Toward Collective Ambidexterity in Public Sector Digital Initiatives: A Case of the Finnish Water Sector

The digital transformation era calls for digital innovations affecting the public sector, but such organizations often focus more on the efficiency and exploitation of current resources, ascribing less attention to explorative innovation. Innovation in new capabilities and technologies often require inter-organizational collaboration. In this context, public sector organizations do not operate in isolation—often engaging in interactions within their service sectors. While the contemporary literature on organizational ambidexterity primarily focuses on the balancing opportunities of a focal organization, radical sector-level development requires collaboration among various actors to achieve collective ambidexterity. Drawing on the research question of the influence of inter-organizational collaboration on the outcomes of digital initiatives, this exploratory case study on the Finnish water sector found that simultaneously operating modes of inter-organizational collaboration can facilitate different aspects of collective ambidexterity in an inter-organizational field and, consequently, the expected benefits from initiatives. Our theoretical contribution lies in demonstrating collaboration modes that can enhance specific aspects of collective ambidexterity. Accordingly, we suggest that collective ambidexterity through deliberately adopted modes of inter-organizational collaboration requires field-level facilitation in parallel with organizational ambidexterity. Furthermore, consequent practical implications for managing such collaborations are suggested for policymakers, interest groups, and the public sector.


INTRODUCTION
Digital technologies and their related capabilities must change, often radically, over time to sustain their relevance and legitimacy [ 37 ].However, organizations tend to focus more on exploiting their existing capabilities and technologies instead of exploring new alternatives, which often implies uncertainty and potentially negative returns [ 39 ].To succeed under such circumstances, organizations must establish a balance between explorative innovation and the efficiency of their existing capabilities and technologies-that is, they must pursue organizational ambidexterity [ 36 , 47 , 49 ].By efficiency, we refer to actions that focus on incremental development and continuous improvement of existing products, services, and processes.Meanwhile, innovation refers to actions that aim for disruptive or radical changes through the acquisition and development of new knowledge (cf.References [ 2 , 37 , 72 ]).
Novel digital solutions-mobile technologies, data analytics, cloud computing, and the Internet of Things (IoT) -can provide revolutionary opportunities to organizations [ 60 ], but seizing their benefits requires innovativeness [ 57 ].While governments worldwide encourage public sector innovation, the demand for stable operations and avoidance of costs and risks largely favor the status quo [ 48 ].As reported recently, this barrier has been encountered by a variety of public sector organizations, for instance, public utilities [ 22 , 65 ], governmental agencies [ 35 , 38 ], regions, and municipalities [ 35 ].
In this context, inter-organizational collaboration may facilitate the development of digital public services [ 71 ].Such collaboration may materialize in varying modes, such as standardization, consortia agreements, or jointly owned service companies [ 27 ].Here, a mode of collaboration "involves a form of collaboration between organizations with an intention to accomplish certain benefits"-separating it from a simple form of collaboration that does not explicitly indicate the benefit objectives (Reference [ 16 ], p. 580).While collaborative development in various public sector digital initiatives is considered necessary (e.g., References [ 1 , 27 , 45 , 59 , 71 ]), urgent attention should be paid to the differing goals, visions, practices, requirements, competence, and know-how, as well as limited resources, to avoid failing initiatives [ 22 , 51 ].Drawing on this discussion, we propose the following research question: r How can the modes of inter-organizational collaboration influence the outcomes of digital initiatives in the public sector?
Our exploratory case study [ 58 , 73 ] specifically focused on six public water utilities in Finland, along with expert insights into the Finnish water sector on a general level.Through inductive analysis and subsequent theoretical positioning of the results, the relevance of ambidexterity in the field-level perspective of inter-organizational development of digital solutions emerged as an explanatory lens to structure our findings.While the literature mostly adopts the stance of a focal organization and concentrates on bilateral relations in the context of the collaborative development of digital solutions [ 6 , 13 , 29 ], the comparatively recent concept of collective ambidexterity focuses on the collective achievement of digitalization by extending the analytical focus beyond the focal organization to the level of organizational networks [ 26 , 50 ].Accordingly, our results suggest that the two complementary views of ambidexterity-organizational and collective (cf.Reference [ 26 ])-can help explain collaborative digital development both from a focal organization perspective and collectively at the field level.
The concept of collective ambidexterity [ 26 ] enabled further theorization of our results on the influence of modes of collaboration on collective ambidexterity in an organizational field.The analysis suggested field-level facilitation of ambidexterity along with organizational ambidexterity.In doing so, this article broadens the understanding of ambidexterity by explaining the ways in which collaboration modes can facilitate collective ambidexterity in developing digital solutions.This study also presents a few practical implications for understanding the effect of different collaboration modes on organizational and collective ambidexterity and the ways in which public sector organizations can achieve ambidextrous development (cf.Reference [ 8 ]).
The rest of this article proceeds as follows.The next section outlines the theoretical background and presents our theoretical proposition.Following this, Section 3 provides a brief overview of the case and explains our research methods and process.Section 4 covers the results, while Section 5 discusses our findings.Finally, a brief conclusion is presented in Section 6 .

BACKGROUND
Concepts like smart government and smart cities for public value require IT-enabled innovativeness as well as connecting the expertise of various domains [ 45 , 67 ].However, the public sector often emphasizes efficiency and risk avoidance, while also encountering organizational barriers to innovation [ 16 , 31 , 37 ].Nevertheless, external changes, such as digital transformations outside the public sector, are rapidly altering the environment, and the expectations of citizens with it [ 41 ].In this context, inter-organizational collaboration can provide a valuable avenue for digital transformation in the risk-averse public sector (cf.Reference [ 27 ]).Based on this motivation, the following sections introduce the theoretical background of inter-organizational collaboration, ambidexterity, and collective ambidexterity, as well as outline the analytical lens adopted in this study to investigate the influence of modes of collaboration on collective ambidexterity.

Inter-Organizational Collaboration
Sector transformation results from collaboration among institutions, actors, and technology [ 12 ].Since changes associated with digital transformation do not occur in isolation [ 18 ], organizational innovation can be embedded within a broader context through networks [ 31 ].Network collaborations-such as ecosystems, alliances, or communities-can yield innovation opportunities [ 68 , 70 ].Furthermore, inter-organizational collaboration fosters knowledge transfer beyond organizational boundaries and can produce better innovation results than intra-organizational efforts [ 4 , 9 ].Moreover, co-designed solutions may have a broader significance in this context, as participants often serve to extend the flow of knowledge [ 52 ].Additionally, organizations require specific capabilities to not only be innovative but also accurately identify the demand for innovations and related external knowledge [ 10 ].Thus, inter-organizational collaboration is necessary for adopting new technologies and realizing the benefits of digital initiatives [ 14 , 27 , 61 ].
Sector-level transformation poses several challenges and barriers to collaboration, including poorly defined objectives and the complexity of development projects and their requirements [ 3 , 22 , 44 , 51 , 62 , 65 ].However, along with providing direct opportunities for the actors involved in the development processes, collaboration can also influence the sector's strategic direction [ 52 ].Moreover, knowledge created in collaboration can equally benefit private sector organizations through the transfer of domain-specific knowledge from the government to its collaboration partners [ 53 ].Collaboration is also necessary for the development of holistic solutions [ 66 ].This implies that public-private collaboration can create value for a wider group of stakeholders, beyond the loci of individual organizations [ 54 ].
While public-private collaboration can enable innovation through interaction, a lack of ability or willingness to participate, along with ambiguous goals, can lead to dead-end projects [ 19 , 40 ].Furthermore, organizations need to first master technology-enabled changes to subsequently realize their long-term benefits [ 41 ].Innovation opportunities may also be hindered by restrictive regulations and the risk-averse cultures of governments [ 65 ].However, regulation may be considered a sub-optimal instrument for innovation in the municipal water sector [ 61 ].Thus, public-private innovation can be a sensitive process, requiring clarity about aspirations, expectations, and capacities to collaborate [ 19 ].

