Key account management and value co-creation in multi-stakeholder ecosystems. A “ market access ” mix

Purpose – The present study aims to contribute to the growing stream of literature about the network perspective of value co-creation via key account management (KAM) by exploring how ﬁ rms, in complex industrial markets, use key account strategies to create value, not only for buyers and sellers of industrial products/services but also, more widely, for larger ecosystems of stakeholders. The research question this paper seeks to address is how the KAM approach promotes value co-creation in multi-stakeholder ecosystem. Design/methodology/approach – To answer this research question, this study uses a qualitative research approach based on data triangulation. This study focuses on the market access (MA) strategies implemented by a multinational UK-based pharmaceutical company within the Italian multi-stakeholder health-care ecosystem over several years. Findings – The results show that KAM in complex networks acts as a catalyst for value creation, through multiple interactions with different actors and an ad hoc con ﬁ guration of ﬁ ve strategic levers: product performance, economic impact, institutional relationships, commercial organization and communication. These levers are able to unlock the appropriate value drivers and form a speci ﬁ c “ market access mix ” implemented by the ﬁ rm to both promote the adoption of the ﬁ rm ’ s products and generate value for all market stakeholders. Originality/value – The study offers an innovative and comprehensive evidence-based model for designing speci ﬁ c MA strategies aimed at co-creating value within multi-stakeholder ecosystems. The proposed MA mix outlines the fact that knowledge, relationships and innovation are not unique factors that can be leveraged by stakeholders to co-create value.


Introduction
For many years key account management (KAM) has been a central subject in the industrial marketing literature (Workman et al., 2003;Zupancic, 2008;Ryals and Davies, 2013). KAM is "an approach adopted by companies aimed at building a portfolio of loyal accounts by offering them, on a continuing basis, a product/ service package tailored to their individual needs" (McDonald et al., 1997). The key goal of this approach is building long-term relationships and stable networks with a selected number of business clients (Ojasalo, 2001). Value co-creation, "the process by which mutual value is expanded together" (Ramaswamy, 2011, p. 195), is one of the main competencies by which KAM teams create value for their own firm (Hakanen, 2014;Kumar et al., 2019) and achieve a network-oriented result.
The network view of KAM, in which customers are conceptualized as nodes in a web of interlinked organizations, is becoming the dominant perspective in the industrial marketing literature (Kumar et al., 2019). However, the early ideas of the network approach in KAM goes back to Shapiro et al. (1984). Georges and Eggert (2003) investigated how KAM can create value for the KA and the supplier company. Using the network perspective, Ivens et al. (2016) analysed how KAM created value for both the focal organization as well as the KA organization. However, the industrial networks of firms does not only consist of buyers and suppliers: sellers can also use KAM to interact with a variety of third-party actors (e.g. government agencies or scientific communities) which are as follows: extremely relevant for the adoption of the industrial product and for the resulting value creation process but; and are not necessarily and directly involved in its purchase and use.
The notion, coming from the value co-creation theory, of multi-stakeholder ecosystems captures such variety (Gyrd-Jones and Kornum, 2013;Pera et al., 2016). Within these heterogeneous networks "value is co-created by the complex interaction of a network of stakeholders each holding specific and individual identities" (Pera et al., 2016, p. 4033).
Despite the increasing interest in the "network view" of KAM (Kumar et al., 2019), prior research has not investigated how such an approach can support value co-creation in multistakeholder ecosystems. Additional knowledge about such an issue would be extremely relevant in developing, for instance, specific theoretical frameworks supporting industrial managers in generating and sharing effectively co-created value within such heterogeneous and complex networks of stakeholders. The present article contributes to the growing stream of literature about the network perspective of value co-creation via KAM (Guesalaga et al., 2018;Hakanen, 2014;Pilon and Hadjielias, 2017;Kumar et al., 2019) by exploring how firms in complex industrial markets use KAM to create value, not only for buyers and sellers but also, more widely, for a larger and more heterogeneous network of industry stakeholders. The research question this article thus addresses is how the KAM approach promotes value co-creation in multi-stakeholder ecosystems.
The article reports the case of MA (Smith, 2012;Schiavone and Simoni, 2019), and a set of KAM practices and programs for value creation in the health-care industry, in a complex multi-stakeholder business ecosystem (Pinho et al., 2014;Schiavone and Simoni, 2019). The results of our study show value co-creation via MA is obtained via the use of five ad hoc "levers": product performance; economic impact; institutional relationship; commercial organization; and communication which facilitate and promote the introduction, adoption and use of the product in the market and co-create value among the different stakeholders involved in the process. The article is organized as follows. After this introduction, Section 2 reviews the theoretical background of the study. Section 3 describes the research method used to investigate the case, with the findings reported in Section 4. Section 5 discusses the findings and the main implications. The conclusions are presented in Section 6.

