Assessing the fragmentation of hybrids' performance management systems

Abstract

Purpose

Fragmentation can inhibit joint goals and performance measures. Therefore, the purpose of this study is to investigate the level of fragmentation between public, private and third sectors in a hybrid organization's performance management systems and the effects of this fragmentation to propose hypotheses and new research methods for future studies.

Design/methodology/approach

The inductive research design was based on a mixed method approach. As empirical data, a survey, interviews and documents were used in this case study examining a hybrid organization called Welfare Alliance.

Findings

The results showed low-level fragmentation in the performance management system of the hybrid. Although the level of fragmentation was low-level, it affected the hybrid's ability to implement joint performance goals and measures. Performance management practices suffered as a consequence.

Originality/value

As a theoretical contribution to research addressing performance management in hybrids, the study proposes new concepts and theoretical hypotheses concerning fragmented performance management systems in hybrids. These theoretical hypotheses propose how performance goals and measures can become fragmented because they isolate service production units and activities from each other. The proposed hypotheses for future studies also attempt to provide explanations for how fragmentation can spread from one management function to another (i.e. from goal setting to performance measurement).

Citation

Rajala, T., Laihonen, H. and Kokko, P. (2021), "Assessing the fragmentation of hybrids' performance management systems", International Journal of Public Sector Management, Vol. 34 No. 3, pp. 312-335. https://doi.org/10.1108/IJPSM-02-2020-0061

Introduction

Hybrid organizations mix managerial features, value systems and institutional logics of markets, state and civil society while delivering public services (Savignon et al., 2018). Although hybrids aim to utilize the best of both private and public worlds, they often meet challenges in creating joint performance management (PM) systems between companies, voluntary and public sector (e.g. Agostino and Arnaboldi, 2015, 2018). Difficulties in establishing joint PM systems in hybrid organizations are understandable, as hybrids are combining different PM traditions of public, private and/or third sector organizations (Johanson and Vakkuri, 2017). In combining the PM traditions, disputes between the three sectors arise from different ownership modes, institutional logics, funding bases and forms of social and institutional control (Billis, 2010; Savignon et al., 2018). PM, which is understood here as target setting and performance measurement (Moynihan, 2005), can become fragmented if these disputes cannot be solved and decoupling of targets and measures of the three sectors becomes modus operandi (e.g. Pache and Santos, 2013).

Unfortunately, fragmentation in PM can be problematic in different types of hybrids (Kurunmäki and Miller, 2006; Hodges, 2012; Alexius and Grossi, 2018), be they state-owned enterprises (SOEs), public-private partnerships (PPP) or social enterprises (SE). PM systems are more or less fragmented in hybrids when they do not connect company staff to government officials and voluntary sector workers in one or more functions of performance management (Rajala et al., 2020). The typical example of fragmentation in hybrids is the situation where the public and private sectors have their own separate measurement systems (e.g. Kurunmäki and Miller, 2006), and commercial confidentiality inhibits the sharing of the necessary performance information between the company and governmental actors (Coghill and Woodward, 2005; Flinders, 2005).

Most often hybrids become fragmented because actors are, or they choose to be, prisoners of the organizational boundaries of public and private organizations (Kurunmäki and Miller, 2006; Rajala et al., 2020). This prisoner mentality limits collaboration in performance management. In general, organizations may cease to cooperate in the design and implementation of hybrid's performance measures if laws, traditions, procedures, norms and habits are against such efforts (Rajala et al., 2020), or if accountability becomes too complex issue (e.g. Bryson, 2006). The previous literature has also recognized many other difficulties in interorganizational collaborations (e.g. Provan et al., 2007). These difficulties include:

1. Failure to identify a common target for the network (Parker, 2007).

2. Absence of an operational management system (Elmuti and Kathawala, 2001).

3. Lack of commitment to joint matters (Zineldin and Bredenlöw, 2003).

4. Failure to fulfill the objectives and needs of the partners (Zineldin and Bredenlöw, 2003).

5. Organizational conflicts (Agranoff and McGuire, 2001).

Fragmentation in hybrids is often unwanted, as it tends to cause problems (e.g. Agostino and Arnaboldi, 2018), such as uneven distribution of benefits between partners (Camarinha-Matos and Abreu, 2007), mission fragmentation (General Accounting Office, 1997), increased costs and undesirable outcomes (Lowery, 1998). In the private sector, fragmentation has decreased competitiveness (Chang and Singh, 2000) and operational efficiency (Weber, 1974).

Despite the increasing academic attention to hybrids and their governance (Grossi et al., 2017; Skelcher and Smith, 2015; Noordegraaf, 2007), there is a shortage of studies that examine how to evaluate the level of fragmentation between sectors in hybrid organizations' PM systems (e.g. Franco-Santos et al., 2012). This study utilizes interviews, documents and survey questionnaires to provide an answer to this research question. By adopting a case-study approach and mixed-methods analysis, this research suggests an evaluation model for examining the level of fragmentation in hybrids' PM systems. As fragmentation can be a barrier to progress in the adoption of joint PM, the ability to evaluate fragmentation is a valuable practical skill in the design and management of hybrids. For academics, the study provides novel hypotheses to be tested in the future.

The remainder of the paper is organized as follows. Section two of this paper reviews the literature to define key concepts and the main theoretical framework. Section three describes our research design and empirical data. Section four summarizes and analyzes the results of our case study. Section five concludes the discussion by reflecting on the results, while last section provides the conclusions.

Hybrid organizations and fragmented performance management

In general, hybrid organizations join different established organizational categories to create new organizational forms (Battilana and Dorado, 2010, p. 1419; Scott, 2008; Kastberg and Lagström, 2019). The focus of this study is in special form of hybrids, namely PPPs. Here, PPPs cover equally government-business interface as well as that of government and civil society. According to the past literature, eight attributes differentiate this type of hybrid organizations from other types of organizations (Billis, 2010; Johansson and Vakkuri, 2017):

1. Mixed ownership between taxpayers, business owners and members of voluntary sector (e.g. Thomasson, 2009).

2. Stakeholders of different sectors (Rajala, 2020).

3. Combinations of governance modes seen in markets, civic society and government (Vakkuri et al., 2021).

4. Goal incongruence between public value, shareholder value and social value (e.g. Kreps and Monin, 2011).

5. Competing institutional logics, that is, profit-seeking logic versus the logic of effective market interventions (Alexius and Cisneros Örnberg, 2015; Skelcher and Smith, 2015).

