police professionals’ analytic knowledge and preventative methods. In our search for new and more effective ways to solve our core tasks, we must involve and engage citizens, other authorities, and businesses. This should ensure us new knowledge, other competences, and more resources to draw on. (D9 Danish National Police, 2016a, p. 7)	Strategic analysis will challenge police professionals’ experience-based knowledge and responsive methods. As a result of the development of crime, the police must increasingly deal with problems related to increased digitization and computer power, encryption, anonymization […]. This means that the police tasks also become more complex and must be expected to require new skills and collaborations. (D11 Danish National Police, 2017, p. 165)

Table 2. Ideal type institutional logics, challenged for multiple meanings.

	Case: The Data-Driven Strategy Of the Danish Police
Institutional logic	Meaning 1	Meaning 2
State logic	Strategic analysis will contribute to the police mission to prevent crime and threats. It is expected that the police work will change gradually in the future so that there—besides traditional case work—also will be increasing demands that the Police is capable of handling and minimizing risks. The police should not only respond to crimes that already have occurred. (D10 Danish National Police, 2016b, p. 79)	Strategic analysis will challenge the police mission to respond to acute crime and fears. The police have not drawn up a strategy or set goals to implement more analysis-based patrolling and thereby achieve a higher degree of targeted patrolling, even though more analysis-based patrolling was one of the goals in the police’s multi-year agreement for 2016–2019 (extended to 2020). (D21 Rigsrevisionen, 2021, p. 3)
Corporation logic	Strategic analysis will contribute to improving organizational performance and efficiency. We support the preventing and crime-reducing work by a targeted optimization of our crime analysis and threat assessments, so we can allocate our resources to those areas where the need is greatest. At the same time, we will strengthen both the cross-district and local investigation. (D9 Danish National Police, 2016a, p. 4.)	Strategic analysis will challenge organizational performance and efficiency. […] It is not that this data basis is without limitations. One reason is that the data are collected for administrative purposes and therefore it is not collected to answer the problems that we may want to examine in the present [strategic] analysis. (D5 Danish National Police, 2015b, p. 6).
Profession logic	Strategic analysis will strengthen police professionals’ analytic knowledge and preventative methods. In our search for new and more effective ways to solve our core tasks, we must involve and engage citizens, other authorities, and businesses. This should ensure us new knowledge, other competences, and more resources to draw on. (D9 Danish National Police, 2016a, p. 7)	Strategic analysis will challenge police professionals’ experience-based knowledge and responsive methods. As a result of the development of crime, the police must increasingly deal with problems related to increased digitization and computer power, encryption, anonymization […]. This means that the police tasks also become more complex and must be expected to require new skills and collaborations. (D11 Danish National Police, 2017, p. 165)

We now elaborate on the different interpretations of each of the three logics.

4.1. The Data-Driven Organization as Expression of a State Logic

The formal purpose of the National Police is to maintain safety, security, peace, and order. This was evident in the Act on Police Activities, Section 1, which also stated that the police should further this purpose through prevention, assistance, and law enforcement (D18, Ministry of Justice, 2015a). The National Police’s business strategy for 2016–2020 was built upon this raison d’être: “We will continuously make sure that our goals and activities ensure the citizens’ safety and security” (D9 Danish National Police, 2016a, p. 3). Thus, the police force’s purpose can be understood as a manifestation of state logic, where citizens’ welfare, democratic rights, and safeguarding society are legitimate purposes (Thornton et al., 2012).

4.1.1. The New Practices

Yet, the police force has been under increasing pressure. New forms of crime and security threats arose, such as terrorism, gangs, and cybercrime, which called for an organization that can act. Politicians and citizens also placed greater demands on the police regarding efficiency and quality, just as citizens were less faithful to the authority than before, questioned the legitimacy of police practice to a greater extent, and required reasons for police actions (Memmert, 2016). The aim of developing knowledge-based practices has emerged in organizational strategy papers, as demonstrated by the Danish National Police: ‘We support the preventive and crime-fighting work by purposefully improving our crime analyzes and threat assessments so we can use our resources where the need is greatest.’ (D9 Danish National Police, 2016a, p. 4)