Ambidexterity
While organizational ambidexterity involves the incremental development of existing technologies and competencies, innovation requires simultaneous experimentation with new alternatives [ 39 ].In this case, the processes related to innovation and efficiency may compete for the same resources [ 49 ].However, the distance between the loci of learning and benefits realization is often greater in the case of innovation than efficiency [ 39 ].Moreover, risk aversion further restricts innovation activities [ 40 ].However, organizational success depends on the ambidextrous abilities of an organization to integrate both activities in the value-creation process [ 46 , 47 ].Although such organizational ambidexterity is considered necessary in the public sector [ 16 , 37 , 63 ], the process of balancing these practices is often heavily tilted in favor of efficiency [ 38 ].However, if organizations focus solely on incremental development, then the resulting inertia would prevent them from adapting to environmental changes [ 46 ], such as digital transformation [ 41 ].
While certain scholars emphasize sequential cycles of innovation and efficiency periods, others regard ambidexterity as the simultaneous ability to innovate and efficiently develop existing competence [ 56 ].This move away from "either-or thinking" toward "both-and thinking" characterizes the idea of contextual ambidexterity, which empowers employees to explore innovative opportunities in their daily tasks, as a complementary view to structural ambidexterity, which involves innovation activities functioning separately, such as in innovation hubs and departments [ 15 ].These two types of ambidexterity vary in their degree of structural separation and employee specialization in innovation or efficiency [ 49 ].Individuals who successfully establish contextual ambidexterity can align activities of innovation with mutual goals and promptly adapt activities to the changing environment [ 15 ].However, individuals in public sector organizations are often under pressure to ensure risk aversion and efficiency, thus indicating the need for political and upper-level support for innovation [ 16 ].
The surrounding environment greatly influences the approach embraced for innovation processes [ 49 ].Structural ambidexterity is beneficial when the perceived distance of opportunities from organizational capabilities is significant, whereas contextual ambidexterity is favorable when opportunities are numerous or uncertain [ 37 , 49 , 55 ].Highly ambidextrous public organizations show evidence of contextual ambidexterity, while moderately ambidextrous ones rely more on structural ambidexterity [ 16 ].At the same time, ambidexterity can be assumed to exist in organizations, in which case management's responsibility is to nurture it [ 37 ].Similarly, employees' personal interests may dictate which development projects are implemented and how [ 21 ].

Modes of Inter-Organizational Collaboration and Collective Ambidexterity
Knowledge creation through external collaborations is vital for developing organizational ambidexterity [ 6 , 7 , 25 , 29 , 30 ].For example, customers benefit from knowledge sharing with vendor organizations [ 25 ], and collective initiatives-such as smart cities-require knowledge of various domains, thereby necessitating knowledge contributions from various partners [ 6 ].In particular, organizations with limited resources that focus on efficiency can benefit from collaboratively supplementing innovation activities [ 6 , 34 ].
Although ambidexterity has been studied in various inter-organizational settings [ 6 , 13 , 24 , 29 , 33 , 64 , 69 ], many of these studies adopted the stance of a focal organization when examining partnerships as a means for developing ambidexterity.Kauppila [ 29 ] rationalized this choice by stating that it is the organizations, rather than the network of organizations, that create ambidexterity by establishing innovation and efficiency partnerships.However, since sector-level transformation necessitates collaboration among institutions, actors, and technology [ 12 ], moving beyond the locus of a focal organization can provide the requisite insight for understanding the influence of ambidexterity on the outcomes of digital initiatives.
Recently, Inoue [ 26 ] outlined that ambidexterity occurs at three levels-individual organization, group of organizations, and platform.He further identified the concept of collective ambidexterity as emerging at the group level, thus shifting the analytical focus from the focal organization to the ambidexterity emerging primarily in organizational networks.Similarly, Page et al. [ 50 ] studied collaborative ambidexterity and identified it requires collaborative advantage and inter-organizational structures.Therefore, this study adopts collective ambidexterity to investigate also collaborative ambidexterity and ultimately synthesizes the two concepts.Consequently, we propose that the modes of inter-organizational collaboration [ 27 ], which exist at the group level, provide the required structures and collaborative advantage [ 50 ] that enable collective ambidexterity at large.In a recent study [ 22 ], four modes of inter-organizational collaboration were identified with regard to digital initiatives in the Finnish water sector: • Autonomous mode , in which ICT acquisitions and development are conducted in-house or through bilateral cooperation with a vendor.• Limited company mode , in which a limited company supports the ICT development of several water utilities.
• Central IT service mode , in which administrative ICT services are acquired from in-house or central municipal IT services.• Standardization mode , in which sector-level standardization of, for example, data models, data transfer, application processing interfaces (APIs), and architectures is pursued to achieve collective benefits.
Among these, the autonomous mode is most commonly present in the contemporary Finnish water sector.Although standardization can be considered an efficiency strategy [ 20 , 39 ] to increase negotiation power and innovations, the collaboration modes-standardization and limited company modes-have been suggested to be parallelly employed to generate benefits beyond those anticipated purely from the autonomous mode [ 22 ].Similarly, Stettner and Lavie [ 64 ] studied the ambidexterity of internal organizations, alliances, and acquisitions to find that if exploration and exploitation are balanced merely within a mode of operation, then the resulting ambidexterity may remain weak.Furthermore, they highlighted that maintaining this balance across different modes of operations facilitates ambidexterity.While Stettner and Lavie [ 64 ] focused on the organizational level, we proposed to pay attention to the field level.Accordingly, we decided to identify how the previously recognized collaboration modes-the autonomous mode, standardization, limited company, and central IT servicescan foster collective ambidexterity in the water sector (cf.Reference [ 26 ]).Hence, we considered the following general-level theoretical proposition as an analytical lens for our research: We propose that different constellations of inter-organizational modes of collaboration (MoC ) [ 22 , 27 ] may influence the variation of focus (efficiency vs. innovation) and, in turn, the balance of collective ambidexterity [ 26 , 50 ] in an organizational field.Thus, MoC stands for multiple modes rather than a mode.The next section describes the methods used to this end and further explicates our analysis of the organizational field of the Finnish water sector.