6. Multiplicity of funding arrangements containing tax money, stocks, bonds, dues donations and legacies (e.g. Hodge and Greve, 2007).

7. Mixed human resources combining public servants with company employees and voluntary workers (Agostino and Arnaboldi, 2017).

8. Blended operational priorities including collective choice, individual choice and mission-oriented choice (von der Heydte, 2020).

9. Parallel usage of economic and social control seen in public, private and third sector (e.g. Power, 2000).

What comes to PM in hybrids, it can be argued that there is no well-established and accepted performance management framework for hybridized government. In contrast, the private sector has the performance prism, the model of the European Foundation for Quality Management, balanced scorecard and so on (Vakkuri et al., 2021). The public sector utilizes the three E's (economy, efficiency and effectiveness) model and its applications widely (Pollitt and Bouckaert, 2011). Although important discussions addressing several theoretical perspectives of PM in hybrids are emerging (see, e.g. Liu et al., 2014), the understanding of PM in hybrids is still in a nascent state (e.g. Battilana and Lee, 2014).

The past research has suggested that the link between organizational strategies and performance measures is ambiguous, and the PM systems are fragmented (Kurunmäki and Miller, 2006; Johanson and Vakkuri, 2017). In fragmented systems, ambiguous performance targets reduce performance information use because it becomes unclear how measured results relate to hybrid's goals (Grossi et al., 2017). In this way, measures get separated from goals and PM becomes more fragmented. Complexity, power and multiple values of the different actors arising from different institutional logics have been often used to explain the fragmentation and ambiguity in PM systems (Johanson and Vakkuri, 2017; Rajala et al., 2020; Campanale et al., 2020), but other reasons have also been emphasized (see Provan et al., 2007).

To understand better the fragmentation in hybrids' PM, a thought exercise concerning the implications of the attributes of hybrid is required. By applying the attributes pointed out by Billis (2010) and Johanson and Vakkuri (2017), Table 1 proposes a set of implications for PM arising from the attributes of hybrids. These implications can lead to conflicts in both goal setting and performance measurement if combining public, private and third sector features fails in the design process of the PM (e.g. Rajala et al., 2020). Failures in combining, on the other hand, lead to fragmentation.

One way to tame the conflicts in PM systems of hybrid organizations is to use decoupling (Alexius and Grossi, 2018). Decoupling can increase or decrease fragmentation depending on the context. For example, decoupling strategic goals from operations may increase fragmentation between strategic and operational levels, but decrease fragmentation of PM at the operational level because it is not necessary to address goal incongruence in service production. Another way to prevent value disputes and fragmentation in PM is to create shared goals by applying the concept of boundary object (Vakkuri et al., 2021). Shared goals motivating day-to-day actions are examples of cohesion, and cohesion is the opposite of fragmentation. A boundary object is an entity that links communities together and enables them to collaborate on a common task (Wenger, 1998). In fact, boundary objects are weakly structured in common understanding, but this structure is adopted in more than one community, making boundary objects recognizable across community borders and enabling collaboration. However, boundary objects are strongly structured and rigorously defined within communities (Star and Griesemer, 1989), and the meaning of boundary objects differs across communities (Star, 2010). A performance goal and performance measure can be a boundary object linking together (Rajala, 2020):

1. All public, private and third sector actors in the hybrid.

2. Subset of actors in the hybrid.

Performance goal as a boundary object for all public, private and third sector actors is a shared goal that is ambiguous by nature. This means that every public, private and third sector actor is attempting to contribute to the goal achievement, but the goal has different meaning to public sector actors compared to private and third sector actors (Vakkuri et al., 2021). For example, the joint goal between public, private and third sector can be enhanced customer satisfaction among elderly people. If the public sector produces housing services, the third sector arranges game nights, and the private sector provides dental healthcare for the senior citizens; the joint goal of customer satisfaction means different things to the three sectors. For the public sector, customer satisfaction is satisfaction toward housing services, whereas the third sector attempts to produce satisfying game nights. The private sector contributes to the goal of customer satisfaction by ensuring that elderly people are satisfied with the dental healthcare services they receive (Rajala, 2020).

Similarly, the measure of customer satisfaction can be boundary object creating common identity between public, private and third sector (Vakkuri et al., 2021). Thus, the measure can be accepted by all three sectors, but public sector measures satisfaction toward housing services, third sector records satisfaction toward game nights and private sector tracks satisfaction toward dental healthcare services. By aggregating the measures used by public, private and third sector, it is possible to measure the customer satisfaction related to the hybrid organization taking care of the elderly people (e.g. Vakkuri et al., 2021). This aggregated measure of customer satisfaction explains how the hybrid is able to respond to the needs of the customers. This aggregate knowledge can be useful to all actors in the hybrid. Because performance measures have different purposes in hybrids depending on who uses them and in what situations (Agostino and Arnaboldi, 2017), turning key performance indicators into boundary objects can become useful.

When performance goal is boundary object to subset of actors, the goal is not shared by all, but it is shared by some actors (Rajala, 2020). Consider a hybrid organization that is a welfare center providing services for two groups of customers: elderly people and young children. This welfare center attempts to achieve two goals: improved customer satisfaction among elderly people and improved children's physical health. The improved customer satisfaction among elderly people is joint goal for public sector providing housing services for elderly people and for the private sector providing food for the elderly people. For the public sector, the customer satisfaction is satisfaction toward housing services, whereas the private sector attempts to produce satisfying food. However, the third sector organization arranges only sports events for children in the same facilities used in the senior citizens' rehabilitation. Therefore, improved customer satisfaction among elderly people is not a performance goal for the third sector organization. It is only a goal for the public and private sectors. In this example, the public and private sectors form a subset of actors to whom the improved customer satisfaction among elderly people is a performance goal in the form of boundary object. The performance measure tracking the improved customer satisfaction among elderly people is also a boundary object to subset of actors (i.e. for the public and private sector).

From a theoretical perspective, it is possible that performance goals and measures are less ambiguous among all actors in hybrid or among subset of actors in the hybrid. However, less ambiguous goals and measures usually relate to collaboration where the different sectors produce the same services. As the case organization examined in this study does not fall into this category, we shall not go through any theoretical examples of shared goals and measures that are unambiguous. Therefore, the last theoretical concepts proposed here are the following two concepts: exclusive performance goal and exclusive performance measure. In the theoretical example relating to the welfare center that had two performance goals, the third sector had an exclusive goal relating to the youth's physical health and measure tracking this goal was an exclusive measure. The goal and its measure were exclusive, as they excluded private and public sectors from the performance management practices related to the goal of improved children's physical health. Exclusive goals and measures are a sign of fragmentation (e.g. Hodges, 2012).