Literature review
2.1 A network view of value co-creation via key account management Existing studies highlight the crucial importance of KAM for firms involved in industrial markets (Workman et al., 2003;Guesalaga et al., 2018;Kumar et al., 2019). Through KAM, firms plan and implement specific services by which they can, for example, increase information quality and meet buyers' needs to establish and maintain long-term business relationships (Guesalaga and Johnston, 2010;Tzempelikos and Gounaris, 2015). KAM includes a series of key resources such as information systems with customers, managers, organizational structures and several intangible resources such as market knowledge and buyer orientation (Guesalaga et al., 2018). KAM value creation is related to specific behaviors (Ivens and Pardo, 2007), such as commitment in relationships and support by KAM teams for creating and transferring knowledge among various actors. The performance of KAM also depends on various intra-organizational characteristics, such as top management involvement, the intensity and proactiveness of the activities and access to marketing and sales resources (Workman et al., 2003). To improve KAM performance, firms must pay attention to both intermediate and final customers (Abratt and Kelly, 2002;Guenzi and Storbacka, 2012;Kumar et al., 2019). Kumar et al. (2019) highlighted the growing attention given by scholars to the network view of KAM. Maintaining long-term relationships through KAM processes means identifying, at the right moment, how the different actors act, understanding needs at different levels and defining how to interact with several stakeholders in business-to-business relationships (Ojasalo, 2001;Vasconcelos and Ramirez, 2011). The need for networking by industrial companies thus pushes KAM teams to minimize the use of predatory strategic behaviors (Deshpandé and Farley, 2002;Ivens et al., 2016). As affirmed by Georges and Eggert (2003), KA managers effectively improve the overall value proposition. Niersbach (2016) discussed interpersonal skills, intra-organizational alignment and commitment and trust within such networks. Previous studies analyzed the structural practices of information and planning systems to satisfy the needs of network partners (Strömsten and Waluszewski, 2012). On the other hand, Grant and McLeod (2007) underlined the importance of supply chain networks in terms of sharing relationships among suppliers and other different actors at various levels. Other scholars attest to the fundamental role of contractual governance in multiple networked relationships (Ivens and Pardo, 2008).
A crucial dimension of the network perspective of KAM is value co-creation. The service marketing literature acknowledges that co-creation networks improve service systems through better matching resources, processes and outcomes (Gummesson and Mele, 2010). Prior research has extensively analyzed how buyers and sellers, tied together via specific industrial transactions, use KAM to co-create value (Gosselin and Bauwen, 2006;Guesalaga et al., 2018;Kumar et al., 2019). Value co-creation competence enables KAM to create value for the firm within industrial networks (Sullivan et al., 2012;Hakanen, 2014). Over time, firms establish and reinforce their business network to achieve value co-creation, for example, by co-developing collaborative solutions which improve the quality and performance of their business relationship (Hakanen, 2014;Kumar et al., 2019). KAM teams play the fundamental role of knowledge integrators for the development of integrated solutions with customers (Hakanen, 2014). Knowledge and resource integration are some of the main practices through which KAM can generate a value network, which "is driven by value creation that is evaluated through value co-creation" (Kumar et al., 2019, p. 287).

