A stakeholder perspective on public sector innovation: Linking the target groups of innovations to the inclusion of stakeholder ideas

Abstract

Public sector innovation scholarship has not yet systematically explored how the target context (or output phase) of innovations impacts the early phases of innovation processes. This study theorizes and tests whether innovating organizations are more sensitive to ideas from particular stakeholder groups depending on the target group of said innovation. Using a large-scale dataset from the Australian Public Service, the results show that innovations with external target groups are more likely to be built on ideas from external stakeholders (compared with internal stakeholders) and—within the group of internal stakeholders—on ideas from managers (compared with non-managerial employees). Practical and scholarly implications are discussed.

Points for practitioners:

Innovations benefit from the inclusion of internal and external stakeholder ideas, both substantively (appropriate knowledge leads to better end products) and symbolically (innovations need to be deemed legitimate, and receive support from the actors that will be primarily impacted by the innovation).

Innovating organizations need to be aware of the perceptions of the stakeholders affected by the innovation, and properly sense, capture and translate the ideas of those stakeholders in the innovation process.

Introduction

Recent decades have witnessed growing pressures on public organizations to improve their performance in increasingly complex, demanding and volatile environments (Christensen and Laegreid, 2007; Pollitt and Bouckaert, 2017). With the gap between (high) expectations and (shrinking) resources ever widening, public organizations are urged to get the most out of the limited resources at their disposal, and to do things differently through innovation. Innovation, therefore, has become a central topic of scholarly and practitioner interest. Innovation can be a major source of improved cost savings, quality of services and overall performance, benefits which positively affect businesses and citizens who rely on an efficient and effective public sector (Cinar et al., 2019; De Vries et al., 2015).

Public sector innovation has, however, long been considered an oxymoron (Bommert, 2010; Torfing et al., 2020), as the concept of innovation was introduced by the economist Schumpeter (1942) as a mechanism to explain economic changes, and public services are far less affected by these dynamics than market actors. Furthermore, the intrinsic bureaucratic nature of most public sector organizations, with rigid regulation, command-and-control operations and resistance to change, fueled the perception that the concept of innovation was ill suited to the public sector (Hartley et al., 2013). Several decades of research (and review articles) into public sector innovation research have, however, provided evidence for the relevance of public sector innovation, and also led to a convergence in the literature regarding the definition of innovation (centered on perceived newness) and the conceptualization of innovation as an increasingly open and collaborative process (Buchheim et al., 2020; Cinar et al., 2019; De Vries et al., 2015).

First, most public sector innovation scholarship bases its definition of innovation on Rogers (2003), who sees it as “an idea, practice, or object that is perceived as new by an individual or other unit of adoption” (Borins, 2000; De Vries et al., 2015). Two important take-aways from this definition are that (a) innovations relate to perceived newness and (b) context is critical. A practice that has become routine in one context may be novel (and therefore innovative) for another (Anderson et al., 2014). Curiously though, while the relevance of contextual perceptions forms the backbone of most definitions of innovation, little is known about how the context for which innovations are developed affects innovation processes.

Second, innovations are increasingly considered as open processes that include the purposeful connection of internal and external ideas (Hameduddin et al., 2020; Mergel and Desouza, 2013; Sørensen and Torfing, 2011). Useful knowledge is widely distributed, both within and outside the organization, and can be identified, connected and leveraged to facilitate innovations (Chesbrough et al., 2006). Little is known, however, about the conditions under which innovations develop from the ideas of particular stakeholder groups. Studying when stakeholder ideas are translated into innovations is important as it gives insight into the conditions under which organizations become “open”, that is, receptive to different types of ideas in the development of innovations.

The core premise of this study is that innovating organizations will be more sensitive to ideas from particular stakeholder groups depending on the target group of the said innovation. Innovations are novel ideas or objects that need to find their place in a particular context. To achieve this, innovations need to be deemed legitimate, and receive support from the actors that will be primarily impacted by the innovation (Lægreid et al., 2011; Osborne, 1998; Verhoest et al., 2007).

We address the following research question: what is the relation between the types of stakeholders (internal and external) at whom innovations are targeted, and the likelihood that these innovations build on ideas from particular stakeholder groups (internal [managerial and non-managerial employees] and external)? Empirically, this study relies on data from the annual survey of Australian Public Service (APS) employees, addressed at the senior executive level (N = 1.369).