Research design and empirical data

This study can be categorized as a case study (e.g. Yin, 2009). The study focuses on fragmentation between different service production units in a hybrid known as welfare alliance. The research design builds on the notion that both quantitative and qualitative studies have benefits and drawbacks (e.g. Johnson and Onwuegbuzie, 2004). Although the qualitative approach is perceived as a more suitable method in some occasions and the quantitative approach is seen to work better in others, using the two approaches together is seen as the most useful approach for the study. Because the mixed-method approaches use qualitative and quantitative analysis techniques, the theoretical results of this type of study are more likely to survive in the practical tests of work-life, where practitioners test theories with qualitative and quantitative approaches.

As a mixed-method study, this article can be categorized as having a partially mixed sequential dominant status design (e.g. Leech and Onwuegbuzie, 2009). The study is partially mixed, as first part of it applies only qualitative methods and the second part applies quantitative methods (e.g. Onwuegbuzie and DaRos-Voseles, 2001). Because the chosen qualitative and quantitative methods are used in a sequential rather than congruent manner, the study is sequential (Senne and Rikard, 2002). Finally, dominant status refers to the fact that most of this study applies quantitative methods, which makes the quantitative approach more dominant (e.g. Collins, 2000). To summarize, because the work is a mixed-method approach, it:

1. Utilizes both theoretical and empirical approaches.

2. Applies both qualitative and quantitative research designs.

3. Employs inductive reasoning in the analysis and abductive and deductive reasoning in the theoretical section.

4. Combines different analysis and data collection methods.

Case: THA

Tesoman hyvinvointiallianssi (THA) translates to Tesoma's Welfare Alliance, and it is the examined hybrid organization in this study. The welfare alliance is located in one of the suburbs of the city of Tampere. This alliance was supposed to provide a wide variety of public services to the 20,000 inhabitants of the suburbs, and its budget was 56,914,399 euros in the first budget term. The length of the budget term was four years, and it covered the years 2018, 2019, 2020 and 2021. THA aims to integrate social services with other welfare services (e.g. cultural and leisure services) and place them all under the same roof. Integrating the services under THA was seen as the best way to promote city residents' health and well-being. Multi-professional collaboration between different welfare services was emphasized, and the attempt was meant to generate a new service production model called the alliance model, wherein the public sector, companies, voluntary organizations and citizens could produce services together. Customer perspective and efficient, effective service paths were at the core of the THA (See Appendix 2, Document 1).

The alliance model was designed to solve problems observed in the purchaser-provider model previously used by the local government. The term “alliance model” referred to the mode of operation, which was based on a joint contract between the public and private sectors and between public and third sector actors participating in the welfare alliance. In the alliance model, all parties involved in the hybrid are responsible for planning and implementation. The parties under the joint contracts form the alliance organization that conducts the actual planning and implementation of the welfare services. The public, private and third sector parties also share benefits associated with operations related to the production of the welfare services (See Appendix 2, Document 2). Risks related to the alliance are shared between public and private sectors (See Appendix 2, Document 6).

The previous purchaser-provider model was based on separate contracts between purchasers (i.e., committees appointed by the municipal council) and various producers, whereas the alliance model executes a joint contract between the purchasers and private sector producer from all services produced in the welfare center. Previously separate contracts were made for each service, but now all the different services (i.e. healthcare services, oral care services and housing services) were in the same contract. Moreover, the alliance model attempts to create joint goals between the purchasers and providers instead of relying on goals set by the purchasers. Blame games and the transfer of risks to the purchasers or providers were also problems associated with the purchaser-provider model. To fix this, the alliance model promotes the sharing of risks and benefits while placing greater emphasis on rewards instead of sanctions. Trade secrets also challenged the use of the purchaser-provider model; therefore, the alliance model was based on open-book principles. Finally, the alliance model switched from reporting to continuous improvement. Overall, the aim of the alliance model is to progress from sub-optimal operation to the optimization of the entire service system comprising public, private and third sector actors (See Appendix 2, Document 3).

The service units operating in the welfare alliance include public, private and third sector units, which points to the existence of mixed ownership, a feature typical of hybrids. Specifically, the units are a library (local government unit), a basic healthcare unit (private sector unit), an oral healthcare unit (joint public and private sector unit), a maternity and children's health clinic (local government unit), a domiciliary care and housing service unit (local government and private sector units), a youth center (local government unit), a walk-in clinic (local government and third sector unit) and a community café (third sector unit). This welfare alliance operates in the healthcare, social care and education policy fields. The tasks it conducts are assigned to local government (i.e. municipality) by laws set in central government. The Regional State Administrative Agencies oversee the quality of the services as a central government actor.

In order to understand the research context in which the welfare center operates, some basic knowledge concerning Finnish local governments is required. According to the Finnish Constitution (1999/731: 121§), “Finland is divided into municipalities whose administration must be based on the self-government of the inhabitants.” Self-government means that municipalities have the right to decide on their own affairs within the boundaries of the national laws. However, the national laws set the mandatory tasks for the municipalities.

The municipal council has the ultimate authority in the municipality, and thus it plays a significant role in exercising self-government. The council is responsible for the activities and finances of the municipality. The council, among other things, approves the municipality's budget. The council also makes decisions on all matters assigned to it by the national laws and regulations. Lastly, appointing members to other municipal bodies, such as the municipal board and committees, is the duty of the council.

The studied welfare center operates under the command of the municipality, although the private actors have been taken into goal-setting sessions in the alliance approach. The administration of the THA is based on two joint organs called as steering group and management group. The steering group manages the operational activities and gathers together every two weeks (Interviewee D). In the steering group, financial matters are reported monthly (Document 4, p. 32) and nonfinancial results are discussed. It has representatives from health center, oral healthcare, maternity and child health clinic, domiciliary care and housing service unit, library youth services and info desk and third sector (Interviewee A).