RESEARCH METHOD AND PROCESS
Innovation can occur at both the organizational and field levels [ 23 ].Notably, the analysis unit of an organizational field comprises a collection of organizations that constitute a recognized area of institutional life [ 11 ].This embedded case study design [ 73 ] involved both these levels of analysis-organizational and field.
The Finnish water sector comprises approximately 1,100 water and wastewater utilities responsible for managing a water network of 107,000 km and a wastewater network of 50,000 km [ 42 ].Although most of these utilities are small, the 20 largest ones provide services to 80% of customers [ 32 ].Even though some of these utilities have active digital initiatives, the majority are only beginning to understand the potential of digital solutions [ 22 , 62 ].While water utilities generally acknowledge the benefits of digital transformation, various barriers, such as legacy systems, vendor lock-in, and limited resources and expertise for novel digital solutions, hinder sectorlevel transformation [ 22 ].However, some utilities have been able to overcome these barriers.Nonetheless, the gap between the most advanced and less mature utilities, as well as between state-of-the-art digital opportunities and the sector's capabilities at large, is considerable.These contradictions offer an interesting empirical basis for studying emerging digital initiatives and the related challenges in the public sector context.Hence, this exploratory case study [ 58 , 73 ] was conducted as part of the WWData project that aims to improve wastewater network data management and digital initiatives in the Finnish water sector.
Six municipal water utilities were analyzed to gain insights into organizational and field-level issues.While four of these utilities-regionally scattered-were part of the project, two more utilities were invited to participate  to gain a better understanding of the opportunities and barriers of digital initiatives for small-and mediumsized utilities.Since the focus of this study is on novel digital solutions, the smallest utilities-such as private associations and local cooperatives-were excluded.Furthermore, 14 vendors provided additional insights into field-level issues, such as opportunities and challenges related to technological development and collaboration.These vendors, who offer or aim to offer products and services to the water sector, varied from startups to mature companies, including product, software, and consulting companies, as well as a network operator.Moreover, although relevant ministries and municipalities are closely related to the water sector, utilities have autonomy regarding digital initiatives [ 22 ].Therefore, ministries and municipalities were excluded from this study.The data collection protocol for the water utilities and vendor organizations was designed in terms of the conceptual lenses of enterprise architecture (business, applications, data, and technology perspectives) and, more broadly, collaborative digital initiatives.Many public organizations in Finland use enterprise architecture to manage their structure of operations, processes, data, information systems, the services they provide [ 28 ], and even organizational transformation [ 74 ].Thus, these concepts of data collection provided the means to capture various viewpoints on the current state and future aspirations, as well as existing and envisioned collaborations, pertaining to digital initiatives in the water sector.
The data collection process included conducting qualitative semi-structured interviews with representatives of the water utilities and vendor organizations, which were carried out in 2020 (Table 1 ).The interviews were recorded (apart from one failed recording) and transcribed.In addition, secondary data (such as research reports; publications of ministries, municipalities, and public interest groups; legislation; financial statements of the utilities; project reports; newspaper articles; and web pages) were acquired for additional information on the water sector and utility operations.Most of these documents, which are publicly available, were used to triangulate the Table 2. Phases of Thematic Analysis [ 5 ] Phase Actions 1. Familiarizing with the data Generating transcriptions of the data; reading and re-reading it.

Generating initial codes
Formulating initial codes describing the characteristics of the collaboration modes (autonomy, standardization, limited company, and central IT services)-for example, current state, future goals, benefits, challenges, required changes, necessary changes, consequences if changes are not implemented, collaboration, inter-organizational aspects.

Generating initial themes
Identifying potential themes-for example, resources and expertise, technological development and complexity, innovation, efficiency.

Reviewing themes
Reviewing the identified themes at the code level, analysis unit level (organizational level), and entire data set level (field level); formulating the thematic map (Figure 1 ). 5. Defining and naming themes Defining and naming the themes (innovation and efficiency) based on the collaboration modes (autonomy, standardization, limited company, and central IT services).

Reporting
Reporting.
data and ensure an in-depth understanding of the phenomenon [ 73 ].Furthermore, secondary data were used to corroborate the informants' statements, for example, about innovation activities and projects in their respective utilities.
The interviews with the water utilities, structured upon enterprise architecture, captured the current state, processes, and organizational aspects of their digital initiatives, including the existing and potential future of inter-organizational collaboration.In general, the interviews aimed at capturing the current state as well as the aspirations of and benefits sought from novel digital solutions involving collaboration among water utilities in the surrounding ecosystem.Meanwhile, the vendor interviews focused on mapping challenges and opportunities as well as obtaining perspectives for the direction of future development within the ecosystem from the industry point of view.The semi-structured format of the interviews enabled discussions on emerging topics during the data collection process [ 43 ].Moreover, while the data collection phase did not assume the theoretical lens of ambidexterity, the interview protocol allowed the informants to describe their current practices and envisioned development plans, as well as the related collaboration, which ensured the construction of a narrative that was not influenced by the theory.
The utility informants included key stakeholders, such as managing directors, network managers, development managers, and engineers responsible for operating, developing, and managing wastewater networks in their respective municipal water utilities.This means that they possessed firsthand knowledge about digital initiatives in water utilities and the sector at large.The duration of the interviews ranged from one to two hours.Two to four interviews were conducted for each water utility to draw various viewpoints on the phenomenon and enhance interviewee reflections [ 17 ].Meanwhile, the vendor participants included managing directors, business managers, and project managers.The key interviewees received the interview themes in advance.They snowballed the experts from their respective organizations to join the interviews, which ensured expertise and knowledge on the topic and aimed at mitigating elite bias [ 43 ].Hence, the informants possessed firsthand knowledge about digital initiatives, as well as the opportunities and challenges of offering such products and services to water utilities.Each vendor organization participated in a two-hour interview, 10 of which were group interviews and 4 were personal interviews.
The data analysis process began with structuring the data according to the previously identified modes of inter-organizational collaboration-autonomy, standardization, limited company, and central IT services [ 22 ].A six-step thematic analysis [ 5 ] (Table 2 ), chosen due to its flexibility and iterative nature [ 5 ], was performed for the organizational-and field-level issues.Iteration cycles were conducted within the six steps, as well as between the steps.The computer-based analysis tool NVivo was used for the data analysis.Our inductive analysis [ 73 ] aimed to identify the barriers and aspirations that instructed the development activities (or the lack thereof).We identified the relevance of the variation in foci (innovation vs. efficiency) regarding desired outcomes.As a result, the theoretical lens of ambidexterity was generated while iteratively analyzing the data.As the analysis progressed, we sought to understand the ways in which the different collaboration modes contributed to innovation and efficiency in water utilities, in particular, and the field, in general.The data concerning the collaboration modes and ambidexterity were triangulated, iterated, and evaluated.The thematic map [ 5 ] in Figure 1 illustrates the themes and relationships identified during the analysis, showing how the analysis of the water utility cases and findings from the vendors contributed to the field-level perceptions.
The data analysis revealed indications of the need for ambidexterity in a broader context beyond focal organizations.In this respect, the research process of this study is consistent with that of a recent study by Page et al. [ 50 ] that identified collective ambidexterity as a concept that is complementary to organizational ambidexterity, even though this relationship was not assumed in the interview protocol.Subsequently, we formulated the analytical lens (MoC → Collective Ambidexterity) for this study and performed final analysis iterations.In this way, our inductive analysis guided us toward novel findings from our data.Furthermore, it helped us discover the importance of a field-level perspective and collective ambidexterity in the context of digital initiatives and the radical transformation of a sector.