The argument in this interpretation of a state logic was that a more systematic knowledge management would help the police become better equipped to deal with an increasingly complex and demanding crime picture and provide transparency in police work. This nurtured the interpretation of a data-driven organization as an adequate tool to strengthen security and safety by gathering information not only from within the police organization but also in dialogue with external stakeholders. One example is the district and local councils that formed the basis for an ongoing dialogue between police and municipalities regarding cooperation in crime prevention. At meetings, the police explained the overall crime pattern in the district, supplemented with municipal authorities’ assessments, and discussed the need for further action. One example is the police’s intention to reduce burglaries in collaboration with the citizens, e.g., to secure their own homes and establish a neighborhood watch, as described in a local police document (D22 Tryg Aalborg, 2014, p. 5). Dialogue with external stakeholders and increased public transparency in crime and threats would strengthen the insight of citizens into how the police works and sustain collaboration between the state and civic society.

4.1.2. The Established Practices

However, a completely different interpretation of the state logic is related to what is viewed in practice as the “nature” of police work. Part of the police mission portfolio requires a quick response to acute incidents to ensure citizens’ safety and security. Therefore, the police need to initiate immediate efforts when a newly emerging problem is recognized. A quick response to burglaries, robberies, inter-personal violence, etc., of the police is very visible to the public and gains direct legitimacy because they “do something”. This adds to the understanding that the traditional perception of good police work is patrol and action work rather than strategic analysis. In contrast, the data-driven strategy of prevention of large-scale, boundary-spanning crimes such as economic crime, cybercrime, international gangs, and terror is often less visible. Therefore, even though preventative efforts may be most effective, the traditional perception of good police work as patrol work and action-oriented became a barrier to a more preventative and analytical approach. Acute events overrule everything (Memmert, 2016).

In sum, we found that a state logic, in which the mission is to ensure the security and safety of all citizens, can be interpreted differently. Data-driven organizing was a new way to accomplish this mission, but it was viewed by the police organization in two conflicting ways: (i) as strengthening the police’s knowledge and prevention of crime and qualify its dialogue with the wider public and (ii) as drawing attention away from the traditional police mission, which is to respond quickly to threats and crime.

4.2. The Data-Driven Organization as Expression of a Corporation’s Logic

The data-driven approach can also be understood as part of a contemporary rationalization agenda in line with corporate logic, where organizational efficiency and performance are legitimate measures (Thornton et al., 2012).

4.2.1. The New Practices

The political agreement on the police and prosecutors’ economy in 2016–2019, concluded by the governing Liberals and the Social Democrats, Danish People’s Party, Liberal Alliance, and Conservatives, included a wide range of initiatives to streamline efforts, such as the following:

Optimizing the operational work by better planning and management of resources, better use of staff’s waiting time and time spent at the station, more analysis-based patrolling, and new IT tools, which can move case management out to the crime scene.

(D19 Ministry of Justice, 2015b, p. 10)

In addition, the agreement included “better case management and a consistent, high-quality investigation by creating common standards, strengthening case classifications, and gathering the investigation in larger units” (D19 Ministry of Justice, 2015b, p. 10). This ambition to optimize the police organization and strengthen knowledge-based practice included target and performance plans, contract management, performance measurement, creating administrative and police operational task communities, and leadership training (Hestehave, 2013). As a superintendent from the HR unit of the National Police explained:

All areas in the police organization need strong analysts—from the traffic area to economic crime. The demand will not diminish over the years—I am thinking, among other things, on the new Intelligence- and Analysis Unit in the local districts, which is one of the flagships of the efficiency process.

(D2 Bjørnholdt, 2013, p. 20)

The strengthening of the police’s analytical capacity and the systematization of police work added to the picture of a police organization capable of prioritizing tasks effectively. The ideal police force was a coherent organization with a tight steering hierarchy, where instructions flowed top-down through plans, contracts, and agreements. Strategic analysis can monitor how the police have tackled current societal challenges and the effects of this effort. This can be a strategic advantage. Previously, politicians set directions for the organization’s goals and activities, giving rise to budgets and targets. However, some political prioritizations can be considered poor investments in potential outcomes. One example is the idea of visible policing (D23 Visionsudvalget, 2005). The extent to which the police are visible in the streets could be the object of great political attention, but in solving crimes, police intelligence and criminological experience show that the results of such efforts represent poor value to how resources might otherwise have been deployed. As Hestehave emphasized: “Thus, it is often politically convenient rather than knowledge-based arguments that determine the goals and target points and thus the contents of police work, as well as how police resources and time should be spent” (Hestehave, 2013, p. 181). Thus, strategic analysis could reduce political interference in priorities and strengthen the police organization’s strategic focus.