Value co-creation in multi-stakeholder ecosystems
A multiple stakeholder ecosystem offers a wider view of the traditional one-to-one business relationship because this conceptualization "encapsulates both the network nature of these relationships and the complex set of subcultures that make up this ecosystem. An ecosystem is normally used to refer to systemic interactions within biological environments consisting of both physical and biological components (Gyrd-Jones andKornum, 2013, p. 1484). An ecosystem provides the context within which interdependent actors interact, adapting their respective strategies and improving their practices (Frow and Payne, 2011). Companies acting in a multi-stakeholder context should take into account the effects of such multiple and heterogeneous relationships, and the network interdependencies that may exist between different groups of stakeholders (Reypens et al., 2016;Lehtinen et al., 2018). For example, the various stakeholders within a business ecosystem dealing with a specific problem tend to gather information from different sources, activate learning dynamics, try to jointly address any conflict between participants and establish reciprocal cooperation (Roloff, 2008).
Value co-creation happens mainly because no single actor owns all the required competence and resources to achieve a competitive advantage (Vargo and Lusch, 2010). Value co-creation in complex systems is enabled not only by the dyadic exchange of resources and interactions between customer and service provider but also by multiple and intensive interactions among multiple types of stakeholders (Pinho et al., 2014). Such co-creation-oriented interactions regularly take place because the customer in multi-stakeholder ecosystems is usually not a passive user but, instead, is an active actor within the knowledge exchange process (Vargo and Lusch, 2008;Singaraju et al., 2016). Prior research has found that different groups of stakeholders co-existing in the same ecosystem decide to co-create value for reputational, relational and experimentation motives (Pera et al., 2016).
Prior research has paid particular attention to the intersection between value co-creation and innovation in multiple-stakeholder ecosystems. Referring to value cocreation in industrial innovation networks, Reypens et al. (2016) argue that this process (at a network level), along with value capture at a stakeholder level, is a crucial stage for the value leveraging process. The authors also identify various dimensions (innovation, knowledge and relationships) and the design of a "value space" within which the stakeholders can position themselves. Gyrd-Jones and Kornum (2013) found that value and culture complementarities were important drivers for achieving the acquisition of successful co-creation outcomes with multiple partners, despite the fact that stakeholders could also provide adjustments to the firm's practice and innovation strategy. In sum, such insights outline the fact that KAM practices by industrial firms in innovationdriven multiple-stakeholder ecosystems can be extremely relevant and play a central role in value co-creation.

Sub-research questions
The paper analyzes how KAM supports value co-creation in complex multi-stakeholder ecosystems. To better understand this analysis, it is worth considering three sub-research questions (Sub-RQ) based on the relationship between ecosystem complexity and KAM. Sub-RQ1 refers to the possibility of rethinking the role and goals of KAM. The presence, in multi-stakeholder ecosystems, of actors who influence the purchasing process without being directly involved in the purchase of the firm's products should lead to a broader view of the scope of KAM. KAM should be considered not only as an approach that directly affects value co-creation but also as a set of management practices that create and enable the necessary conditions for other stakeholders to create value. Accordingly, KAM programs should focus on promoting, organizing and supporting interconnected decision-making between multiple parties, which drives the adoption of the company's products or services.
Sub-RQ2 concerns the value drivers of value co-creation in complex ecosystems. The heterogeneity of stakeholders within such ecosystems should increase the variety of value drivers and, therefore, also of the value co-creation mechanisms that are relevant for all parties involved. Such variety should lead to the establishment of both value-creating relationships, with high levels of complementarity between the stakeholders (Gyrd-Jones and Kornum, 2013), and relationships that are less oriented to value co-creation. As a result, the design of KAM practices and programs should be much more sophisticated and complex for corporate top management. KAM should face the relevant challenge of understanding, in advance, which relationships within the ecosystem will lead to value co-creation and which will be less oriented to this outcome. Moreover, different value drivers may require different managerial approaches, in that they may increase the difficulty of this challenge.
Finally, Sub-RQ3 refers to the specificity and complexity of multi-stakeholder ecosystems. Indeed, the need to address various value co-creation processes, by leveraging heterogeneous value drivers that are relevant for stakeholders playing different roles, should make typical KAM levers less suitable for, and applicable to, all stakeholders. For example, in innovation-driven ecosystems, as suggested by Reypens et al. (2016), KAM should design new (or slightly revised) levers both to promote the co-creation of value with other innovators and completely support the value captured by other stakeholders. Such a condition obliges firms to redefine their KAM policy and activities and customize the usual KAM levers for its complex multi-stakeholder ecosystem.