RESULTS
Our data revealed that while efficiency is considered the underlying goal of public water services, with efficiencyfocused digital development taking place in all six water utilities, innovativeness is increasingly being considered necessary to realizing benefits, ranging from advanced analytics to ensuring high-quality services in the future.However, various barriers that hindered innovation and held up, if not prevented, digital transformation at large were observed.
This chapter first outlines the results of the data collected from the six water utilities and then presents the field-level results.Table 3 summarizes the general-level goals of efficiency against innovation, as assessed from the digital initiatives undertaken in terms of the collaboration modes.The chapter concludes with a summary of how the identified collaboration modes influence the establishment of ambidexterity in the water sector.Notably, our results and analysis insist on looking beyond organizational boundaries to consider the field at large.

Organizational Ambidexterity in Water Utilities
Water Utility 1 preferred autonomy in its digital initiatives on wastewater solutions.According to the head of the network department, only the operative departments of the utility possessed sufficient expertise to acquire, develop, and manage wastewater-related digital solutions.Furthermore, although central in-house IT services  were available, they were only used for general solutions and services and, therefore, would be rendered inefficient if each department acquired the same general-level digital expertise.Notably, in-house expertise in the operational department enabled the agile digital development of the utility.According to the interviewees and various project documentation, innovation projects, hackathons, and experimentations were conducted to create new knowledge.Furthermore, the management encouraged self-guidance and experimentation with novel solutions along with daily work, as noted in the quotation below: I would put it so that [. . .] digital transformation is just a tool, and the goal is to get those processes as short as possible, and people could analyze the operations themselves, like self-directing to make the operations leaner.I do not quite believe that there should be a great command from above, and then everyone starts to skirmish happily to carry out that command, but perhaps it would be that people would have the opportunity to see for themselves what went wrong, what worked well, and then develop that well-working line forward.(head of the network department) Experimentations were accepted and encouraged to ensure sustainable development, which could only be maintained by establishing a permissive and supportive atmosphere.Therefore, a clear vision, adequate resources, and new knowledge acquired through multiple channels enabled systematic digital development and innovation in this utility.
Water Utility 2 exclusively used the autonomous mode in its digital initiatives related to wastewater operations.According to the interviewee, the varying requirements of different utilities complicate projects if they are conducted jointly.The utility's interest in developing a joint data platform was expressed, although not initiated at the time of the interviews.The utility acquired new knowledge by hiring individuals with prior experience in digital initiatives, experimenting with new solutions, and participating in an innovation cluster with access to ongoing innovation projects and partner networks.Furthermore, expertise in various new technologies was acquired and, based on the interviews, the vision for their utilization was clear, as apparent from the following quote: The platform is the be-all and end-all, a place where all this information can be obtained and mathematical models can be created.It also enables computing in the background [. . .] to get the entire wastewater system under control.[. . .] We need a platform where we can use computing and bring machine learning.One place [where data are] accessible and available.[. . .] One way is to munch [analyze] the existing data.Then again, to get them all in one place would have tremendous possibilities.(water and wastewater network manager) Enhanced expertise not only encouraged learning but also contributed to more systematic and efficient digital initiatives and innovation activities, such as ongoing experimentation to understand the benefits of new IoT technologies.The interviewee further emphasized that enhanced data assets with advanced analytics would enable better customer service in the current turbulent environment and enable new business opportunities.Moreover, expertise in digital acquisitions in the water sector would improve system interoperability and the development of novel solutions (insufficient competence in acquisition serves to complicate, or even prevent, system interoperability).
Water Utility 3 acquired its domain-specific digital solutions autonomously, while solutions provided by central municipal IT services were used whenever possible.For example, the utility acquires cloud services for analytics and data visualization from central IT services.According to the interviewees, coherent field-level standardized practices increase the comparability of different water utilities, and general-level recommendations regarding network parameters could help different utilities in better asset management.Furthermore, uniform solutions offered by a software vendor (cf. the limited company mode) enabled this utility to participate in joint development with other utilities based on common requirements.Development projects, corroborated by annual reports, produced new knowledge on novel solutions.According to the interviewees, vendors propose solutions to specific challenges, while utilities require comprehensive and interoperable solutions.The interviewees envisioned new and increasingly open business markets enabled by digital transformation.Thus, they deemed development in the water sector necessary for attracting new actors to the field.In this regard, one of the interviewees expressed the following: If we do nothing, it is already now challenging for us to get any players [vendors] as we have so poorly data available.[. . .] And then there is this certain pressure on us to enable new business markets.Whether we want it or not, there is constantly this card on the table that we should promote transparency and thus enable business from it.(network manager) Water Utility 4 also acquired its digital solutions autonomously, with limited resources restricting its feasible opportunities.This utility was a partner in a wastewater treatment plant (W W TP) limited company that provided them with an opportunity to implement a joint super visor y control and data acquisition (SCADA) system.However, changing the water domain systems was found to be challenging.In this regard, an interviewee expressed the following: When the decision is made, then you are married to it.(managing director) Consequently, the utility continued using the existing SCADA.The utilization of standardized data, such as energy consumption data, was regarded as beneficial.Furthermore, although new knowledge on novel solutions was acquired, the solutions often became too costly for a small utility, providing little, if any, return on investment.For this specific utility, novel digital solutions could provide multiple benefits, such as improved data exploitation, but limited resources restricted its development.As a result, the focus was primarily on incremental development coupled with tacit knowledge in combination with the data received from the network.The interviewees predicted that external pressure and rising costs would likely drive change in this utility.
Water Utility 5 regarded joint acquisition projects with other utilities as essentially complex.Hence, it preferred the autonomous mode of digital acquisition.This utility was also a partner in the W W TP limited company, which enabled the implementation of a joint SCADA.However, since changing the essential system was found to be too complex for this utility, the idea was disregarded, as noted in the quotation below: So, basically, all the way from there, when we started to do that automation (in the 1990s), the challenge has been that it is much easier for us to follow the development path of the vendor and proceed that way.[. . .] It is kind of so that [the vendor] who has been able to cop the deal has almost a bit of a monopoly position.So it would be a pretty crazy thing to do if we were to think about changing our entire data collection.Incremental development In addition, central IT services for the W W TP partners were anticipated to be problematic due to varying proprietary systems.The interviewees confirmed that identifying new opportunities offered by the novel solutions and applications appeared challenging to them.Under such circumstances, benchmarking could lead to the creation of new knowledge on novel solutions.Meanwhile, tacit knowledge was used in the aggregation of existing operative systems.Digital development projects often require vendor collaboration, and it took some time for this utility to persuade its current vendors to develop legacy systems.This was further complicated by a lack of interoperability, which the interviewees mentioned should have been addressed better.Nevertheless, legacy system development is necessary to implement novel solutions.With vendors offering specific applications without having a comprehensive view, it was considered challenging to ascertain the specific requirements from novel solutions and the options that could benefit from them in the long run.
We have made attempts to think about how the data from pumps and electricity consumption and all could help us anticipate something or future costs or anticipate defects.There is a need for this kind of thing, but this is just that, when we live everyday life, and if we have been doing this [working in the water sector] for longer, I have noticed that we tend to take a narrow perspective.People wonder what it [insights from the data] could be.So we know it would be great to know more, but we don't know what we should know so that we would know.(managing director) Water Utility 6 also preferred the autonomous mode of implementing digital initiatives.The current vendors of this utility were reluctant to develop systems, or the costs seemed too high, as noted by the interviewees: Our SCADA vendor [anonymized] has been reluctant to develop interfaces to other systems, for example, to the Network Information System.These are the interface problems.(planning director) The product [SCADA] has remained the same for a long time, and the supplier [anonymized] does not seem to have its development group for this that would continuously develop this product.(operating engineer) Furthermore, this utility considered collaboration with other utilities as a means of gaining bargaining power, although doing so would still be challenging.Other functions (district heating and energy) of the utility further affected water-related digital investments.Benchmarking was considered to create knowledge related to different novel solutions, and the efficient use of data was envisioned to produce added value and enable data-driven decision-making.In addition, it was observed that better interoperability and data quality would improve data management, while improved data analysis could provide new information for planning and management.However, at the time of the interviews, no such implementations had been made to enable these envisioned benefits, although the utility was expecting a quote from a vendor for such a system.Meanwhile, tacit knowledge, along with existing systems and data analysis (e.g., using Excel), was employed.According to the interviewees, challenges related to wastewater network conditions and maintenance and the need to allocate investments more efficiently represented the main drivers of change.Table 4 summarizes the perceived aspects of organization-level ambidexterity in the six utilities.