The management group is the main decision-making body in the hybrid organization. The duties of the management group include the following (Document 1):

1. Resolve disagreements and exercise the ultimate authority in the hybrid organization.

2. Accept changes in goals and provide strategic guidance to the hybrid organization.

3. Define the authority for the service unit managers and specify procedures.

4. Appoint a manager for the hybrid organization and accept proposals from the service unit managers who participate in the steering group.

5. Communicate with the purchaser and stakeholders.

6. Review and approve any changes to the service plan and management system described in the contract used to establish the hybrid organization.

7. Monitor the policies and performance of the hybrid organization.

All in all, the case organization suits well to the study design because it attempted to create a coherent PM system by including public and private sector actors to design the process of the PM. The goal-setting in the alliance model deviated from goal-setting in contracting out, outsourcing and public-private partnerships because it included creating joint risks and benefits between public and private actors. Due to the joint risks and benefits, a unified PM system was needed, although the hybrid included various actors producing services in different policy fields, and there were multiple values present indicating that fragmentation in PM system would be present. The third sector organization was recruited later, and the recruiting process ensured that the goals of the third sector were compatible with the goals of the hybrid.

On the one hand, the case offered examples of cohesive PM solutions arising from the attempts to generate a unified PM system. On the other hand, it was also possible to see more fragmented areas of PM because the variety of actors in the hybrid made it difficult to create a completely cohesive PM. Therefore, the case offered a chance to test how well the methods developed in this study could detect different fragmentation levels in the case organization's PM system. Studying other hybrids that use joint risk and benefits models to unite actors producing services in different policy fields would provide an opportunity to compare the results of this study.

The research method

In the data collection stage, researchers interviewed employees from three different levels of the hybrid. We carried out nine semi-structured interviews (e.g. Blee and Taylor, 2002) during autumn 2018. The interviewees were the heads of service provision responsible for purchasing services from the hybrid (n = 2), project managers responsible for the design and implementation of the hybrid (n = 2) and management group members who ran day-to-day management (n = 5). The interviewees were interviewed only once. Each interview lasted about one hour, was conducted face-to-face, audiotaped and transcribed verbatim. Either one, two or three researchers participated in the interview, and after the interviews finished, researchers who were not present in the interviews read the transcripts to identify any biases identified by Brown (2001). No biases arising from the interviewers were observed. The interview questions are in Appendix 1.

In addition to interviews, we conducted an online survey (e.g. Blee and Taylor, 2002) asking which of the 11 performance goals and 12 measures of the hybrid were considered to be relevant in the day-to-day management of the service unit, according to the manager leading the service unit. The managers could choose between four response categories: not relevant at all, relevant to some extent, extremely relevant or I do not know how to answer this question. The survey was sent to seven managers to cover all the eight service units in the welfare center. All managers send their replies. Item nonresponse (e.g. Bethlehem et al., 2011) was not a problem, as every manager responded to each question asked. However, only manager from adult oral healthcare participated. Therefore, children's oral healthcare is not covered. Lastly, we collected documents from the hybrid organization to comprehend the current status of the hybrid. The 44 documents were collected in 2018–2020.

In the analysis stage, the performance goals and measures were analyzed and coded by using Atlas.ti software (see Friese, 2019). The coding categories for qualitative content analysis and coding examples are reported in Table 2 and column 2. In the inductive content analysis (e.g. Elo and Kyngäs, 2008), the survey answers and interviews were used to determine the relevance of the performance goals or measures to the different service units. Relevant to some extent and extremely relevant answer options chosen by the survey respondents meant in our interpretation that the service unit of the respondent considered the performance goal or measure as a relevant one in day-to-day management. The not relevant answer option in the survey meant that the goal or measure was not considered as relevant in service unit in which the respondent worked. The coding was conducted separately to goals and measures in a different file in Atlas.ti software. The interviews were used to triangulate the findings of the survey and they confirmed the same findings.

The evaluation framework used to assess the level of fragmentation was developed inductively because the authors could not find an existing evaluation framework developed for analyzing the experienced fragmentation level of PM systems in hybrids. The quantitative analysis utilized descriptive statistics and co-occurrence tables. There were eight types of descriptive statistics used in this study, and c-coefficient measures the strength of a relationship between two codes (Virues et al., 2019). The c-coefficient is similar to the correlation coefficient, and the range of c-coefficient is between 0 (two code categories presented in column 1 in Table 2 never co-occur) and 1 (two code categories presented in column 1 in Table 2 co-occur whenever they are used) (e.g. Friese, 2019). All the quantitative measures used in the study are presented in Table 2 and column four.

The logic behind the descriptive statistics measuring totally isolated service units is that a service unit that has no shared goals and measures with other service units in the hybrid is a sign of how organizational boundaries generate fragmentation (e.g. Kurunmäki and Miller, 2006). Isolated service unit conducts PM within its organizational boundaries and does not participate in joint PM activities in hybrid, as none of the goals and measures are relevant to it (e.g. Rajala et al., 2020).

Lack of fragmentation is measured by using descriptive statistics measuring the proportion of performance goals or measures that are boundary objects. The logic here is that if all service units consider all performance goals or measures in hybrid as relevant to them, then the maximal level of cohesion has been achieved among the service units in the PM system. Cohesion in PM is here considered as the opposite of fragmentation. The percentages informing from the nonexistent dyadic relationships are measures of fragmentation. Dyadic relationships refer to the interaction between two actors in network theory (Rowley, 1997), and this study examines networks created with PM systems. An example from a dyadic relationship in this context is the following: the basic healthcare unit has a common goal with the oral healthcare named as customer satisfaction. Through the goal of customer satisfaction, the basic healthcare can form a dyadic relationship with oral healthcare (Rajala, 2020). The percentages informing from the nonexistent dyadic relationships tells how many of these type of dyadic relationships are not present in the hybrid. As there are 28 possible dyadic relationships between the different service units in the examined hybrid, the reported percentage tells how many of these dyadic relationships from the 28 possible relationships do not exist.

Evaluating the level of fragmentation in examined hybrid

Although there were no severe conflicts between actors in THA goal setting, the process of objective setting was not easy. In general, the objectives of THA relate strongly to a national-level development agenda, wherein cutting the costs of the public sector is necessary, raising productivity is a virtue and customer value is emphasized. Focusing on customers is a way to overcome some of the well-known problems originating from siloed thinking in service provision. The goals and performance measures seen in Tables 3 and 4 were established in the following decision process described by one interviewee.