This study contributes to the public sector innovation literature in the following ways. First, and in contrast to the private sector literature where innovation has been an established field of study for decades, public sector innovation studies often fail to provide a theoretical underpinning for predicting and explaining innovation dynamics (Clausen et al., 2020; De Vries et al., 2015). While definitions of innovation consistently connect the concept to the implementation context in which the innovation is obtained (De Vries et al., 2015; Rogers, 2003), and scholarship recognizes the role of different stakeholders as potential facilitators of innovation (Sørensen and Torfing, 2011), we lack a theoretical understanding of when stakeholders impact innovation processes. This study complements existing research that distinguishes innovations by the organizational practices that are affected (governance, processes, products/services; Buchheim et al., 2020; De Vries et al., 2015). A stakeholder-centered perspective distinguishes innovations on the basis of the target group from whom support for the innovation needs to be achieved. Depending on the target group—that is, the stakeholders that will directly observe, use and value the innovation—organizations are expected to rely on ideas from these target groups to maximize the likelihood that the innovation process includes relevant knowledge and adheres to expectations.

Second, we contribute to the public sector innovation field that requires a rebalancing from “a qualitative dominance to using other methods, such as surveys, experiments and multi-method approaches” (De Vries et al., 2015: 146). We use large-scale data from the APS. The Australian government has a longstanding interest in promoting innovation and has been seen as a frontrunner in several public sector reform doctrines (Demircioglu, 2019; Halligan, 2007). The importance of innovation has been highlighted by successive Australian governments since the mid-2000s. In 2015, prime minister Turnbull declared that “This is a century of ideas, this is a time when Australia's growth, when our living standards, when our incomes will be determined by the human capital, the intellectual capital that all of us have. By unleashing our innovation, unleashing our imagination, being prepared to embrace change, we usher in the ideas boom” (Australian Government DIIS, 2015, p. 2).

In the remainder of this article, we first present our theoretical framework and hypotheses. Next the methods and data are introduced to empirically test the hypothesis. The results are then presented, after which we discuss our findings in light of existing theories, and summarize our core findings and limitations in the concluding section.

Theoretical framework: a stakeholder perspective on innovation

Innovation is defined as the generation and practical adoption of new ideas or behaviors (Amabile, 1988; Damanpour and Schneider, 2008). The newness of these ideas or behaviors is evaluated against their adoption context, that is, innovations are considered as such when they are new for the specific context (Anderson et al., 2014); cf. Rogers’ (2003, p. 12) definition of an innovation as an “idea, practice, or object that is perceived as new by an individual or other unit of adoption”.

Few studies, however, consider the actors present in this implementation context as a primary factor to explain how innovations come about. Nonetheless, research has indicated the importance of stakeholder expectations for innovation. Osborne (1998), for instance, shows that innovations result from a pressure to adhere to prevailing expectations within the institutional field. Verhoest et al. (2007: 488) provide additional support for this observation: “[Flemish government organizations] oriented themselves quite strongly toward what they perceived as the expectations of their customers, interest groups and, sometimes quite indirectly, their political principals (or some factions of these principals) in order to enhance their legitimacy”.

This study takes a stakeholder perspective that distinguishes innovations on the basis of the target group from whom support needs to be achieved for the innovation to be successful. At the level of the dependent variable, we distinguish two types of innovation through the stakeholders that it mainly targets:

“Internal innovations” are developed for internal stakeholders such as their own workgroup, department or organization. Process innovations that focus on internal business processes, such as the digitalization of service delivery processes, are examples of these innovations (De Vries et al., 2015; Walker, 2014).

“External innovations” are directly aimed at external stakeholders such as citizens, businesses, other organizations, etc. New products or services that are provided to certain users are good examples of this type of innovation (Damanpour and Schneider, 2008), as are process innovations aimed at citizen–state interactions (de Vries et al., 2015), e.g. innovative ways to redirect traffic during the construction of roads (Callens et al., 2021).