Field-level Issues of Collaboration and Ambidexterity
Based on evidence from the vendor and utility interviews, the autonomous mode for digital initiatives appeared as the most prominent mode prevalent in the Finnish water sector.This indicates a focus on efficient local acquisitions and development, which enabled the achievement of organization-specific requirements and customizations.However, customer-specific customizations tended to complicate the systems and increase the relative cost of autonomous development, as noted below: Digital leap in the water sector is seen as software acquisitions.[. . .] The interface architecture must be crystal clear in order to exploit all of them [software] efficiently.Otherwise, connecting the systems may be too expensive or even impossible.[. . .] A new kind of expertise is needed at the water utilities to ensure interoperability.(water utility informant) Legacy systems decelerate such progress by demanding vendors to consider existing systems.These issues, together with inadequate resources and expertise, limit autonomous development opportunities and create barriers to radical changes at the field level.While interest in digital initiatives was observed, many utility and vendor representatives considered the perceived costs hindering progress.Irrespective of the identified challenges, many vendor representatives mentioned observing recent changes in the field, reporting that a few utilities were actively experimenting with and implementing novel solutions, while some, although marginal, were involved in systematic research and development activities.The general view of the vendor representatives was that changes were taking place, albeit cautiously.
Notably, in connection to rare acquisitions, the autonomous mode poses challenges and can lead to vendor lock-ins-where a utility depends on the products and services of one vendor.This sentiment was also conveyed by one of the interviewees as follows: There is the problem with the fact that [. . .] public procurement process is quite tough, and then the water utility does not have such IT expertise, especially a longer-term vision.[. . .] The industry does not move forward if the same guys do the same job for 20 years, and the technology would have developed considerably further, but the systems cannot be replaced smartly.(vendor informant) Therefore, although various solutions to ensure efficient data usage and digital transformation exist, autonomous development hinders the realization of benefits and interoperability in the long term.Furthermore, more emphasis on expertise in digital solution acquisitions could have been beneficial for this sector.Interviewees from two water utilities emphasized the importance of utilities to understand the larger picture, while several vendors emphasized that change must emerge fundamentally from the utilities.In contrast, interviewees from two other utilities expected holistic solutions from vendors rather than the narrow applications that they offered.In this context, limited access to relevant data and lack of visibility of the overall circumstances of the utilities further complicate the situation, as reported by the vendor representatives: If a design is ordered for a wastewater network [. . .], it may be that the customer sends a stamp-sized piece of the network map as the source information.It [the design] probably can be done based on it, but the customer may not realize that if you provide the whole material, it could tell the designer something more, i.e., "Aha, this should be done in a completely different way than it was ordered."(vendor informant) Consequently, our evidence proposes that while new knowledge about digital initiatives is necessary for water utilities, vendor organizations also require equally comprehensive knowledge of the situations prevalent in the utilities to ensure effective collaboration.may not have them.So it kind of feels like you don't necessarily want to demand them if the answer is "we don't have them."Then, it's gonna be a bit awkward.(vendor informant) This indicates that mutual knowledge sharing could enable a more holistic approach to digital initiatives in the field, thereby suggesting the need for trust-based ecosystem collaboration.
Considering the challenges of the prevailing autonomous mode, individual organizations could participate in informal collaboration, knowledge sharing, and joint innovation activities within the sector.We found that the utilities considered in this study participated in knowledge sharing in various formats, such as through joint innovation projects with private sector companies, universities, and the Finnish Water Utilities Association.Meanwhile, the vendors envisioned that joint innovation projects could yield valuable benefits, such as new innovations, new use cases, new customers, and, in general, a better understanding of the overall situation and goals of the water utilities.In turn, the utilities anticipated benefits like improved understanding of the implementation process of new solutions, new knowledge and knowledge sharing on digital initiatives, and improvements to leakage management.Furthermore, while benchmarking is already used in the water sector and in comparing the district heating and energy sectors, more systematic benchmarking was anticipated to be helpful.A solution suggested by one of the vendor interviewees was coordinated benchmarking.Regardless of the informal collaboration, there was a clear consensus-instead of individual utilities developing capabilities for their legacy systems, more benefits would be received from standardized solutions that fit many utilities.
The interviews highlighted the urgent requirement for standardization in the water sector.Although various standards and protocols exist, no coherent field-level standard has been able to gain sufficient ground, as noted by a vendor, Sure, there are data transfer protocols and interfaces, but they seem to be unfamiliar in the water sector.Around five years ago, I contacted SCADA vendors, and they did not know about the standards I had asked about.In the energy sector, those were commonly used.(vendor informant) The incoherent adoption of standards prevents vendors from efficiently offering field-level solutions and restricts utilities from obtaining benefits from digital transformation.Furthermore, overall interoperability was considered unsatisfactory-the existing interfaces were insufficient or non-existent, while legacy systems prevented data sharing.Thus, it is evident that standardization is crucial for better interoperability, as expressed in the quotations below: For this type of solution that we do, data analytics, the challenge is to get good quality data.[. . .] They are pretty much in silos, the data, and they are behind those other vendors.There is always a great effort to make a new data transfer or interface, which is not very smart for water utilities or new companies either, because then water utilities cannot get better solutions.(vendor informant) For a service provider, data availability is essential.(vendor informant) While most vendors wished for standardization, only a few interviewees supported its application through regulation.Coherent field-level standards would enable more vendors to enter the field, new innovative solutions , economies of scale, and overall interoperability.In general, standardization would ensure better fitting and compatible solutions to secure operations between utilities, as well as enable new innovations in the field.
When the requirements of multiple utilities are largely similar, the limited company mode can offer effective solutions and economies of scale to a group of partners or customers.However, for municipally organized utilities, holding the shares of limited companies could become complicated due to legal constraints.Under such circumstances, commercially produced standardized solutions could ensure secure operations and outsource services among utilities or a third party, as expressed by an interviewee: New system made it possible that utilities can work together and stand in for each other.Therefore, although the limited company mode enables the efficient development of digital solutions, incorporating various requirements from different utilities may complicate this development, as stated by one vendor: The idea was to make a small and simple system, but when experts came into it and billing clerks, there was a barrel of ideas and opinions which would be nice to get.And it started to get bigger.So, it's not small and simple but very wide.But configurable, and in the end, that is good.(vendor informant) Central municipal IT services ensure efficiency and support for solutions that are not specific to water service operations.In other words, they provide general-level expertise to utilities.Development projects may require the collective involvement of the central IT service provider along with the utility in question and the vendor(s).However, while this mode of collaboration can offer generic expertise in digital solutions to utilities, specific knowledge in water-domain-related digital solutions would be essential at the operational level.Overall, the consensus was that this expertise needed to be improved.In summary, it was concluded that although central IT services could likely provide efficiency to digital initiatives, they fail to capture the entire expertise required for the novel digital solution requirements of systems related to the water domain.
Taken together, the viewpoints of most vendor and utility representatives identified the need for a paradigm shift and new knowledge at the field level.This is because while interest in novel solutions and data-driven decision-making has grown, tacit knowledge has often been preferred over adopting, for example, advanced analytics solutions, as highlighted by a vendor interviewee: It may be that it [advanced analytics] is much more laborious to implement there [in small utilities] than to recall how the network was built in the '80s.(vendor informant) In addition, current operations and development activities are primarily focused on efficiency and the incremental development of existing capabilities and technologies.For example, advanced data analytics was considered mainly as a management control tool rather than a tool for significantly influencing innovative water services or work processes, as indicated in the following comments: It would be more of a management and planning tool.It analyzes and compiles data based on which it is easier to create an overall picture of the situation.(water utility informant) The changes would bring only one new system from which work is received.So there would not be much change in practice.(water utility informant) Despite recognized objectives in favor of innovative digital initiatives, limited time and resources in public sector organizations restrict their development opportunities, which essentially indicates the predominant focus on efficiency in this sector.Hence, it is evident that the goals for digital development are primarily directed by the efficiency of operations.
Furthermore, data sharing and interoperability were considered necessary for field-level efficiency and experimentation with innovative solutions.One of the vendor representatives reflected that a customer-focused mindset would require a new focus from management.This observed change is slow, fundamentally requiring initiative from utilities.In this context, new acquisition expertise is also considered necessary.Furthermore, another critical aspect of innovative development is empowering individuals, as noted by an interviewee.
More encouragement to experimentation and acceptance of different experiments is needed.Increased understanding that individuals must be interested in developing, it cannot come from the top management, but the one who wants to develop their own practices needs to be interested.(water utility informant) While representatives of the water utilities generally pointed out the need for novel solutions, such as those enabled by IoT and analytics, realizing benefits from these digital initiatives could take a long time, which may,  to a certain extent, hamper the willingness to invest.However, experimentation and interest in IoT solutions have increased recently, and vendor representatives expect this trend to continue.