We had a working group focusing on goals and outcome measures. There were experts from Tampere [the municipality], Mehiläinen [the private company], region of Pirkanmaa [regional government], and Sitra [an independent trust fund set up by the central government to promote future development and economic growth]… in these working groups, we started defining the key performance areas that are most vital to the [welfare] alliance… these [performance areas] were in line with our customer segments… Then we started to think about performance measures for … key performance areas… We had to work a lot to find measures that were reliable enough and from which we could actually get some data. There were unrealistic ideas about universal measures for well-being, but eventually, we landed on a more realistic level… Every performance area was taken to a more concrete level with the defined performance measures, and we defined the formulas for the measurement and the target levels for the measured results… (Interviewee B)

The municipality and the third sector organization had a separate contract between them. The contract defined what was expected from the community café, and it guaranteed the goal and mission alignment between the third sector organization and other actors in the hybrid. In THA, special attention was given to the whole service system instead of individual organizations (see Tables 3 and 4).

As lack of zeros in the count columns show in Tables 5 and 6, THA succeeded in creating shared objectives between all service production units. Thus, zero percent of service production units were isolated by the chosen goals. The count columns also show that every service unit was linked to some other unit by at least six performance goals (this is 54.5% of all the goals). This indicates that the percentage of nonexistent dyadic relationships was zero. Five of the goals were boundary objects between all the service units (goals 2, 8, 9, 10 and 11 in Tables 3 and 4). Performance goals, such as customer and collaborator satisfaction, connected all the units. Each unit was linked to every other unit by approximately eight performance goals, which is 72.7% of the strategic goals used in the hybrid.

Some units seem to be more isolated than others based on the performance goals, but all the dyadic relationships existed (see Tables 5 and 6). For example, youth center and domiciliary care were linked together by six goals (goals 2, 4, 8, 9, 10 and 11 in Tables 3 and 4). In the PM system, this was the weakest connection between the two units. Basic healthcare, community cafe, library and walk-in clinic were linked together by 11 performance goals. These units were the most connected with c-coefficient value of 1, while the youth center was least connected to other service units. On average, the youth center had seven goals that connected them to other units. Some of this evident fragmentation is understandable, as it can be difficult, for instance, to link the activities of the youth center to the activities of basic healthcare.

Because the community cafe coordinated the operations of the walk-in clinic, it was tightly connected to it strategically. The walk-in clinic did duties of the basic healthcare, and this explains the connection between the walk-in clinic and the basic healthcare unit. Through the walk-in clinic, the community café was connected to the basic healthcare unit.

In Tables 7 and 8, it is evident that THA has created joint performance measures for different service production units. The lack of zeros in the count columns in Tables 7 and 8 reveal that every service unit was linked to some other service production by at least two performance measures. Thus, zero percent of service production units were isolated by the chosen measures, and the percentage of nonexistent dyadic relationships was zero in the performance measurement context. Fragmentation is, therefore, not evident.

There were two boundary objects between all the units. The smallest number in the count columns in Tables 7 and 8 is two, which means that the weakest linked service production units were connected by two measures (i.e. two boundary objects linking together all units). For example, youth service and domiciliary care were only linked by two performance measures, whereas nine measures linked together basic healthcare, community cafe and walk-in clinic.

Each unit was linked to every other unit by approximately five performance measures, which is 41.6% of the strategic measures used in the hybrid. Performance measures, such as collaborator satisfaction survey and input (i.e. cost) measure connected all the units, whereas the measure such as the number of treatment plans made for people who need one was only used by four units. The library, youth services and oral healthcare were less connected to other service units compared to the basic healthcare, walk-in clinic and community cafe, and this demonstrates slight fragmentation in the PM system.

Even this slight fragmentation had consequences. “With the current measures, it was difficult to determine how library succeeded in its functions.” As a result, the performance management practices suffered (Interviewees B and E). Similarly, managing youth services' performance was left in the background while “healthcare and social care issues dominated” the PM (Interviewee B). This domination of healthcare and social care issues meant that the PM system suited well to adult oral healthcare, although not all goals and measures were directly relevant to it (Interviewee B and C).

Interestingly, the fragmentation in performance goals contributed to the fragmentation of the performance measures to some extent (see Tables 5–8). This indicates that fragmented performance goals present a barrier to shared performance measures and effective use of performance information. For example, the goal of collaborator satisfaction has less fragmentation than the goal of the welfare of elderly people. Therefore, performance measures tracking collaborator satisfaction can connect more units together than the measure related to the welfare of elderly people, if the collaborator satisfaction measure is designed to be as inclusive (in terms of relevance to different service units) as the goal related to collaborator satisfaction. Overall, the study results are summarized in Table 9.

Discussion

The evaluation framework created and described in Table 2 provides five novel methods to locate and describe the fragmentation seen in hybrids' PM systems. This framework can be used in problem-solving if fragmentation introduces unnecessary costs (Lowery, 1998), inhibits the desired outcomes, prevents joint performance measurement and information use (Kurunmäki and Miller, 2006; Hodges, 2012), causes uneven distribution of benefits between partners (Camarinha-Matos and Abreu, 2007) or generates mission fragmentation (General Accounting Office, 1997). Once the fragmentation causing problems is identified by using the evaluation framework, one can start to create shared performance goals and measures by either applying the concept of boundary object (e.g. Rajala, 2020; Vakkuri et al., 2021) or by creating less ambiguous goals and measures that are understood similarly across different organizations. In general, the created evaluation framework (see Table 2) and the concept of boundary object (e.g. Star, 2010) provide techniques to break the information silos between different sectors described by Kurunmäki and Miller (2006) and Hodges (2012).

As fragmentation in PM systems has been well-documented (e.g. Johanson and Vakkuri, 2017; Agostino and Arnaboldi, 2017; Alexius and Grossi, 2018), there is a need to developed both concepts and methods that help us to understand, describe and control the fragmentation and its consequences. The evaluation framework fills a research gap arising from the fact that past studies have not adequately addressed the analysis of fragmented PM (e.g. Hodge and Greve, 2007; Thomasson, 2009; Liu et al., 2014). According to the suggested framework, the fragmentation of the PM system can be assessed from the hybrids' goals and measures because joint objectives and their measures can unify public, private and third sector actors (e.g. Battilana and Dorado, 2010; Kastberg and Lagström, 2019; Rajala, 2020), whereas lack of shared goals and measures isolate units from each other (e.g. Campanale et al., 2020). The concrete definitions of fragmented PM given in this study help academics and practitioners to identify the phenomenon and discuss it.