Our central argument is that the implementation phase, and the stakeholders active in the implementation context, is closely connected to the idea-generation phase of innovations. In the next section, we will outline how external innovations will be more likely to be built on the ideas of external stakeholders, while the ideas of internal stakeholders will more be likely to be at the basis of internal innovations. We adopt a stakeholder perspective that distinguishes between external and internal stakeholders, by linking the input phase (i.e. where ideas for innovations come from) and the output phase (where innovations end up). In doing so, we build on previous studies pointing to the motivations for (Godenhjelm and Johanson, 2018; Rogers, 2003) and managerial challenges of (Cinar et al., 2019; de Vries et al., 2018) stakeholder involvement in innovation processes. A recurring and shared theme in these studies is that organizations need support from partners in terms of resources to sustain innovations. One way to achieve such support is to build on the ideas of stakeholders that are affected by the innovation. Cinar et al. (2019, p. 284) state that “the content of innovation should be considered and cautiously designed in accordance with the needs of organizations and citizens”. According to Bommert (2010), the involvement of relevant stakeholders is critical, not only in terms of innovation creation, but also to ensure institutional embeddedness. Put differently, the inclusion of relevant stakeholders benefits innovation projects, for both substantive reasons (e.g. accessing local knowledge) and symbolic reasons (e.g. creating a sense of ownership, demonstrating support). ¹

Hypotheses

The study proceeds in two steps as (a) it addresses how ideas from internal and external stakeholder groups are related to innovation targets (RQ1), after which (b) the ideas from different internal stakeholder groups (managers and employees) are related to innovation targets (RQ2) (see Figure 1 and see online supplementary documents).

External innovations refer directly to novel and improved processes, outputs and services toward external stakeholder groups. The group of external stakeholders is diverse, and ranges from political actors to other public, non-profit or for-profit organizations and associations to users and citizens. Each of these stakeholder groups comes with its specific knowledge and expertise, but also demands and attitudes regarding an innovation (Cinar et al., 2019). A shared characteristic of external innovations, however, is that they will be used by external stakeholders, and that their value is ultimately decided by external feedback (Hauser et al., 2006; Karlsson and Skålén, 2015). This gives organizations (and the stakeholders in question) an incentive to build on ideas from external stakeholders to maximize the likelihood that resulting innovations are in accordance with external demands, and have sufficient problem-solving capacity for external actors.

While the bases of support and type of knowledge may differ, this general mechanism is expected to hold for different external stakeholder groups. First, innovations targeted at citizens and end-users often aim to address societal needs and improve service delivery (de Vries et al., 2018). In such cases, citizens and end-users bring local knowledge and support, while demanding more personalized and tailored innovations that are consistent with their existing values, easy to understand and use (Albury, 2005; Rogers, 2003). Innovating governments are therefore incentivized to include these end users to benefit from their local expertise and avoid opposition. An example from the APS concerns the efforts from the Australian Taxation Office to make it easier for non-residents to meet their Australian Goods and Services Tax through a new, simplified and user-oriented reporting and payment system. The system was developed using extensive user research and feedback—not just from taxpayers, but also from tax agents, businesses and international organizations—and has led to an strong increase in uptake by users (Public Sector Innovation Network, 2019). Second, radical ideas that significantly break from the status quo require political support and authorization and often originate at the political level. In a Federal government system such as Australia, intergovernmental competition and political renewal (through a change of government or even a single minister) may fuel innovation. An example from the APS relates to the outsourcing of employment services introduced in 1996—one of the first comprehensive attempts internationally to apply market principles to the provision of active labor market assistance for job seekers—an innovation that was introduced by political actors and then implemented by civil servants (Australian Public Service Commission, 2010). Third, in innovations that target business actors too, government actors tend to enter into partnerships—as opposed to more top-down hierarchical relations—with said businesses. An example from the APS is the Solar Panel Validation Initiative, in which the problem of unapproved and potentially poor-quality solar panels being installed on Australian homes and businesses was tackled by a joint initiative between the Clean Energy Regulator and the solar industry (Public Sector Innovation Network, 2019).

Internal innovations, in contrast, are developed for internal target groups (own workgroup, department or agency). The underlying objective of innovations toward internal targets is often related to efficiency in the manner in which internal processes are designed and conducted (Un and Asakawa, 2015). According to de Vries et al. (2018), innovation adoption is frequently driven by internal organizational wishes, such as improved organizational productivity and performance, or an increase in employees’ job satisfaction. In contrast to external innovations, then, the innovativeness and success is valued internally by managers and employees after evaluating whether changes have established cost reductions and process improvements. From the change management literature, we know that innovators need to take the potential of internal resistance into account—caused by high-risk aversion, rent-seeking culture, increasing workload owing to innovation, power struggles, lack of knowledge about innovation, unclear responsibilities and complexity of innovation (Cinar et al., 2019). The likelihood of these factors coming into play increases as innovations are developed without adequate involvement of staff (Keast and Brown, 2006). Different attitudes arise when staff are more involved or when innovations build on ideas and feedback from internal stakeholders that are affected. For instance, de Vries et al. (2018) reported how the Works Council—a “shop-floor” organization that represents employees—was asked to give feedback, or even formal consent, to teleworking innovations, which led to changes in formal working practices. This example illustrates a wider belief in the literature that potential barriers of resistance can be regarded as opportunities to design and modify the innovation so that it becomes more appropriate to the particular context (Borins, 2014; Cinar et al., 2019).