Research method
To answer the research question, this study adopts a qualitative method using case study analysis (Yin, 1994(Yin, , 2017. This approach is consistent with the exploratory nature of our research which "focuses on understanding the dynamics present within single settings" (Eisenhardt, 1989, p. 534) and addresses a "how" research question. The lack of previous studies further strengthens the case study method choice (Strauss and Corbin, 1990;Mills et al., 2010).
We explored the case of MA to understand the role of pharmaceutical drugs in the Italian multi-stakeholder healthcare ecosystem. Market access (MA) is a particular KAM approach that uses a set of elements and activities to promote the adoption and diffusion of a new drug among different stakeholders in the health-care ecosystem (Toumi, 2017;Schiavone and Simoni, 2019).
The study was based on the triangulation of data. Multiple sources of information were considered including: (1) expert interviews; (2) direct observations; (3) official archival records; and (4) online documentation.
As the first source of information, we analyzed the market access strategies (MAs) implemented by a UK pharmaceutical company which had just entered the Italian market. This company had a "market access organizational unit" that coordinated actions at different levels of the company to support the diffusion of new drugs. Hence, an analysis of the market access strategies of this pharmaceutical company was appropriate to answer our research questions because it enabled us to better understand the multiple stakeholder relationships in the Italian health-care ecosystem and the kind of KAM approach that should be adopted to promote value co-creation in this context.
We presented a semi-structured questionnaire to five different types of MA managers (MAMs) in the Italian healthcare sector: 1 Head of Market Access (Southern Italy). 2 Regional Access Manager (RAM). 3 Field Access Manager (FAM). 4 Key Account Manager (KAM). 5 Product Specialist (PD).
We organized the in-depth interviews using a semi-structured questionnaire with open-ended questions about account management decisions and the specific use of market access practices at different levels (global, national, regional, local). During the interview, specific questions covered the conditions created by MA to enhance relationships and to promote the cocreation of value in the multi-stakeholder Italian health-care ecosystem. The average duration of the interviews was about 60 min. The interviews were audio-recorded and then transcribed verbatim.
After various rounds of open coding, we identified aggregate dimensions moving from first-order descriptive terms to second-order themes that aggregated dimensions (e.g. conceptual coding of market access strategies) (Gioia et al., 2013). Triangulation with other sources of evidence helped us to reinforce our analysis (Yin, 2009). We used additional primary information and secondary data to strengthen our interpretation of MA activities at different levels of the Italian health-care ecosystem. We followed a two-step approach.
First, through direct observations, we documented activities and various aspects of MA processes using our senses without having to depend on the actor's capacity to answer questions. Indeed, the MA practices were observed and monitored from 2015 to 2018 in accordance with the latest European Medicines Agency related document (2015), related to the application of EU marketing authorization for medicinal products for human use.
Second, we performed a desk search of secondary data online (e.g. reports, specialized trade press) and official archival records. We collected evidence from several official health-care information sources and from various groups of actors involved in MA processes (e.g. health-care organizations, public institutions, regions, patient associations). Triangulating direct observations with interviews, online documentations and the investigation of records retrieved from official pharmaceutical company documents, enabled us to clarify the typical value co-creation mechanisms within the Italian multi-stakeholder ecosystem and in particular, the specific value drivers for each stakeholder. Through this data, we assessed the impact of MA, as a form of KAM, in terms of the promotion of value co-creation via an effective exchange of information between institutions, pharmaceutical companies and public and private bodies at a global, national, regional and local level.

Results
Drawing on Gioia et al. (2013), we coded information in subsequent steps to identify the main aggregate dimensions of the study to tackle the three sub-research questions. Three dimensions emerged that are proposed and explained in the remainder of this article: 1 multiple stakeholder relationships; 2 value drivers in a multi-stakeholder ecosystem; and 3 MA levers. Figure 1 illustrates the data structure and coding process together with some examples of first-order terms.