Conceptually, the study contributes to the PM literature concentrating on hybrid organizations (e.g. Alexius and Cisneros Örnberg, 2015; Grossi et al., 2017) by identifying the following new concepts: fragmented PM, coherent PM and exclusive performance goals and performance measures leading to isolated performance regimes (i.e. strategically isolated service units). As the modern forms of governance combining public, private and civil forms of institutional action become more common (Savignon et al., 2018), these new concepts provide valuable analytical tools for practitioners to deal with fragmentation in management arising from various social forms of control (Power, 2000), goal incongruence (Kreps and Monin, 2011) and different institutional logics of markets, state and civil society (e.g. Pache and Santos, 2013; Skelcher and Smith, 2015). The suggested theoretical concepts welcome academics to develop more novel concepts to describe fragmented PM systems and their functioning, as the current conceptual frameworks are lacking in this area (e.g. Moynihan, 2005; Pollitt and Bouckaert, 2011). Developing the theory of fragmented PM would push the PM theories concerning hybrids to search for more concepts describing both dysfunctional and functional PM systems. Here, we propose that adverse fragmentation relates to dysfunctional PM, while beneficial fragmentation is associated with functional PM.

Lastly, we would like to point out that it might not be possible to determine the right level of fragmentation in hybrids' PM systems that would work everywhere. The right type of fragmentation is equally difficult to define. The case might be that different contexts require different solutions because people experience fragmentation and its effects differently depending on the situation. Therefore, we encourage scholars and practitioners to ask themselves what is the right balance between common goals and measures on the one hand and strategic fragmentation on the other hand (e.g. Rajala et al., 2020). Despite the difficulties in determining the optimal level of fragmentation, we think it is useful to understand how fragmentation can be increased or decreased, as these modifications can help managers generate shared performance measures and effective performance information use in hybrids.

Conclusions

This article asked previously unexamined questions addressing how to evaluate the level of fragmentation between sectors in a hybrid organization's PM systems (see Billis, 2010; Grossi et al., 2017; Johanson and Vakkuri, 2017). As an answer, an evaluation framework described in Table 2 was created and tested in a case organization called THA. In the empirical testing of the evaluation model, isolated service units, nonexistent dyadic relationships, c-coefficients and lack of goals and measures operating as boundary objects were used to describe the fragmentation. The evaluation results mostly show low-level fragmentation in the examined hybrid, but even this fragmentation caused some frustration in the more isolated units. The study also showed that creating fragmented goals is also a barrier to shared performance measures. Overall, the evaluation model's empirical testing was successful, demonstrating that fragmentation can be analyzed with the proposed evaluation model.

The study had many limitations. We examined only one case, indicating that only future research questions and preliminary hypotheses can be derived from the results of this study. Although we controlled responded bias (e.g. Brown, 2001) by using triangulation, taking part in the day-to-day management activities while observing the actual use of performance goals and measures in case organization would have provided more in-depth knowledge about the existence of the fragmentation. One limitation relating to the empirical data was that we did not interview anyone from personnel, which means that it is impossible to examine whether fragmentation of PM was more evident among the personnel providing the services. What comes to the analysis of fragmentation, there is no established interpretation of the intensity of fragmentation (e.g. Kurunmäki and Miller, 2006; Hodges, 2012; Rajala et al., 2020). Therefore, we recommend that researchers and practitioners consider the results as our interpretation of the numbers. Further academic debates are needed concerning how these numbers should be interpreted. Regardless of the interpretation, the numbers are valid, and these can be used in future studies. Thus, the interpretation does not prevent the accumulation of knowledge.

The study also identified how goals and measures could generate dyadic relationships between different hybrid actors and create cohesion in the PM systems. As these relationships have remained mostly unexamined (e.g. Agostino and Arnaboldi, 2018; Savignon et al., 2018), we propose the following two hypotheses for future studies:

H1.

Fragmented goals lead to fragmented performance measures.

H2.

Performance measures can create fragmentation, although performance goals generated cohesion between public, private and third sector actors.

Future studies could examine the benefits and disadvantages of fragmentation. For example, does isolated performance regimes in hybrids lead to poor collaboration between public, private and third sector organizations or is it sometimes better to isolate different actors from each other? Moreover, academics should explore whether fragmentation results from the inability to consider the hybridity in performance management design. Lastly, further developments on how to measure fragmented PM are also needed.

Table 1

Implications of the hybridity to PM

An attribute of the hybrid	Implications of the attribute to goal setting	Implications of the attribute to performance measurement
Mixed ownership	The goals of the taxpayers, business owners and donors of the voluntary sector must be considered	Measures need to track the achievement of the goals of the taxpayers', business owners' and members' of the voluntary sector
Combinations of governance modes	Public elections, share ownership and private elections produce goals	Public elections, share ownership and private elections require performance information produced with measuring
Goal incongruence	Value disputes between public value, shareholder value and social value (i.e. societal betterment through social inclusion in the field of activities lacking government or private sector production according to Akingbola et al. (2019))	Conflicts about the right measures or alternatively conflicting metrics
Competing institutional logics	Goals concerning mission alignment, market shares and effectiveness of market interventions	Measures of competitiveness, public sector cost-effectiveness and mission achievement
Distinctive human resources	Goals need to motivate public servants, private managers and members of a volunteer association	Measures need to motivate public servants, private managers and members of a volunteer association
Multiplicity of funding arrangements	Goals related to the appropriate use of tax money and donations as well as goals driving returns for investments	Budgetary accounting is needed as well as measures tracking wealth increases of the owners and misuse of donations
Operational priorities	Collective choice, individual choice and mission-oriented choice all introduce goals	Collective choice, individual choice and mission-oriented choice propose demands for performance measurement
Forms of social control	Goals needed to establish corporate governance, political and administrative control and rules for membership	Measures needed to establish corporate governance, political and administrative control and rules for membership