To summarize, we expect that the implementation context of innovations will be related to the stakeholder input that innovations build upon, as both the innovating organization and the affected stakeholder group benefit from the collaboration to develop an innovation that is in line with stakeholder demands. We propose the following hypothesis:

Hypothesis 1: Ideas originating from external stakeholders are more likely to be the source of external innovations (compared with ideas from internal stakeholders).

It is important to note at this point that our expectations should be understood in relative terms (this also holds for the later hypotheses in the study), that is, by saying that external ideas are generally more related to external innovations, we are neither saying that external parties are never the source of internal innovations (to give but one example, external consultants are often involved in optimizing process operations—see e.g. Caloghirou et al.’s [2016] study on private sector involvement in the development of a standardized electronic service scheme for Greek municipalities), nor that internal ideas are never the source of external innovations. In fact, our second research question pertains to precisely this question: how are ideas from internal stakeholder groups related to different innovations (see RQ2 in Figure 1). In the next paragraphs, we recognize the internal heterogeneity of organizations and contrast two perspectives. Because of the lack of prior empirical evidence to substantiate these perspectives, we will refrain from formulating hypotheses.

A first perspective would suggest that external innovations are more likely to result from managerial ideas. External innovations have direct repercussions for an organization's legitimacy, as they are ultimately evaluated against and decided by the feedback of external stakeholders (Karlsson and Skålén, 2015). Concerns with the organization's legitimacy in a wider social system are strongly in the realm of management activities and role conceptions (Brunsson, 2002; Thompson, 1967). The ideas of managerial employees are therefore more aligned toward the assumptions and expectations of the external environment. Additionally, managers are more concerned with the strategic positioning of an organization in its environment and the benefits of innovation for this strategic positioning (Danneels, 2002; Un and Asakawa, 2015). As such, this perspective expects that this type of strategic management and environmental awareness is more likely to be present among managerial staff. In contrast, non-managerial employees’ main concern lies with activities related to executing the technical task, e.g. processing materials or other inputs, or handling clients (Thompson, 1967). The cores of non-managerial employees’ knowledge bases are shaped by operational knowledge that develops from experience (Hartley and Rashman, 2018). These experiences contribute to an intimate understanding of internal processes and routines, which makes non-managerial employees well suited to delivering ideas for innovation that have a higher likelihood of achieving cost reductions and process improvements (Un and Asakawa, 2015).

A second perspective formulates some counter arguments. The service innovation literature has argued that frontline employees are well suited to contribute to organizations’ service innovations (de Jong and Vermeulen, 2003; Karlsson and Skålén, 2015; Lusch et al., 2007). Ordanini and Parasuraman (2011) argue that service innovation without frontline employee involvement may make the implementation difficult and the innovated service lacking in customer orientation, thus hampering the responsiveness and problem-solving capacity of the innovation. External stakeholders’ ideas are important for external innovations (cf. H1), yet their inputs are often latent and not easy to capture and identify (Matthing et al., 2004). Non-managerial employees are the organizational members most proximate to end users of an organization's products and services. They serve as linking pins that couple expectations from the external environment with their own local routines, and are best placed to sense, capture and translate these expectations in innovative ideas. Managerial staff, in turn, can be argued to possess important ideas to realize successful innovations for internal target groups. Managerial staff can be expected to have a broad understanding of what makes the organization work as a whole. Managerial staff have an important role in mediating between an organization's operational staff and the external environment, e.g. by procuring the resources for carrying out the technical functions and by controlling (Thompson, 1967). Managerial staff is tasked with instituting effective control mechanisms to monitor progress toward organizational and project goals. This role requires managers to prioritize that business operations are run efficiently and to focus on details in addition to the broad picture, all qualities which are expected to benefit internal innovations more than external innovations.