Summary of the Results
This study was initiated to investigate the influence of inter-organizational collaboration modes on the outcomes of digital initiatives in the public sector.The results suggest that, while a paradigm shift and new knowledge are required in the water sector, water utilities in Finland mainly focus on the pursuit of efficiency, with a narrow focus or limited opportunities to innovate.While some of the utilities were able to innovate and maintain their balancing practices to achieve organizational ambidexterity through their preference for autonomous development, the remaining utilities merely pursued efficiency gains from digital initiatives.Furthermore, the collaboration modes-such as standardization, limited company, and central IT services-provided (or were envisioned to provide) only efficiency opportunities for the utilities.At the field level, it was observed that while the autonomous mode could provide innovation and efficiency gains to water utilities, the lack of standardization hindered vendors from offering field-level solutions and innovating.Considered together, the preference for autonomy (providing narrower utility-level benefits) and the lack of standardization pose barriers to deriving the benefits of digital initiatives.Accordingly, the need for efforts to introduce alternative modes of collaboration to increase innovation capacity in the water sector emerges.Figure 2 summarizes our observations on the dominant modes of digital development and collaboration, as well as those pertaining to the organizational-level versus field-level ambidexterity.
At the organizational level, we found that (1) the autonomous mode enables innovation and, consequently, ambidexterity for certain water utilities but not for all.Regarding the modes of collaboration, we discovered that (2) while the autonomous mode was largely preferred and primarily used for digital initiatives, the lack of standardization limited innovation opportunities for the organizational field.Hence, the overall finding pertaining to the field level proposes that (3) the lack of ambidextrous balancing by one set of actors within a field can cause an imbalance in another set of actors and, consequently, in the entire field at large, thereby hindering the process of implementing sector-level radical changes and transformation.In summary, new knowledge is needed in both water utilities and vendor organizations and can be achieved through new constellations of collaboration modes.

DISCUSSION
This section discusses the theoretical and practical implications of the findings of this study.