Multiple stakeholder relationships in the Italian health-care ecosystem
In the Italian health sector, the regions have a large degree of decision-making autonomy. The public authorities not only regulate access to the market, but also participate in the supply chain by acquiring new devices or drugs for public institutions. Several different actors play a role in the process, which extends from the initial application for acceptance of a firm's new medical device or drug to commercialization of the product and its usage by doctors and patients: national health-care authorities, regional health-care authorities, local health-care organizations (the so-called ASL, usually managing one or more hospitals), hospitals, specialized operating units within hospitals, doctors and patients.
Enormous differences in the criteria adopted by northern and southern regions to assess health-care products and treatments were observed. Indeed, some regions had a regional health technology assessment (HTA) department, such as "Emilia Romagna" and "Veneto", whereas "Lombardia", which is geographically proximate, adopted a completely different methodology. Therefore, close relationships with regional offices were essential for companies. All pharmaceutical companies in the MA department had a territorial network of people, whose job was to demonstrate the therapeutic and economic value of their drugs to regions, departments and hospital pharmacists. This activity was particularly difficult and required mutual trust between all stakeholders, especially when the drug was new and, for the first time, the firm sought authorization for the drug to be included in a Regional Pharmaceutical Manual (PTOR). The main actors involved at different levels (global, national, regional) were the European Medicines Agency (EMA) and, in the Italian health sector, the Italian Medicines Agency (AIFA), the Ministry of Economic Development, the Ministry of Health, the Italian National Health Service and the regions.
Details on the types of actors involved and their role are reported in Figure 2.
These actors, together with other health organizations (e.g. payers, patients' associations), participated in value co-creation and had a different impact on the decision-making process related to the adoption and use of the product. As highlighted by the RAM, regional decision-makers have a large degree of decisionmaking autonomy, but relationships among different stakeholders in the regional business ecosystem played a key role in achieving the desired results. In the following section, we identify the factors that most affect value creation at the level of each stakeholder.

Value drivers in the multi-stakeholder Italian health-care ecosystem
Starting from the international dimension, we find primary activities for drug commercialization. In this context, the relationships between the company and the international regulatory agencies was fundamental to place pharmaceutical products in the market. As affirmed by the EMA-related document of 2015: [. . .] companies should inform the Agency in writing approximately 7 months in advance of their intended submission date. About the same time, a pre-submission meeting with the Agency's product team may be requested. It is strongly recommended that applicants take this opportunity to obtain procedural, regulatory and legal advice from the Agency These activities refer to the registration phases, with the presentation of exhaustive documentation of the therapeutic class. Another important phase is the Pricing & Reimbursement (P&R) negotiation. The submission of the negotiation request belongs to the company concerned. The aim is to establish the price and obtain the reimbursement of a drug after the release of the marketing authorization. This phase of the process is very long and complex (sometimes even very conflicting), given the opposing interests of the two actors involved and should be based on the criteria as risk/benefit ratio or cost/effectiveness ratio.
National activity is included in the global context. A majority of activities refer to the price negotiation phases with the national agency (in Italy: the AIFA) and inclusion in the registers. First, the company has to prepare a series of activities aimed at acquiring all the information that will support the introduction of the drug.
The RAM remarked on this aspect: Market Access structure refers mainly to the size of Pricing & Reimbursement in correlation with the relationship with the Regions, which have a greater decision-making power at the home of "devolution". The obstacles refer to not seizing the opportunities, as some drugs without a specific market access structure could never see the market and the best price category, also carrying out an analysis of costs and benefits.

Figure 1 Data structure and coding
At the regional level, most of the challenges that the pharmaceutical companies must face are related to the broad decentralization of decision-making (e.g. different groups of decision makers). First, it is necessary to apply for inclusion in the Regional Pharmaceutical Handbook (PTOR), as the company must relate to multiple health systems which have different characteristics. Naturally, at this stage, as well as at all levels, the company must develop dossiers to demonstrate the value of the product through clinical evidence and therapeutic outcomes, to build greater credibility for different groups of stakeholders who have a decision-making role in defining PTOR.
The last level is the local one, where the activity of the company must be focused on payers such as the ASL (Italian health local units), hospitals and so on. Pharmaceutical companies have to work on drafting files, collaboration programs and meetings with stakeholders. Table 1 provides details on the main drivers that create value for each of these stakeholders.

Market access levers
As described above, regulatory impositions result in more unstable and delicate buyer-seller interactions and, thus, companies are at risk of not being able to fully exploit the potential value that their products can create, especially when

Risk/benefit ratio
Evaluation of the clinical trial as the most fundamental tool for the evaluation of the benefit/risk profile of drugs Observation of ethical, practical and economic issues such as the recognition of value that not only concerns clinical outcomes, but also improvements in the quality of life of patients and the sustainability of the system National National agencies (e.g. AIFA) and firms

Financial impact Epidemiology
Evaluation of the dossier for obtaining reimbursement and access to hospital handbooks. Evaluation of financial outcomes for the national agencies following the adoption of drugs and vaccines. Budget impact model planning Regional Regional decision-makers (e.g. Campania) and firms