4.2.2. The Established Practices

However, another view of the new strategy questioned the data quality upon which the analysis was built. The data were crucial because they were used for managerial steering and for the local police districts that produce annual strategic crime analysis, which together formed the basis for national strategic analysis. However, the police realized there was room for improvement, apparent in the annual police business strategies, such as those for 2016 (D9 Danish National Police, 2016a). Strategic Analysis 2015 was based partly on external sources such as statistics on Danish society published by Statistics Denmark and partly on the police’s own data and statistics systems POLSAS, HS2, and POLIS. The first two were case management systems, whereas the latter was a management information system that pulled data from the other two systems. The data included (in the Strategic Analysis chapter on crime) were the cases that received a so-called sharp record number. This means that the case could not be defined as merely an inspection or an event, but it was considered to possibly involve prosecution. From the policies’ reflections, it appears that this data registration created the following limitations:

One reason is that the data are collected for administrative purposes and therefore it is not collected to answer the problems, we may want to examine in the present analysis.

(D5 Danish National Police, 2015b, p. 6)

Furthermore, data registration was a limitation. Data records are often fraught with errors, which makes it difficult to search for data from police records (Memmert, 2016). This affects the quality of the strategic analyses. The analysts depended on the fact that the employees who wrote the review reports were careful about entering relevant information. The evaluation of the latest police reform raised concerns that some officers are sloppy with registration because it is difficult and perceived as a waste of time (Balvig et al., 2011). The data challenge was visible in the centralized steering of local police districts. In the “Goal and performance plan of Funen Police 2015”, for instance, there was a fairly detailed agreement on the prioritizing of activities that could support knowledge-based practices and, more specifically, the quality of data registration:

It is also crucial that the quality of the data used for both the business management and the operational control is accurate. It is here the police districts’ responsibility to ensure that the registered data in, for instance, the case system and the time-registration system is correct.

(D6 Danish National Police, 2015d, p. 3)

This meant that the data were not seen as providing an adequate picture of crime patterns because they were intended to measure and monitor how the organization meets performance management goals.

Ideally, the strategic analyses should include police knowledge ‘from the street’ (Sætre, 2006). However, this is not an obstacle. Police organizations are characterized by silent knowledge, which is transmitted only orally and cannot be included in the analyses unless shared with analysts (Memmert, 2016). For example, the analyst would not be able to seek knowledge that the intelligence department may have from observations if it were not written down and made searchable. The result would, therefore, be that the strategic analysis data did not provide a complete picture of the situation, which could lead to the analysis being rejected by, for example, the intelligence departments, as they had more adequate knowledge than the analysis could provide.

Thus, viewed from a corporate logic perspective, the new strategic analysis of the National Police aims to optimize organizational performance by improving the use of data and developing evidence-based interventions. This fits well with the police organization’s aspiration to set its direction with less political interference. However, the strategic analysis also nurtured another negative view, emphasizing that the data quality was insufficient and that the organization needed to allocate many resources to ensure data quality, which may conflict with the initial goal of increasing efficiency in crime fighting.

4.3. The Data-Driven Organization as Expression of a Profession Logic

The emphasis on a data-driven organization could also be understood as part of a change in police work from reactive to preventative methods, with new demands for professionals in the organization and challenging the interpretation of professional logic.