Theoretical Implications
Our insights suggest that neither inter-organizational ambidexterity nor the focal organization perspective provides a comprehensive view of the ambidexterity required to enhance digital initiatives and bring about radical changes at the field level.As noted, for example, by Bresciani et al. [ 6 ] and Tiwana [ 69 ], inter-organizational and alliance ambidexterity are often viewed as dyadic relationships, which would benefit from the expansion of the network perspective.As a result, we adopted the concept of collective ambidexterity [ 26 ] and argued in favor of the following theoretical implications.
(1) Our study presents (hitherto relatively scarce) empirical insights into sector-level governmental services and explains how the prevailing autonomous mode of digital development falls short of enhancing the outcomes of the digital initiatives of a sector as a whole.
(2) As a unit of analysis, the organizational field advances the theoretical understanding of the modes of collaboration.Accordingly, an ensemble of modes of collaboration (cf.Reference [ 27 ]) may need to be adopted to gain sector-level benefits beyond those anticipated from a single mode of collaboration or by a single organization.(3) Consequently, we proposed a theoretically novel analytical lens (MoC → Collective Ambidexterity), which suggests that varying constellations of modes of inter-organizational collaboration (MoC) would influence how collective ambidexterity [ 26 ] emerges from the public sector digital initiatives.(4) This elaboration extended the theoretical foundations of collective ambidexterity [ 26 ] concerning the three-level view of collective ambidexterity in a particular field (Figure 3 ).
In the context of digital initiatives in the public sector, public organizations must inevitably collaborate with various public and private actors to acquire and develop new knowledge.As observed in this study, this knowledge also needs to be extended beyond organizational boundaries to spread across an entire field.While the autonomous mode prevailed in our case study, limited innovation opportunities caused an imbalance in the field, which hampered the sector-level outcomes of digital initiatives.While a few utilities could achieve organizational ambidexterity autonomously, attaining collective ambidexterity would require the combined success of various actors [ 26 ].Such field-level imbalance, in turn, hinders the outcomes of digital initiatives in the field as a whole.
Adopting a field-level perspective made it possible to extend the understanding of how a combination of inter-organizational modes of collaboration (cf.Reference [ 27 ]) can contribute to the collective ambidexterity of an organizational field.Consistent with our findings, the literature identifies standardization as an efficiency strategy [ 20 , 39 ].However, as observed in our analysis, the current lack of standardization hinders vendors from offering field-level solutions and even prevents innovation.Notably, our field-level analysis revealed that the standardization mode, in combination with other collaboration modes, may be considered a means of innovation.This viewpoint was previously inaccessible to studies conducted from the perspective of a focal organization.Thus, instead of vendors providing customized solutions to individual utilities, combining standardization with other collaboration modes could allow innovative development to benefit a broader customer base.By identifying the advantages of multiple collaboration modes, our study complements Stettner and Lavie [ 64 ], who addressed the matter from a focal organization perspective.While an organization can obtain various advantages from a single collaboration mode, a combination of these modes can produce collective ambidexterity beyond that which can be anticipated by a single organization.Hence, understanding collective ambidexterity [ 26 ] in a sector requires parallel foci on the organization and the organizational field as a whole.
The preference for the autonomous mode-which nonetheless provided benefits to individual water utilities [ 22 ]-coupled with a lack of standardization, limited innovation opportunities in the water sector.While governmental organizations are faced with an increasing need to enhance their innovation capabilities, their attempts may be counteracted by certain organization-level governance mechanisms, such as emphasizing sound investments and bureaucracy [ 38 ].Since the autonomous mode (pursuing organizational ambidexterity) alone does not lead to the development of collective ambidexterity in a sector, with a combination of collaboration modes being envisioned to produce benefits beyond those anticipated purely from one mode, we can establish that a constellation of collaboration modes is necessary for sector-level ambidextrous development (i.e., MoC → collective ambidexterity).Notably, this proposition complements the conceptualization of collective ambidexterity [ 26 ].
Consequently, we proposed an analytical framework comprising three levels to better understand collective ambidexterity (Figure 3 ).In Figure 3 , the left side indicates the overall three-level view, while the right side illustrates the central elements of collective ambidexterity connected to a focal organization and the field at large.Overall, the figure illustrates how the MoC is realized at the group level, thereby providing the necessary structures and collaborative advantage [ 50 ] for attaining collective ambidexterity, which, in turn, impacts and is impacted by organizational and field-level mechanisms (cf.Reference [ 26 ]) and balancing practices (cf.Reference [ 38 ]).While a focal organization perspective provides an overview of balancing across MoC related to the focal organization, the field-level view provides a perspective across the MoC of multiple organizations.Thus, our study provides insights into appropriately balancing across MoC throughout the entire field.Overall, this analytical framework illustrates the requisite parallel foci of organizations and organizational fields to achieve collective ambidexterity.
Additionally, our framework complements the concepts presented by Inoue [ 26 ] and Page et al. [ 50 ] and extends the current understanding of collective ambidexterity to the organizational field.According to DiMaggio and Powell [ 11 ], the organizational field is not merely an interactive network but covers the totality of relevant actors.From this perspective, an ecosystem may be considered a type of organizational field [ 68 ].Furthermore, this study describes and theorizes how MoC can generate collective ambidexterity in a specific field.As discussed, the autonomous mode can produce innovation and efficiency benefits at the organizational level.Similarly, standardization can be considered a part of both an efficiency strategy [ 20 , 39 ] and an innovative approach.In addition, limited company and central IT services can render efficiency gains.In other words, innovation and efficiency are not required in equal measures to achieve ambidexterity [ 7 ]-a few collaboration modes can remain focused on efficiency to achieve collective ambidexterity, while others must also aim for innovation.These findings largely align with Stettner and Lavie [ 64 ]: Ambidexterity is enhanced by balancing across rather than within the MoC.
In summary, while ambidexterity in alliances [ 6 , 24 , 33 , 64 ] and other inter-organizational settings [ 13 , 25 , 29 ] has been studied extensively, this study offers a theoretical understanding of collective ambidexterity.Moreover, this study expands the analytical spectrum beyond organizational boundaries by providing a novel field-level perspective on ambidexterity (Figure 3 ).The subsequent findings advance the understanding of the achievement of ambidexterity in a field and its influence on the outcomes of digital initiatives.