Epidemiology, Effectiveness
Organizational aspects Assessment of the drug profile for obtaining inclusion and access to the regional therapeutic handbooks (PTOR). Evaluation of effectiveness on clinical evidence and therapeutic outcomes, to build credibility and create value at regional level Local Local payers and firms Financial aspects Epidemiology Projects duration Evaluation of an innovative project to include the adoption of drugs in local therapeutic handbooks (PTO), to support primary care and prevention these products are radically new. At an organizational level, this prompts the industrial marketing department to create a specific unit and hire specialized human resources that are completely devoted to monitoring and managing regulatory affairs and the entire set of relationships with different players of the multi-stakeholder health-care ecosystem (e.g. public government, decision makers, national agencies and so on).
A majority of current pharmaceutical companies have a department dedicated to MA. According to an interview with the Head of Market Access about the roles covered by a MAM, MA is characterized by a particular need for economic and organization skills, rather than scientific skills: The market access manager (MAM) is a person who is paid by a pharmaceutical company and works half of his time in the company and half for the National Health Service. He is a sort of external technical consultant, a good prompter, who tries to fill those times and problems to ensure the best patient care. A person who has an economic background in this role. He also has a greater capacity than a subject that has a pharmaceutical background, as in every company there is pharmaceutical training but not economic training.
A fundamental MAM issue refers to the difficulty the departmental organization has in defining the different figures that operate within it, who have a different hierarchy depending on the account relationships with which they interact: In most pharmaceutical companies there is a hierarchical scheme with a Pricing & Reimbursement department at the top, which communicates with the Health Economics office and the Health Technology (HTA) team and follows a territorial group of SAMs which is often divided into RAMs (Regional Account Managers) and KAMs (Key Account Managers) in which the first interfaces with the regions while the interlocutors of the second are not regional decisionmakers but hospitals and local health services Accordingly, MA is essentially aimed at designing, implementing and managing value co-creation processes among the key stakeholders that operate in the business ecosystem, allowing the introduction and adoption of a firm's products at different levels.
To promote value co-creation, MA managers use different levers that address the value drivers of each stakeholder in the ecosystem. This addresses the clinical, economic and sustainability aspects of their products. Accordingly, we define the "market access mix" as the set of levers that the companies arrange and monitor to promote value co-creation in the multi-stakeholder health-care ecosystem. The levers of the MA mix are described in Figure 3.

Product performance
Product performance demonstrates the positive results obtained from clinical research which includes the better use of medicines and appropriate use of behavioral and organizational interventions (e.g. health therapies). In the initial phase, this multidimensional lever can be used at the global level with reference to specific drugs or processes (e.g. clinical procedures) or services (e.g. data analytics) with specific global actors (e.g. EMA). For example, real-world evidence can observe data obtained outside the health-care ecosystem of randomized controlled trials and implemented during routine clinical practice. The presence of many actors involved in the health ecosystem is, per se, a source of complexity for companies launching new pharmaceutical products. Indeed, for a new product to be adopted and diffuse, it must be positively evaluated by all stakeholders (e.g. international agencies, decision-makers, regions, payers, physicians, scientific societies, patients and so on) according to their respective evaluation criteria. As a result, the same innovation must meet many very different expectations. For example, after the evaluation of outcomes research, the HTA report presents a valid instrument of MA that is able to evaluate a health technology that delivers results at the national level (e.g. AIFA) to demonstrate the value of a specific innovative product (e.g. a new drug) by a KA manager. Finally, analytics and real-world evidence allows for an analysis of the databases for evaluation of the health cost of diseases and drugs. These are tools to support governance for the implementation of innovative projects at a local level (e.g. a hospital). Thus, in the health care market, the response time is a direct function of both the effort put into promoting the innovation and the specific contingencies of the adopting organization.

Economic impact
The price measures the economic impact in relation to the savings in expenditure to the health-care ecosystem that the new product can induce. These savings are derived not only from the lower price the product may have but also, and especially, by the reduced indirect costs. The lower cost may be derived from diminished therapeutic activities or reduced side effects obtained from the adoption of the innovation. A multidimensional MA strategy prepares the P&R dossier, which is a tool used to obtain reimbursement and access to regional hospital prescriptions. At this stage, MAMs have to monitor national (e.g. AIFA) and regional (e.g. the Campania region) regulations. Pharmaceutical companies consider the budget impact, which helps to evaluate and demonstrate the financial outcomes by a KA manager for the national stakeholder and other actors following the adoption of drugs (regions and local hospitals).