4.3.1. The New Practices

New positions for strategic analysts have been established and anchored in local police districts. One of the new local Intelligence and Analysis Units expressed that the local strategic analysis 2017 was like a first foundation to be built upon: “In the end, there should be a finished house. A house that symbolizes the Midt- and West Jutland Police District works more analytically and knowledge-based than before” (D16 Midt- and West Jutland Police, 2017, p. 5). Strategic analysts should have master’s degrees in political science, statistics, analysis, sociology, or criminology. The strategic analyses, then, not only entailed new work functions but also a new type of professionalism based on expert knowledge. The strategic analyst’s job was to collect data on specific crime areas, analyze the data, and suggest initiatives to prevent and develop methods. Where the focus in recruitment was earlier on police officers with skills to react and make quick decisions, the current focus is moving towards police officers with analytical and reflective skills (Hartmann, 2014). In the North Jutland Police’s job advertisement from 2013, it appeared that the task was to collect, develop, and combine multiple sources and data types:

In your analysis, you are capable of involving both police professional information and the environment, including socio-economic conditions, political conditions at local, regional and national level.

(D4 Danish National Police, 2013, p. 1)

They intended to create cross-professional collaboration. The strategic analyses, then, not only entailed new work functions but also a new type of professionalism based on expert knowledge.

4.3.2. The Established Practices

However, strengthening analytical competencies challenged police officers’ established perceptions of good police work. According to Memmert (2016), the strategic analysts in the police felt they should defend their existence in the organization and struggle to be accepted by the ‘traditional’ police workers.

Emphasis placed on the police’s quick response to a problem or threat counteracted the ability of the police organization to work analytically. In addition, in the job ad (mentioned above), the conflict between the types of knowledge that strategic analysts and the established police force possess was noticed. It said:

You’re good at setting the direction while you are also thinking of coherency and interested in working interdisciplinary, in an environment where you have the courage to stick to your professionalism and thereby create quality and value and include and use the knowledge which is already present in the Northern Jutland Police.

(D4 Danish National Police, 2013, p. 1)

Thus, there were traces of a general skepticism towards introducing academics into the police and the “academization” of policing (Hestehave, 2013). This skepticism may be rooted in the fact that most police officers think they are more operational and action-oriented (James, 2013). It feels more right for a police officer to catch criminals than to stop them: “This stress on quick thinking in emergence situations suggests that policing is a profession dedicated to action rather than to reflection” (Brodeur & Dupont, 2006, p. 12). Thus, patrol work was viewed as a core competence in line with the approach that “… non-coercive activities, talk, listening, community policing, partnerships, and other current fashionable programmatic approaches are disdained, sabotaged, avoided, and assigned to low-status officers”. (Manning, 2013, p. 63).

The difference between analytic and experience-based knowledge led to negotiations, for example, “whether statistical analysis or ‘street level information’ should be used to decide where operational resources should be allocated” (Johannessen, 2015, p. 174). Hartmann, who studied the Danish police, pointed out:

The industrial reorientation of the police into the elite branch of criminal investigators and the more recently invented cadre of ‘investigators of organized crime’ has generated internal tensions in the police …

(Hartmann, 2014, p. 60)

These tensions transcended all levels of the organizational hierarchy. Police employees were promoted after some years in the organization and came to stand as the heads of the hierarchy. The analytic strategists’ production of new insight into the organization’s way of working challenged the management’s traditional form of knowledge based on several years of experience in the service and the management’s formal monopoly to explain what is going on in the organization. It was perhaps not surprising that the new analytic units would report directly to the local chief supervisor, allowing for closer management of the analytical approach to policing.

In 2013, a new master’s policing course was established, and in 2014, a basic police training program was accredited. However, both programs were closed again. Basic police officer education was shortened from three to two years, and even shorter cadet-training programs were established. The aim is to increase the number of police officers involved. Therefore, along with the trend towards more analytical competencies, competing political ideas about the development of the police profession existed.

In sum, as demonstrated above from professional logic, implementing a data-driven strategy can be understood in two ways. On the one hand, it promotes the analytic competencies of police professionals through recruitment strategies and new educational programs to identify and prevent complex crimes. On the other hand, the police profession traditionally valued the ability to react quickly to crimes and threats, which differed from studying and preventing crime. These two viewpoints put pressure on the police’s professional identities. It was no longer a given knowledge and method needed to form the basis for police work: analytic, preventative, experience-based, and responsive.

5. Discussion

In contrast to renowned studies that have shown how a practice is the manifestation of a single logic (Lounsbury, 2007; Thornton & Ocasio, 1999), how a practice has changed because of a change in logic (Zilber, 2002), or how a practice is emerging in response to institutional complexity (Greenwood et al., 2011; Pache & Santos, 2010; Waldorff et al., 2013), this study’s findings revealed that a new strategy could evoke latent tensions representing a particular type of intra-institutional complexity that arise when organizations interpret conflicting demands within the same institutional logic. We offer two contributions to the literature on intra-institutional complexity.