Practical Implications
The practical implications of our study provide insights into the innovation and organizational learning necessary for realizing benefits from digital initiatives and therefore are valuable for organizations, policymakers, and interest groups.The following are the three main practical implications of our results: (1) The autonomous mode enables innovation and, consequently, ambidexterity in a few water utilities but not all.(2) While the autonomous mode was largely preferred and primarily used for the digital initiatives of water utilities, more standardization is required to allow opportunities for innovation in the organizational field.(3) A lack of ambidextrous balancing in one set of actors within a field can cause an imbalance in another set of actors and, ultimately, in the field at large, thereby hindering the process of the digital transformation of a sector.
The autonomous mode was observed to enable efficiency as well as innovation in certain water utilities, whereas others exhibited evidence of merely undertaking incremental development.Moreover, the most ambidextrous utilities demonstrated both contextual and structural ambidexterity (Table 4 ).A clear vision of future digital development was deemed vital, along with continuous development, innovation, and experimentation.Furthermore, new ideas and knowledge on various digital solutions were produced by offering a permissive and supportive atmosphere, allowing experimentation, and recruiting experts with relevant skills.These findings largely support the established view of contextual ambidexterity [ 15 , 49 ]-it requires individuals to be capable of aligning themselves with adaptability.However, the challenge with organizations constructed on efficiency goals lies in initiating contextual ambidexterity when the surrounding environment calls for radical changes [ 49 ].
New knowledge and a paradigm shift were deemed necessary changes that require immediate action in the water sector to enable beneficial outcomes from digital initiatives.However, the efficiency-focused utilities (Table 4 ) considered these opportunities irrelevant or unattainable either because of their limited resources or their limited level of operations.Consistent with the literature [ 22 , 31 , 37 ], it has been observed that public organizations that focus on incremental development can obtain only limited, if any, innovation gains using solely the autonomous mode.As a result, tacit knowledge is used instead of learning new digital solutions.Enforcing contextual ambidexterity in such an environment may endanger new initiatives [ 49 ] if personal interests, or the lack thereof, direct the development process [ 21 ].It may even compromise efficiency activities if the organization ends up being mediocre in both strategies [ 46 ].Thus, structural ambidexterity is essential to avoiding cultural clashes and developing new capabilities [ 49 ].
Structural ambidexterity-in the form of innovation projects, schemes, and innovation clusters-was observed in the utilities capable of receiving innovation gains from the autonomous mode (Table 4 ).Our analysis noted that structural ambidexterity allowed these utilities to gain the new knowledge needed to initiate transformation and create an organizational environment where contextual ambidexterity could flourish.Furthermore, a high level of centralization [ 8 ] and a structural approach [ 49 ] can equally benefit vendor organizations [ 33 , 53 ].Thus, it may be assumed that to initiate a paradigm shift, structural ambidexterity is required to gain external knowledge and initiate top-down cultural change in an organization.
The second practical implication of this study is that while the autonomous mode was found to be the most preferred mode for digital initiatives, the lack of standardization limited innovation opportunities in the organizational field.Data availability, process automation, and standardization play significant roles in transforming the water sector [ 66 ].Our data suggest that establishing common standards for digital solutions and data parameters could improve the comparability of utilities.Such field-level coherence would enable more vendors to enter the field, new innovative solutions, economies of scale, and improved interoperability.Moreover, along with increasing efficiency [ 20 , 39 ], standardization also contributes to innovation opportunities.Hence, we argue that if water utilities continue to depend solely on the autonomous mode without considering standardization as a means to establish rules and routines in the sector, innovation opportunities in the field will continue to encounter barriers.
Innovation activities in the water sector are interconnected with a broader social and technical context [ 31 ], making new knowledge and innovation vital for water utilities and equally necessary for vendor organizations [ 53 ].However, the observed field-level imbalances persuaded us to focus beyond the focal organization to obtain our third crucial empirical finding-the lack of ambidextrous balancing among one set of actors in a field can cause an imbalance for another set of actors and, in turn, in the field at large, thereby hindering the process of digital transformation of a sector.Hence, when radical changes are required in an organizational field, a focal organization's ambidexterity might be insufficient.In such a situation, field-level collective ambidexterity is necessary.Consistent with Inoue [ 26 ], the field-level view explored in this study provides grounds to comprehend how collective ambidexterity results from the combined success of various actors.We believe that this field-level view of collective ambidexterity has important practical implications for policymakers and the sector at large in terms of attaining a better understanding of ways to collectively improve the outcomes of digital initiatives.
Overall, while experimentation and innovation have increased in recent years, the resulting progress has been cautious.New technologies continue to fuel innovation [ 4 ] but, as observed, limited resources can make such opportunities irrelevant or unattainable.In this context, because individual water utilities may not be able to capture the full potential of innovation independently, they would benefit from collaboration among various actors within the field [ 22 , 31 ].Furthermore, such an innovation process among organizations, consisting of both inward and outward knowledge flows [ 9 ], can benefit both utilities and vendor organizations and, consequently, the field at large.Hence, it can be assumed that collaboration in the innovation process can provide leverage at the field level, thereby creating a productive environment for collective ambidexterity and sustainable development.In this context, the parallel use of additional modes of collaboration along with the autonomous mode can provide the requisite support to enable collective ambidexterity.These insights improve the understanding of the role of innovation in the public sector, which is vital for realizing the outcomes of digital initiatives.In summary, this study highlights how public sector organizations can attain ambidexterity-a concept that has so far garnered limited attention in the literature [ 8 ].

Limitations and Avenues for Future Research
Despite our attempts to thoroughly comprehend the phenomena under study, our research has certain limitations.First, we presented a case from a single sector in a single country, which may limit the generalizability of the study results to different institutional settings in different countries.Although the case study design (which contains heterogeneous embedded units of analysis and data collection from various companies from different sectors) took steps to mitigate this risk, future studies should seek to verify our findings in the context of different settings.Second, since the digital transformation of the water sector in Finland is still at an early stage of public sector digitalization and the standardization mode has not yet been extensively utilized, the interviewees' views likely represented merely envisioned ideas rather than actual outcomes.However, this provides a promising avenue for future research .We recommend conducting further studies focusing on the evaluation of how the standardization mode facilitates innovation in the public sector and contributes to collective ambidexterity in a particular organizational field.We also urge academics to study the modes of collaboration for collective ambidexterity in sectors that have already achieved an advanced level of digital transformation.Furthermore, it is essential to better understand the governance of collective ambidexterity, particularly the influential mechanisms (cf.Reference [ 26 ]) and balancing practices (cf.Reference [ 38 ]) that serve to enhance its effects.

CONCLUSIONS
Digital technologies and their related capabilities require ambidextrous development to sustain their long-term relevance and legitimacy [ 37 ].While public sector organizations are compelled to maintain risk aversion and efficiency [ 16 ], mere incremental development does not provide adequate support when the surrounding environment changes radically [ 46 ].In this context, inter-organizational collaboration could ensure better innovation results and long-term sustainability [ 9 ].Drawing on this context, the current exploratory case study provided novel insights into collective ambidexterity, as conceptualized by Inoue [ 26 ], and identified its relevance in securing expected outcomes from digital initiatives in the public sector.Additionally, this study extended the theoretical understanding of collective ambidexterity by theorizing how the MoC (providing the required structures and collaborative advantage [ 50 ]) can describe the formation of collective ambidexterity in an organizational field.Our three-level analytical framework described collective ambidexterity from both the focal organization and the organizational field perspectives.The observed field-level imbalance hindered benefit realization from digital initiatives at large, indicating the need for a parallel focus on the organization and the organizational field as a whole to achieve collective ambidexterity.In addition to our theoretical contribution, our findings have significant relevance to organizations, policymakers, interest groups, and the public sector in general in improving their understanding of the role of innovation and the ways in which ambidexterity can support the outcomes of digital initiatives.Therefore, our study contributes to the limited literature [ 8 ] on clarifying how public sector organizations can achieve ambidexterity.
Data are often missing or of poor quality [. . .].We have recognized that our employees might not dare to press for them.[. . .] The customer may not realize that the data are needed, and some customers Digital Government: Research and Practice, Vol. 4, No. 4, Article 20.Publication date: December 2023.

Table 1 .
Data Sources

Table 3 .
Summary of Observed Objectives of the Collaboration Modes

Table 4 .
Summary of Perceived Organizational Ambidexterity