Institutional relationship
A multidimensional MA structure establishes and maintains continuous relationships over time with the key stakeholders in the market to both promote the new product adoption and diffusion, and to follow progress (e.g. marketing authorization obtained from the AIFA). As noted by the managers interviewed, each actor is a stakeholder representing a specific mix of different interests and objectives: the efficiency of the entire healthcare system, the effectiveness of public health policies (e.g. public/government affairs), the quality of therapeutic protocols, the financial equilibrium of regional healthcare systems, of the local system or of the single hospital/ operating structure, the quality of life of patients. In the health context, this challenging task is achieved with the concurrent efforts of the different actors. KA managers must be able to map the stakeholders involved in the decision-making process. They support patient needs and establish a long-term relationship with several payers and decision-makers. That is, each actor (e.g. the Campania region) which decides to adopt the drug becomes an active promoter of the product, thus sharing information and co-creating value for other players (e.g. payer engagement).

Commercial organization
This lever aims to create an appropriate commercial strategy that complements the information needed to achieve the formal approvals and long-term projects of different authorities in a multidimensional perspective. For example, a MAM can interface with AIFA's Committee on Prices and Reimbursement to fix the drug price and the conditions of reimbursement.
Furthermore, pharmaceutical companies aim to obtain a higher degree of homogeneity among actors regarding their knowledge about the new products' features and their evaluation. To increase the relevance and reliability of the information provided to all actors, the commercial organization invariably has a MA unit. Because of this specificity, a KA manager involved in the MA unit of a pharmaceutical company has a territorial-based organization in terms of nations and possibly regions and local contexts.

Communication
Companies communicate their products following cultural and sustainable principles at each level (global, national, regional, and local). For example, patient centricity and ethical duty are the key elements of any effective innovation-related communication used by a MAM (es. KA manager) with different stakeholders to obtain marketing authorization. With reference to these conditions, the value proposition offered by the innovative tool of MA in the current health-care ecosystem is recognized.
Thus, a multilevel MA strategy (the MA mix) represents a useful approach to involve the different groups of actors at each dimension (e.g. international agencies, decision makers, regions, payers, physicians, scientific societies, patients and so on). After the preclinical and clinical test, pharmaceutical companies must obtain marketing authorization from the national agency. In this step, the account managers should know which levers are best used in which positions (e.g. product performance and communication). Furthermore, they negotiate the price for the drug's classification according to the reimbursement regime (the lever of economic impact). The RAM affirmed that regional decision makers have a large degree of decision-making autonomy, thus institutional relationships and commercial organization play a fundamental role in achieving the intended results, as seen below: The Regional Health Service is two sides of the same coin, on the one hand there is drug policy and on the other there is the contracting authority. The regions insert this into the PTOR but do not buy it, the contracting entity buys it. In the case of the Campania region it is called "Soresa Subject Aggregator". They enter a race [. . .] in the case of Puglia there are 15 aggregators, in Sicily one, in Calabria one. Thus, there is a person who decides that the drug is good and one that you buy it and often do not have relationships. Market access is the solution Thus, the MA mix places itself at the center, between drug policies and the contracting entity, to speed up this process and synchronize their commitment.

Discussion and implications
First, the findings of the present article contribute to the recent literature on network value co-creation (Abratt and Kelly, 2002;Guenzi and Storbacka, 2015;Kumar et al., 2019), demonstrating that, as the complexity of the network increases, due to purchasing decision processes that require the interdependent concurrence of several actors playing different roles, KAM must accordingly be rethought. In particular, our results show that KAM in complex networks, on one hand, as a direct value creation mechanism, acts through multiple interactions with different players, consistent with the previous literature (Deshpandé and Farley, 2002;Ojasalo, 2001;Vasconcelos and Ramirez, 2011). On the other hand, KAM, as an enabler of conditions for value creation, acts through levers that at different levels are able to unlock the appropriate value drivers. This suggests a broadened perspective on KAM's studies that go beyond the logic of network value co-creation to embrace the wider idea of business ecosystem design.
Similarly, the five MA levers outline the fact that knowledge, relations and innovation are not the only possible dimensions that stakeholders can leverage to co-create value (Reypens et al., 2016). Our study shows, for example, that other elements, such as interest for tangible outputs (e.g. economic costs, clinical care and so on), are equally important in addressing the phenomenon.
Finally, we contribute to the recent literature on MA (Smith, 2012;Schiavone and Simoni, 2019), providing a multidimensional interpretative key to MA, which allows for a correct and adequate understanding of the sense and of the conceptual and thematic articulation of the entire study. The findings demonstrate that MA can promote value co-creation at different geographic levels, referring to different groups of stakeholders (e.g. international agencies, decision makers, payers, scientific societies, patients and so on). By using a MA mix, each stakeholder can access information simultaneously in the interdependent processes of the different innovation adopters (e.g. national agencies or regions). Furthermore, MA implemented in the multi-stakeholder health-care ecosystem (Pilon and Hadjielias, 2017) is intended as a set of dimensions that promote the diffusion and adoption of drug value by the different subjects. This implies a prior understanding of the institutional context, making use of several professional actors positioned at different levels.