Data

In order to test the above hypotheses we make use of the APSC’s 2016 annual survey of APS employees. The APS is the federal civil service of the Commonwealth of Australia responsible for the public administration, public policy and public services of the departments and executive and statutory agencies of the Government of Australia. It houses a large variation of agencies ranging from agencies active in policy development, such as the Department of Finance, to agencies active in policy implementation, such as the Bureau of Meteorology, to regulatory agencies such as the Australian Communications and Media Authority, to agencies providing specialist advice to government, such as the National Audit Office. Within the APS, the APSC is a small policy agency within the portfolio of the Department of Prime Minister and Cabinet. One of the APSC's recurring activities is the organization of the APS Employee Census, an annual survey sent to all APS employees, which is used to collect confidential attitude and opinion information from APS employees on important issues in the workplace, and serves as the data source for this article. The survey captures employee attitudes on important issues such as wellbeing, innovation, leadership, learning and development, and the engagement of the APS workforce. Although several waves of this survey exist, the 2016 data focus extensively on innovation (understanding and measuring innovation as well as the ideas and targets of innovations).

The 2016 APS employee census was administered to all available APS employees from 105 agencies across the APS. This census approach provides a comprehensive view of the APS and ensures that no eligible respondents are omitted from the survey sample, removing sampling bias and reducing sample error. The census ran from 9 May to 10 June 2016. Overall, 96,672 APS employees responded to the employee census, a response rate of 69%. The questions regarding innovation were asked solely to the senior executive service (SES), which reduced the number of respondents from 96,672 to 2155 (yet which kept organizational variety). These data were found to be representative. However, owing to item non-response, our sample further declined to 1369 observations. ² When comparing this sample with the initial, representative sample of 2155 observations, no significant differences could be observed.

There are several motivations for relying on SES-level answers. First, the SES has the best vantage point for viewing the entire organizational system, and is therefore best suited to answering questions on innovation for the workplace as a whole. Second, a “position-based” approach is the most efficient way of doing comparative research where organizational differences are large (Aberbach et al., 1981). Third, the perception of SES about actual innovation within the organization will also influence actions and the way in which senior staff manage their agency. In short, and although focusing purely on SES perspectives regarding innovation drastically reduces sample size, it is a well thought out approach leading to several advantages.

Dependent variable: target groups of innovation

After respondents were asked to think about the most significant innovation that had been implemented by the workgroup in the last 12 months, the target of innovation was measured using the question “Who was the main target for this most significant innovation? ”, with the following answer options: “Your workgroup, department or agency”, “government ministers”, “other government organizations”, “individual citizens”, “businesses”, “not-for-profits”, “non-governmental organizations”, “interest groups” and “business associations”.

This question was operationalized as a dummy, set to zero if the respondent answered “Your workgroup, department or agency” (i.e. internal innovation) and set to one otherwise (i.e. external innovation).

Main independent variable: ideas from stakeholder groups

In order to examine the sources of innovative ideas, respondents were asked “Where did the idea for this most significant innovation primarily come from?” This question was coded as idea from internal stakeholder if the respondent answered “higher level management”, “staff at levels below your leadership group”, “your leadership group” or “other branch of your department or agency”; and coded as idea from external stakeholder if the respondent answered “businesses as suppliers, clients or consultants”, “professional organizations”, “participation in conferences by your or your staff”, “not-for-profit organizations, “non-governmental organizations, interest groups, or business associations”, “feedback and comments from citizens”, “universities and public research organizations”, “other government organizations outside your department” or “your minister or minister's office”.

Control variables and descriptive statistics

Apart from including the source of the innovation, we also include control variables such as gender, age (under 40 years, 40–54 years and 55 years or older) and organizational size (small, i.e. less than 250 employees; medium, i.e. 251–1000 employees; and large, i.e. 1001 or more employees).

The descriptive statistics as well as a Pearson correlation matrix are presented in Table 1 (see online supplementary documents). The table shows that roughly 49% of innovations are targeted at external stakeholders, while the rest are aimed at improving the workgroup and organization itself. Nonetheless, roughly 90% of innovations build on internal ideas, while 10% come from external parties (5%) or the government (5%).