Table 2

Assessing the fragmentation in hybrids' PM systems: the evaluation framework

Evaluation aspect	Coding categories used	Coding example	Evaluation criteria for fragmentation
Performance goals	Coding category 1: a goal or measure for basic healthcare unit Coding category 2: a goal or measure for adult oral healthcare Coding category 3: a goal or measure for library Coding category 4: a goal or measure for maternity and children's health clinic Coding category 5: a goal or measure for domiciliary care Coding category 6: a goal or measure for youth center Coding category 7: a goal or measure for walk-in clinic Coding category 8: a goal or measure for community café	The goal called satisfaction to collaboration was a boundary object between all units, and it was coded with eight labels as each unit shared this goal	The percentage of totally isolated service units that have no shared performance goals (i.e. boundary objects) with other service units The percentage of performance goals that are boundary objects between all units The percentage of nonexistent dyadic relationships from the total amount of dyadic relationships (there are 28 possible dyadic relationships between eight service units) The average number of goals that connect any two service units together C-coefficient value of zero in the context of performance goals shows the nonexistent relationship between two service units in the hybrid
Performance measurement		The measure called satisfaction to collaboration was a boundary object between all units, and it was coded with eight labels as each unit shared this measure	The percentage of totally isolated service units that have no shared performance measures (i.e. boundary objects) with other service units The percentage of performance measures that are boundary objects between all units The percentage of nonexistent dyadic relationships from the total amount of dyadic relationships (there are 28 possible dyadic relationships between eight service units) The average number of performance measures that connect any two service units together C-coefficient value of zero in the context of performance measures shows the nonexistent relationship between two service units in the hybrid

Table 3

First set of goals and measures for the hybrid organization

Goals and their relevance

Measures and their relevance

The welfare of the children and youth •Relevant to –Basic healthcare and a walk-in clinic run by the community cafe (these actors promote the well-being, treat physical pains, handle mental problems and help with sexual health issues) –Library (promotes the well-being and provides information about health) –Youth center (promotes the well-being and health of children and teenagers) –Maternity and child health clinic (promotes the well-being while advocating healthy habits of the children and their families) •Not relevant to –Domiciliary care (provides services only to elderly people) –Oral healthcare for adults (provides services only for adults)

Life satisfaction, perceived health and oral health of fourth and fifth graders (relevant to basic healthcare, library, maternity and child health clinic and youth center because children are their service users) Perceived health, symptoms and sicknesses, sexual health, oral health and functionality of family life (relevant to basic healthcare, library, youth center community café and walk-in clinic because teenagers are their service users) • The content of these measures is aggregated from the below measures –Life satisfaction of fourth and fifth graders (descriptive statistics) –National survey reporting how satisfied pupils are with their lives –Perceived health of fourth, fifth, eight and ninth graders –The proportion of pupils who self-assessed their health status as poor –The proportion of pupils who frequently have • Neck and shoulder pain • Abdominal pain • Sleep disorders • Headaches • Sore throat • Flu • Cough •Itchy or watery eyes –Oral health of fourth, fifth, eighth and ninth graders (descriptive statistics) –A national survey reports how often do pupils brush their teeth –Symptoms and sicknesses of eighth and ninth graders (descriptive statistics) –The proportion of pupils who frequently have • Neck, shoulder and lower back pain • Abdominal pain • Sleep disorders • Headaches • Fatigue and weakness • At least two different symptoms every day –The proportion of pupils who have diagnosed with a long-term medical condition –Sexual health of eighth and ninth graders (descriptive statistics) –The proportion of pupils who • Have had sex • Have not used contraception in their latest sexual intercourse • Need someone to talk about relationships and sexuality • Need more information about the body • Need free condoms • Need cheaper birth control • Need more information about sexually transmitted diseases • Need more information on how to get pregnant • Are in a relationship • Need information about chlamydia tests • Need services related to sexual health at weekends and evenings –The functionality of family life of the eighth and ninth graders –The proportion of pupils who • Think that the financial situation of the family is good, bad and so on • Work alongside studying • Have troubles communicating with parents • Have a good dialogical connection with their parents • Looks after a family member or other close relative every week • Live mostly with one parent • Live with two parents that live in different addresses

Source(s): Documents 5 and 6

Table 4

Second set of goals and measures for the hybrid organization

Goals and their relevance	Measures and their relevance (in general, measures are relevant if they relate to service users of the production unit)
(2)Coverage of major national diseases • Relevant to all service units because every unit promotes health (3)Coverage of major national mouth diseases • Relevant to all service units because every unit promotes health (4)Welfare of treated patients who have at least one disease belonging to the major national diseases • Relevant to all service units promoting health except youth center and maternity and child health clinic because the patients are not service users of these two units	(3)Treatment plans are made for coronary heart disease and diabetes patients between the ages of 18 and 65 (relevant to basic healthcare, domiciliary care, community café and walk-in clinic) • Measures how many patients with a certain diagnose have treatment plans (reported in percent) (4)Treatment plans are made for adult oral healthcare patients (relevant to basic and oral healthcare, community café and walk-in clinic) • Measures how many patients using the services of oral healthcare have treatment plans (reported in percent) (5)Number of diabetes and coronary heart disease patients with good therapeutic equilibrium divided by all diabetes and coronary heart disease patients (reported in percent and relevant to basic healthcare, domiciliary care, community café and walk-in clinic)
(5)Welfare of elderly people (6)Health service use of the elderly (7)Elderly's ability to live home • All three goals are relevant only to basic healthcare, walk-in clinic, library, domiciliary care and community café as these provide services to elderly people and face the costs of services	(6)Elderly's ability to function and live home (relevant to basic and oral healthcare, domiciliary care, community café and walk-in clinic) • Number of service users with low need for services divided by all service users in elderly care (reported in percent) (7)The reduced need to use emergency services (relevant to basic healthcare, domiciliary care, maternity and child health clinic and oral healthcare) • Changed cost per customer in emergency services (8)Changed per-patient cost of hospital services (relevant to basic healthcare, domiciliary care, maternity and child health clinic)
(8)Curbing the cost of special medical care • Relevant to all service units because every unit promotes health and faces the costs of care	(9)Per-patient cost of special medical care (relevant to basic healthcare, maternity and child health clinic, domiciliary care, community café and walk-in clinic)
(9)Customer satisfaction (10)Collaborator satisfaction • Both goals are relevant to every service unit because these units have customers and collaborators	(10)Customer satisfaction for services (relevant to all units except domiciliary care as they have their own customer satisfaction measures) • Net promoter score (NPS) –The survey measures the willingness of customers to recommend public services of the THA to others – As a follow-up question proposed to the dissatisfied customers, the survey asks what should be improved – As a follow-up question proposed to the satisfied customers, the survey asks what was good (11)Collaborator satisfaction (relevant to all units) • NPS survey send to collaborators, such as associations, condominiums, companies and residents living in the suburb –The survey measures the willingness of collaborators to recommend collaboration with the actors of the welfare center to others – As a follow-up question presented to the dissatisfied collaborators, the survey asks what should be improved –As a follow-up question presented to the satisfied collaborators, the survey asks what was good
(11)Stay within the budget • Relevant to all service units because their expenses are part of this budget	(12)The combined overall cost of units in the welfare alliance (relevant to all units)