Managerial implications
The study results offer at least two relevant managerial implications for firms that operate in complex multistakeholder ecosystems. First, a suitable professional profile of KAM in such contexts is not likely to be found or easily taught. Indeed, these managers should possess a diversified set of managerial, economic, social and interpersonal skills and capabilities to effectively co-create value within such a complex business ecosystem. Thus, companies need to plan and implement specific ad hoc recruitment processes and employee training.
Second, KAM activities are required to implement holistic and multidimensional approaches and strategies to co-create value with (and among) the different actors in a multistakeholder perspective. Only through such approaches can organizations enable the creation of value at different levels and for different stakeholders. Accordingly, firms operating in the healthcare sector can also redefine their MA activities in broader terms as KAM of a complex multi-stakeholder business ecosystem.

Theoretical implications
As clearly emerges from our explorative study in healthcare, an expanded role for KAM is not simply an opportunity for firms but a necessity to address and manage interdependencies among actors that, according to their specific role, may promote, sustain and facilitate the adoption of a firm's products, but may also inhibit, hinder or oppose such adoption (Schiavone and Simoni, 2019).
The findings of the present article also offer theoretical implications to the growing stream of literature about cocreation in multiple-stakeholder ecosystems (Reypens et al., 2016;Gyrd-Jones and Kornum, 2013) by better detailing the heterogeneity of actors, their different backgrounds and the subsequent need for specialization of KAM practices in this context. The five MA levers by which healthcare organizations promote value co-creation extend the set of theoretical mechanisms and outcomes already known in the literature. Our results show that not only complementarity between values and stakeholder culture is important for the co-creation of value (Gyrd-Jones and Kornum, 2013) but also tangible elements, such as the economic impact of health products on the ecosystem or adherence to industry standards, are crucial, both to achieve bilateral co-creation of value and to enable other stakeholders (not directly involved in business relationships) to capture value.

Conclusions
The paper analyses how KAM promotes and supports co-creation of value in multi-stakeholder ecosystems, exploring the case of MA in the health sector. The study shows that the value system of healthcare is complex in nature and that to fully understand KAM's role in such a complex system, it is necessary to rethink and extend the scope, objectives and levers of KAM.
The results of the study contribute to several literature streams: co-creation of network value, KAM in complex ecosystems and MA management. The paper also has managerial implications for companies operating in contexts where multiple stakeholders contribute to the decision to adopt a product, having the power to trigger or inhibit the adoption process.
However, our study has several limitations. First, this research paper highlights the case of a pharmaceutical company offering pertinent insights and evidence in business and management studies. However, the complexity of the healthcare ecosystem (e.g. decentralization of decision-making power) might create various issues about the value of case study when this method is used to tackle the research question. A second limitation regards the small number of actors interviewed. This issue could lead to a limited comprehension of the phenomenon under analysis.
Although the KAM model proposed in our work is the result of a case study based on a specific sector, the theoretical implications of the model can be extended to other sectors and contexts in which multi-stakeholder ecosystems are particularly complex (e.g. aerospace, telecommunication equipment, investment finance, digital security). However, this generalization may require further research that should point in two different directions. First, some features of our KAM model could be idiosyncratic to the analyzed case, i.e. MA in the Italian health-care system. Therefore, future research could assess whether in other countries the MA shows the characteristics and critical aspects necessary to support our model. Secondly, the presence in other industries of a different distribution of stakeholders' decision-making power may make it necessary to investigate whether the same relationship logic, value drivers and KAM levers identified by this study can be fully applied or should rather be adapted.