First, based on our findings, we theorize three forms of intra-institutional complexities: the “mission dilemma” associated with the state logic, the challenge of “resource allocation” associated with the corporation logic, and the tensions around “identity pressure” associated with the profession logic.

Table 3. Three types of intra-institutional complexities.

Institutional Logic	Intra-Institutional Complexity Type	Tensions in Logic Interpretations	Case Example
State	Mission dilemma	When logic interpretations diverge regarding the organization’s purpose in society	Police’s mission is crime prevention versus crime response
Corporation	Resource allocation	When logic interpretations diverge regarding whether the use of resources facilitates or constrains organizational developments	Police’s use of data creates efficiency, whereas data quality is poor and inadequate
Profession	Identity pressure	When logic interpretations diverge regarding what profession and what expert knowledge should be given priority in organizational practices	Police officers’ identity builds on analytic knowledge versus street-level knowledge

provides an overview of the three types of intra-institutional complexity.

The mission dilemma refers to the complexity that may arise when there is ambiguity or a shift in the organization’s purpose in society. The state logic facilitates divergent perspectives on what the organization’s mission should be and multiple practices that are not aligned. The resource allocation complexity arises when the corporate logic nurtures doubt in the organization about whether the use of more resources will facilitate or constrain organizational developments. The identity pressure emerges when professional logics guide organizational actors in multiple directions regarding what types of professions and expert knowledge are valuable in the organization.

These intra-institutional complexities are important. Based on our case study, we argue that the different views of an organization’s mission (state logic), resource allocation (corporate logic), and identity (professional logic) indicate that there are more profound differences in how the logic elements are interpreted, including their core principles, legitimacy, authority, and strategy (see Table 1). For instance, when it comes to the state logic that now legitimizes crime response as well as crime prevention, it is no longer a given how to enhance the common good and safeguard social and economic stability. The basis of legitimacy is democratic participation, but what does this mean? Is it representative democracy where elected politicians set goals for the police, or is it participatory democracy where local stakeholders collaborate with the police? Furthermore, is bureaucracy the best authoritative way to organize police work, or could other forms, such as partnerships and networks, exist? These intra-institutional complexities are also related to corporate and professional logic.

Thus, we add to the discussion on the intra-institutional coherence of logics. Institutional logics build upon substances that are “unquestioned, constitutive interiors, the sacred core of each [institutional] field, unobservable, but socially real” (Friedland, 2009, p. 61) or an unobservable but essential value’ anchoring an institutional logic’ (Friedland, 2013, p. 34). Yet, although the institutional logics of the state, corporation, profession, etc., depict specific values, they each develop different organizational principles, identities, or practices depending on the specific context in which they reside. For instance, the profession logic may manifest differently, as there are different types of professionals with different views on values and practices (Berg Johansen & Waldorff, 2017). Market logic may manifest differently in countries with liberal or planned economies (R. E. Meyer & Höllerer, 2014). This discretion paves the way for variation, intra-institutional complexity, or near-decomposability, as Thornton et al. (2012) stressed, who explain how near-decomposability ensures agency and historical specificity without erasing what makes logic “institutional”. Despite his emphasis on core institutional content, Friedland recognizes “a certain modularity in the logics perspective, but simply warns the reader that the relation between the ‘institutional logics and its practical order may vary, they are not, however, arbitrary’ (Friedland, 2013, p. 38 in Madsen & Waldorff, 2019)”. Thus, while it may be impossible to define exactly how much variation in interpretations it takes to erase the “institutional” aspect of a logic, our study contributes to discussions about what are the core elements of institutional logics, and how these core elements vary and change. Our findings point towards more research and empirical cases of how intra-institutional tensions affect the core elements of logic.