Methods

To examine the degree to which innovation that targets external parties depends on the origin of the idea, we make use of a logit model. This kind of approach is preferred as it is able to account for models whereby the dependent is operationalized as a dummy, as is the case in our situation (“0” = internal innovation/“1” = external innovation). Using this method, we calculate odds ratios; for a unit increase in $x_k$ , the odds of a lower outcome compared with a higher outcome are changed by the factor exp ( $-{\beta }_k$ ), holding all other variables constant. This makes the interpretation of our results different compared with a standard regression. For instance, an odds ratio of 2 for variable gender (female) means that the odds of having an innovation that targets external stakeholders are 2 times higher for women compared with men. See Long and Freese (2006) for a more thorough discussion. Furthermore, the assumptions of homoskedasticity and normality were not violated.

Results

The results are presented in Tables 2 and 3 (see online supplementary documents). Each table presents the odds ratios (see above for the interpretation of these values), standard errors, p-values, confidence intervals and a visualization of the significance level (asterisk). To gain an insight into our first research interest on the relation between the target groups (internal/external) of innovations and the origin of the idea for said innovation, we turn to Table 2. Consistent with H1, the most important result is that ideas from external stakeholders are more likely to be related to external innovations (compared with ideas from internal stakeholders). Ideas from external stakeholders are 3.7 times more likely to relate to external innovations (with a p-value of 0.000).

In order to address our second research interest—that is, how ideas from different types of internal stakeholders are related to external innovations—Table 3 expands the detail regarding the types of internal stakeholders. A distinction is made between “higher level management”, “your leadership group”, “staff at levels below your leadership group” and “other branch of your department or agency”. The findings show that when innovation ideas originate from the higher echelons (upper management, the leadership group), the likelihood increases that innovations target the external environment. When ideas originate from staff below the leadership group or from other branches of the department or agency, innovations are more likely to target internal stakeholders. These findings (predictive margins) are visualized in Figure 2 (see online supplementary documents). Note that respondents could answer with other branches of the department or agency when the idea did not come from their own workgroup and he/she therefore did not have information regarding the hierarchical level. As this category will include all hierarchical levels, this also explains the large confidence interval for this category.

Discussion and conclusion

This study sought to contribute to public sector innovation scholarship by addressing the relation between the stakeholder groups at whom innovations are targeted (and from whom support for the innovation needs to be found) and the reliance on ideas from different types of stakeholders.

Our findings are suggestive of two main conclusions. First, our results demonstrate that external innovations are more likely to be built on ideas originating from external stakeholders, and internal innovations on those from internal stakeholders. This finding is in line with theoretical expectations that innovations benefit from ideas of those stakeholders that will be primarily affected by the innovation (Cinar et al., 2019; Godenhjelm and Johanson, 2018). Both the innovating organizations and the affected stakeholders benefit from the observed connection between the output of innovations (where innovations end up) and the input of innovations (where innovations originate from). For innovating organizations, the involvement of affected stakeholders in the early phases of idea generation has substantive (e.g. absorbing local knowledge) and symbolic (e.g. gaining support, avoid public opposition) benefits (Bommert, 2010; de Vries et al., 2018). For stakeholders, early involvement maximizes the likelihood that the developed innovations are in line with their needs and expectations.

Second, two opposing lines of argument were presented for the relation between targets of innovation and ideas from internal stakeholders (managerial vs. non-managerial employees). Our results indicate that the ideas of managers and non-managerial employees contribute differently to internal and external innovations. As external innovations have direct consequences for the organization's legitimacy and strategic positioning, and securing these aspects is an essential managerial activity (Brunsson, 2002; Danneels, 2002; Karlsson and Skålén, 2015; Thompson, 1967; Un and Asakawa, 2015), the ideas of managers are influenced by the perceptions and assumptions of external stakeholders and, as such, contribute to a larger extent to external innovations than to internal innovations. In contrast, non-managerial employees have an intimate understanding of internal processes and routines, and their perceptions and knowledge bases are shaped by operational knowledge (Hartley and Rashman, 2018; Thompson, 1967; Un and Asakawa, 2015). This makes their ideas exceptionally beneficial for internal innovations, which explains why their ideas contribute to a larger extent to these innovations.

