Perceived success of hybrid microorganizations in a contested category

ABSTRACT The organizational literature privileges objective performance indicators often selected by researchers. There is scarce research focusing on legitimacy-challenged hybrid and microorganizations and on perceived success under exigent conditions. To address this scarcity, this study, conducted among complementary and alternative medicine (CAM) microorganizations, explores success as a subjective measure originating from managers’ perceptions. For the purpose, it integrates cognitive mapping and multiple criteria decision analysis (MCDA)—a methodological contribution to construct a subjective-success framework that can be helpful for contested hybrid microorganizations. Seven factors emerged, of which human capital is recognized as critical, while external factors are considered unimportant.


Microorganizations;
subjective perceived success; hybrid organizations; complementary and alternative medicine (CAM); cognitive mapping; multiple criteria decision analysis (MCDA) The liability of smallness is a well-known phenomenon (Bruderl & Schussler, 1990). Inserting "micro" and "hybrid" prefixes to organizations adds an extra layer of precariousness. Most microorganizations have extremely limited resources (Gorgievski et al., 2011;Lonial & Carter, 2015) and, when present, legitimacy hurdles pile on even more constraints (Ruffo et al., 2018;Wang et al., 2017). Legitimacy deficiency leads to stakeholders questioning the very existence of the organization ensuring limited customers, financing sources, and community support. As a result, small organizations from contested market categories are stuck in a microframework, which turns into a constant struggle for survival (Galvin et al., 2004;Ruffo et al., 2018). Hybrid microorganizations are confined by lack of both resources and legitimacy. They have several goals and merge diverse institutional logics leading to legitimacy ambivalence (Battilana & Lee, 2014;Doherty et al., 2014). The combination of multiple organizational forms departs from socially accepted templates leading to unique obstacles for hybrid CONTACT Marta Bicho marta_liliana_bicho@iscte-iul.pt IPAM Lisboa and Instituto Universitário de Lisboa (ISCTE-IUL), Business Research Unit (BRU-IUL), Lisboa, Portugal organizations (Battilana & Lee, 2014;Hahn & Ince, 2016). Consequently, hybrid microorganizations constitute the most disadvantaged and contested group across the spectrum of organizations. As such, they have been overlooked by researchers, too.
Further, there is no consensus about the appropriate assessment of organizational success (Gorgievski et al., 2011;Maltz et al., 2003;Reijonen & Komppula, 2007;Singh et al., 2016). Past research tends to use either objective or subjective indicators, as opposed to bringing all indicators together (Staniewski, 2016). Moreover, most studies focus on organizational performance or success from the researcher's point of view (e.g., Amato et al., 2017;Gunasekaran et al., 2017;Singh et al., 2016;Staniewski, 2016). In contrast, only a small number of studies examine perceived success from the managers' point of view (e.g., Gorgievski et al., 2011;Wach et al., 2016). The literature is even less vocal when it comes to microbusinesses (Gherhes et al., 2016).
This study targets the gap of knowledge of helpful frameworks for disadvantaged firms by investigating hybrid microorganizations' perceived path to success from the managers' point of view in a category that lacks legitimacycomplementary and alternative medicine (CAM). It may be particularly taxing for hybrid microenterprises to assess the fit between internal capabilities and external market conditions at two levels (commercial and noncommercial) and yet be consistent with their dual purpose (Battilana & Lee, 2014;Reijonen & Komppula, 2007;Sharma et al., 1990). Defining a set of success factors leading to survival and longevity would be challenging for managers of hybrid microorganizations in low-legitimacy market categories (Dobrev & Gotsopoulos, 2010;Ruffo et al., 2018). Because they face clashing demands-the reconciliation of multiple and often conflicting objectives is an arduous task-individual managers may not know whether they are succeeding.
This research gives voice to marginalized decision-makers' notion of success by using techniques that capture the criteria's complexity. The perceived-success-factors framework is built by integrating cognitive mapping and a multiple criteria decision analysis (MCDA) approach. The joint use of these constructivist tools has been shown to be helpful to decision-makers to think through and discuss multifarious problems, guiding them in finding the best options in complex environments (Belton & Stewart, 2002;Ferreira et al., 2015). The decision to employ cognitive mapping and MCDA is motivated by the intricate dual-identity nature of hybrid microorganizations operating in contested categories.
The study's contribution is twofold. First, it sheds light on the most overlooked type in organizational research-hybrid microorganizations in contested markets-and responds to Battilana and Lee's (2014) call for the construction of hybrid performance measures as one of the most prominent challenges to researchers in the field. Second, it demonstrates how an MCDA approach can be useful in exploring perceived success drivers by combining it with cognitive mapping. The study creates a unique map of key success factors that can be of particular value to managers of contested hybrid microorganizations. The features of the method (e.g., interactivity and learning-oriented processes) allowed the participating decision-makers to debate and structure the evaluation framework in an open atmosphere, identifying fundamental points of view (FPVs). The joint use of these techniques further contributes to practice through the framework's real-world application. The participants were managers with limited resources who found the evaluations of their organizations, grounded in the constructed framework for success, very helpful and insightful.
The remainder of this article is structured as follows: The next section outlines the theoretical background; the third section presents the context of the study; the forth section describes the methods and the data, followed by the main findings and a discussion of the findings. Finally, conclusions, limitations, and future research are presented in the last section.