Second, we advance the notion of institutional complexity, explaining how tensions arise not only due to conflicting demands in the macro external environment but also because of the intra-institutional complexities that the organization itself evokes at the meso level. We argue that institutional complexity may arise because organizational actors’ construction of different meanings related to a new concept evokes the presence and relevance of different institutional logic interpretations. This is in line with the conceptualization of organizations as contexts in which actors make sense of, interpret, and enact institutional prescriptions (Dopson et al., 2008; Binder, 2007; Zilber, 2002). The logic comes alive—so to speak—due to the organization’s deployment of them (Lindberg, 2014).

Our approach differs from most institutional complexity literature that assumes complexity is a result of incompatible field-level institutional logic. Greenwood et al. (2011) emphasized that the structure of organizational fields, whether emerging or mature, impacts the nature and extent of the institutional complexity facing organizations. This potential ambiguity may create space for organizational entrepreneurship. Or as R. E. Meyer and Höllerer (2016) explain, organizations may strategically produce ambiguity to neutralize institutional complexity. This literature furthermore stresses that institutional complexity does not equally affect all organizations. As shown by Pache and Santos, “We begin by identifying the contexts in which conflicting institutional demands are likely to arise and be imposed on organizations. We then explore the way in which conflicting institutional demands are experienced by organizations” (Pache & Santos, 2010, p. 456). Besharov and Smith (2014) developed a framework to understand how multiple logics manifest in organizations as contested, estranged, aligned, and dominant due to different logic compatibility and centrality combinations. Our findings support the idea that institutional logic is powerful in affecting actors’ views, but we also show that inside the organization, actors can interpret each logic in different or even opposite directions.

6. Conclusions

A data-driven strategy and its implementation are inevitably a process of dialogue, discussion, and negotiation. Various actors will seek to develop it in a direction that fits their values and norms. We found different, competing interpretations of a data-driven strategy perceived through the lens of three distinct institutional logics. Our analysis showed tensions in implementing this data-driven approach in the Danish National Police, owing to different logical interpretations. The novelty of our study is that we highlighted how institutional complexity may arise not primarily from a struggle between multiple logics but from the different interpretations of single logics that led to intra-institutional complexity.

Intra-institutional complexity was generated in state logic by two different interpretations: one concerning the data-driven strategy as strengthening safety and security by prioritizing preventative or responsive work; the other was critical because it may draw away attention and resources from more traditional police work, which was to respond quickly to acute threats and crime. Similarly, corporate logic facilitates a double interpretation that questions resource allocation. As a data-driven organization, the police may be able to work more systematically with knowledge and analysis to tackle societal challenges effectively. However, doubts arose as to whether there was enough capacity and good data to improve organizational performance and efficiency. Finally, professional logic created identity pressure because the data-driven strategy challenged police identity to shift from experience-based and responsive methods to analytical knowledge and preventative methods. Within the police organization, tensions arose regarding which forms of knowledge were legitimate and who should and could make decisions. Overall, the data-driven strategy challenged the perception of what it meant to be police.

This article offers practitioners an analytical lens by highlighting the multiple tensions that may arise in an organization when managers implement a new strategy. Our findings contribute to their understanding of the tensions deeply embedded in the organization’s logic, expressing different but equally legitimate understandings of the organization and organizational change. In many ways, the underlying institutional logics are powerful structures that connect meanings and practices in organizations, but they are also unstable entities and, therefore, facilitate different interpretations. Thus, examining intra-institutional complexities will help managers understand when and why change is difficult and help them navigate strategically.

The limitations of this study must also be acknowledged. First, the case study context may constrain the generalizability of our findings. Specifically, the institutional logics identified—state, corporation, and profession—are likely reflective of a traditional public sector organization such as the police and may not readily apply to other contexts. Second, we prioritized examining a wide range of documents from multiple sources, but the number of interviews conducted was limited, and the interview data may not adequately represent the full spectrum of organizational roles and perspectives. Third, this study focused exclusively on identifying institutional logics and their various interpretations, without exploring the tensions and dynamics between these logics in depth. To address these limitations, we propose three avenues for future research. First, the propositions developed in this study should be empirically tested and refined in different geographical and disciplinary contexts. Second, future studies should incorporate larger and more diverse interview samples that capture a broader range of relevant actors. Third, although our analysis spans approximately a decade, further research is needed to investigate how interpretations of institutional logics evolve over time and to examine the influence of key actors in shaping data-driven strategies.