Source(s): Documents 5 and 6

Table 5

First co-occurrence table relating to performance goals

	Basic healthcare		Community cafe		Domiciliary care		Library
	Count	c-coefficient	count	c-coefficient	count	c-coefficient	count	c-coefficient
Basic healthcare	–	–	11	1.00	10	0.91	11	1.00
Community cafe	11	1.00	–	–	10	0.91	11	1.00
Domiciliary care	10	0.91	10	0.91	–	–	10	0.91
Library	11	1.00	11	1.00	10	0.91	–	–
Maternity and child health clinic	7	0.64	7	0.64	6	0.55	7	0.64
Oral healthcare	8	0.73	8	0.73	7	0.64	8	0.73
Walk-in clinic	11	1.00	11	1.00	10	0.91	11	1.00
Youth center	7	0.64	7	0.64	6	0.55	7	0.64
On average	9		9		8		9

Table 6

Second co-occurrence table relating to performance goals

	Maternity and child health clinic		Oral healthcare		Walk-in clinic		Youth center
	Count	c-coefficient	count	c-coefficient	count	c-coefficient	count	c-coefficient
Basic healthcare	7	0.64	8	0.73	11	1.00	7	0.64
Community cafe	7	0.64	8	0.73	11	1.00	7	0.64
Domiciliary care	6	0.55	7	0.64	10	0.91	6	0.55
Library	7	0.64	8	0.73	11	1.00	7	0.64
Maternity and child health clinic	–	–	7	0.88	7	0.64	7	1.00
Oral healthcare	7	0.88	–	–	8	0.73	7	0.88
Walk-in clinic	7	0.64	8	0.73	–	–	7	0.64
Youth center	7	1.00	7	0.88	7	0.64	–	–
On average	7		7		9		7

Table 7

First co-occurrence table relating to performance measures

	Basic healthcare		Community cafe		Domiciliary care		Library
	count	c-coefficient	count	c-coefficient	count	c-coefficient	count	c-coefficient
Basic healthcare			9	0.75	8	0.67	5	0.42
Community cafe	9	0.75			6	0.55	4	0.40
Domiciliary care	8	0.67	6	0.55			2	0.18
Library	5	0.42	4	0.40	2	0.18
Maternity and child health clinic	7	0.58	4	0.33	5	0.50	4	0.50
Oral healthcare	6	0.50	5	0.50	4	0.40	3	0.38
Walk-in clinic	9	0.75	9	1.00	6	0.55	4	0.40
Youth center	5	0.42	4	0.40	2	0.18	5	1.00
On average	7		6		5		4

Table 8

Second co-occurrence table relating to performance measures

	Maternity and child health clinic		Oral healthcare		Walk-in clinic		Youth center
	count	c-coefficient	Count	c-coefficient	count	c-coefficient	count	c-coefficient
Basic healthcare	7	0.58	6	0.50	9	0.75	5	0.42
Community cafe	4	0.33	5	0.50	9	1.00	4	0.40
Domiciliary care	5	0.50	4	0.40	6	0.55	2	0.18
Library	4	0.50	3	0.38	4	0.40	5	1.00
Maternity and child health clinic			4	0.44	4	0.33	4	0.50
Oral healthcare	4	0.44			5	0.50	3	0.38
Walk-in clinic	4	0.33	5	0.50			4	0.40
Youth center	4	0.50	3	0.38	4	0.40
On average	5		4		6		4

Table 9

Results of the evaluation

Evaluation aspect	Evaluation results
Strategic goals	The percentage of totally isolated service units that have no shared goals with other service units: 0% Five goals were boundary object between all the service units (45.5%) The percentage of nonexistent dyadic relationships: 0% On average, eight goals connected one service unit to another in the hybrid The c-coefficient values were between 0.55 and 1 indicating rather good coherence in PM
Performance measurement	The percentage of totally isolated service units that have no performance measures with other service units: 0% Two of the measures were boundary object between all the service units (16.6%) The percentage of nonexistent dyadic relationships: 0% The average number of performance measures that connect any two service units together was 5 The c-coefficient values were between 0.18 and 1, indicating low level of fragmentation in PM

Appendix 1

Interview questions

1. Could you describe how the hybrid operates now?

2. Goal setting

   • Can the management model reduce the diversity of objectives and the lack of compatibility between different objectives?

   • What kind of process was used to determine the goals?

   • Are there goal conflicts and if there are how are these resolved?

   • How do the chosen goals serve all the service units, and are goals serving the purposes of THA?

   • How is the principle of shared value seen in everyday operations of THA?

3. Organizing

   • Are the responsibilities defined clearly and who defined them?

   • Do different actors in hybrid know what their responsibilities are?

   • How do you know that those responsibilities are clearly defined?

   • Why was the coordination group formed the way it was formed?

4. Commitment

   • How does THA promote commitment and how can it be seen in operations?

   • What principles are financial and non-financial incentive systems based on?

   • How are different actors in the hybrid taken into consideration in the feedback system, and does information reach different parties?

   • How would you evaluate the commitment level of different actors now?

5. Measurement and information flows

   • How are performance measures defined and why did you choose the ones you chose?

   • How do the chosen measures help explain why THA succeeded or failed? How do you demonstrate the shared value?

   • Do the current measures inhibit optimization?

   • How do you follow the goal achievement?

   • Is the information available to all, and are the information systems compatible with each other?

   • Do you have common information systems?

Appendix 2

Document 1. Description of welfare alliance model: a PowerPoint presentation by the local government

Document 2. Alliance model used in the welfare alliance and for provincial social care: a PowerPoint presentation by public manager

Document 3. Goals and measures of the welfare alliance: an attachment from a contract between the hybrid members

Document 4. Service plan: an attachment from the contract between the hybrid members

Document 5. Implementation contract: a contract between the hybrid members

Document 6. Commercial model: description of the contract used to form the welfare alliance