These findings contribute to public sector innovation scholarship that used a stakeholder perspective, e.g. to examine stakeholder-specific perceptions about innovations (de Vries et al., 2018) or how stakeholder involvement impacts the degree of innovation (Godenhjelm and Johanson, 2018). More broadly, this study contributes to the innovation literature, which widely recognizes the contextual nature of innovation (Amabile, 1988; Damanpour and Schneider, 2008; De Vries et al., 2015; Osborne, 1998; Rogers, 2003), and in which empirical studies have indicated how stakeholders in the implementation context affect public service innovation (Osborne, 1998; Verhoest et al., 2007).Yet it remained unclear how the stakeholders in implementation context influence the innovation process. From the premise that idea-generation and idea-implementation phases are closely connected with each other (Anderson et al., 2014; Van de Ven, 1986), we conjectured that the stakeholders present in the implementation context might affect the ideas used in the idea-generation phase.

In terms of practical implications, our study adheres to a belief that great ideas will hardly ever settle into effective innovations as long as the implementation context does not accommodate their realization. Our findings show that innovators should be aware of the perceptions of the stakeholders affected by the innovation, and properly sense, capture and translate the ideas of those stakeholders in the innovation process. Innovations are, at least partially, socially constructed, and should not only be functional but also match the social reality of the implementation context. Innovation is not the enterprise of a single entrepreneur. Organizations are ever directing their efforts at mobilizing the creative potential of all its human capital (different types of employees, citizens and stakeholders) to contribute big new ideas and help drive the organization and, ultimately, society forward.

By theorizing and testing how stakeholders affect innovation in public sector organizations, this study contributes to a previously established lack of theorization in the public sector innovation literature (Clausen et al., 2020; De Vries et al., 2015). Furthermore, the field has been characterized by a dominance of qualitative methodological approaches. This study relied on a large-scale employee census from the Australian Public Service which allowed us to test the expected effects on a sample of respondents in a variety of organizational settings.

Our study is not without limitations. First, our approach is limited in terms of directly studying the mechanisms that connect our main independent (stakeholder ideas) and dependent (innovation targets) variables. Future studies can use methods that are better suited to studying these mechanisms (e.g. process tracing, experimental methods). Such approaches may also do more justice to the potential heterogeneity in the group of external stakeholders. This study sought to uncover a common and broadly valid mechanism, that is, innovations are more likely to build on respective stakeholder ideas when they are targeted at these stakeholders. Yet it is very possible that this basic argument can be more refined when taking into account potential differences between innovation processes that primarily involve citizens, businesses, other government organizations, etc. (see Cinar et al., 2019 for an excellent overview of the respective barriers). Furthermore, we recognize that our focus was explanatory rather than normative: the present study's strengths lie in demonstrating the link between the inclusion of certain stakeholder ideas and the innovation's implementation context. Future studies can take a more critical stance on the political nature of stakeholder involvement and potential democratic implications.

Second, our study probed respondents to consider the most significant innovation in the workgroup. While this forced respondents to be focused in their answers, readers should be aware that our study did not include all innovations that were developed in each work setting.

Third, future research may tackle some potential issues with external validity. To start, while we have formulated several motivations for relying on respondents from the SES level, and we see no reason to expect any consistent bias in, e.g., overstating their influence on particular types of innovation, it would also be useful to include operational staff perceptions. Concerning the Australian context, the APS is well suited to examining the theorized effects in a setting that has long recognized the critical importance of pursuing innovation (Demircioglu, 2019). We believe that the underlying mechanism of interest—that is, innovations will more likely be built on ideas from particular stakeholders in the early phases of innovation processes depending on the target context for substantive knowledge-related and symbolic legitimacy-related considerations—will probably hold for other governments. Nonetheless, the APS does come with a particular Anglo-Saxon political-administrative culture which may set it apart from other administrative cultures, and we therefore urge future studies to apply our framework on non-Anglo-Saxon cases. In addition, our empirical focus was restricted to public sector organizations. While believe that the mechanism is sufficiently generic to also hold in a for-profit context, future studies could confirm this while taking into account dynamics that are more specific to the private sector (such as the role of market competition).

Supplemental Material

sj-docx-1-ras-10.1177_00208523211043704 - Supplemental material for A stakeholder perspective on public sector innovation: Linking the target groups of innovations to the inclusion of stakeholder ideas

Supplemental material, sj-docx-1-ras-10.1177_00208523211043704 for A stakeholder perspective on public sector innovation: Linking the target groups of innovations to the inclusion of stakeholder ideas by Jan Boon, Jan Wynen and Chesney Callens in International Review of Administrative Sciences