Organizational success
Organizational success is tracked by operational criteria and financial measures. Operational criteria are nonfinancial indicators that might lead to financial performance and often require qualitative assessment by managers (Kotey & Meredith, 1997;Maltz et al., 2003;Reijonen & Komppula, 2007;Singh et al., 2016). Financial measures are at the core of the organizational effectiveness, reflect the fulfillment of the economic goals of the organization (Maltz et al., 2003;Reijonen & Komppula, 2007;Singh et al., 2016), and are the most frequently used measures by managers (Gorgievski et al., 2011;Kotey & Meredith, 1997;Lonial & Carter, 2015).
Several authors have discussed the difficulty of adequately assessing how to evaluate success (Gorgievski et al., 2011;Reijonen & Komppula, 2007;Singh et al., 2016). The choice of success factors is complex as they may be based on the organizations' goals and objectives (Kotlar et al., 2018;Singh et al., 2016), context (Kotlar et al., 2018), and characteristics of the managers (Bouchikhi, 1993;Gorgievski et al., 2011;Kotlar et al., 2018). Nevertheless, studies have mainly focused on a single performance goal-profitability (Kotlar et al., 2018). Some researchers have looked at success criteria and measures from their own perspective. For instance, Maltz et al. (2003) suggest five dimensions for assessing organizational success: financial measures, customer/market measures, process, people development, and preparing for the future measures; while Gorgievski et al. (2011) list 10 criteria of business owners' success: profitability and growth, innovation, firm survival/continuity, contributing to society, personal satisfaction, satisfied stakeholders, good balance between work and private life, public recognition, and utility or usefulness.
Ultimately, success is a subjective measure-managers have their own perceptions of the meaning of success, but research on subjective organizational success is rare (Gorgievski et al., 2011;Reijonen & Komppula, 2007;Rogoff et al., 2004;Wach et al., 2016). One way of defining success is "generating an effective firm in the long term" (Bouchikhi, 1993, p. 561). To adequately address the demands of various stakeholders (Maltz et al., 2003;Rogoff et al., 2004), the definition can include both subjective and objective elements (Staniewski & Awruk, 2019, p. 434). Wach et al. (2016) define managers' perception of organizational success "as the individual understanding and assessment of the achievement of criteria that are personally important" to him or her (p. 1099). This is especially relevant for managers of microorganizations who may have different notions of success and who may not be interested in growth (Gorgievski et al., 2011;Reijonen & Komppula, 2007;Staniewski, 2016). For example, a manager can see its organization as successful in achieving a given purpose, while from a financial point of view, the firm might not be that successful (Gorgievski et al., 2011;Staniewski, 2016). Subjective measures have been used in past research, mostly applied to performance in organizational (Singh et al., 2016) and entrepreneurial contexts (Wach et al., 2016;Wang et al., 2017). The current study is closer to the subjective measures used to access "entrepreneurial subjective success often examined through the self-reporting of an entrepreneur's satisfaction with the business's performance, growth and status" (Staniewski and Awruk 2019, p. 434). Although objective performance measures are of extreme importance as they identify dimensions that might not be obvious to managers, "organizational success" as a subjective framework, holistically capturing managers' perceived success factors, is also critical as most managerial decisions are based on it rather than on a comprehensive list of objective indicators.
If microorganizations are hybrid, the understanding of success becomes even more problematic. Hybridity has been defined as "the mixing of core organizational elements that would not conventionally go together" (Battilana et al., 2017, p. 129) and "combinations that violate institutionalized rules about what is appropriate or compatible" (Battilana et al., 2017, p. 138). Hybrids "draw on at least two different sectoral paradigms, logics and value systems" (Doherty et al., 2014, p. 418). Battilana et al. (2017) conceptualize hybridity as a matter of degree rather than type, because a growing number of organizations exhibit some degree of hybridity such as nonprofits commercializing for financial sustainability (Dees, 1998) or state organizations implementing business-like procedures (Fotaki, 2011;Hayllar & Wettenhall, 2013). When organizations have multiple identities/goals, it may be hard to reconcile them (Battilana & Lee, 2014;Hahn & Ince, 2016). Such consensus deficiency creates a "blurred" vision of success (Battilana & Lee, 2014). In the case of social enterprises, for example, success is determined by both their social impact and financial success (Hahn & Ince, 2016;Moss et al., 2011). Hybrids must combine institutional logics, each providing a different "set of assumptions and values, usually implicit, about how to interpret organizational reality, what constitutes appropriate behavior, and how to succeed" (Thornton & Ocasio, 1999, p. 804). Thus, hybrid organizations face conflicting institutional demands (Santos & Eisenhardt, 2009), which are bound to limit their potential for success. A hybrid nature augments the hurdles of microorganizations that can face even more difficulties if they operate in contested categories. Therefore, the question whether high failure rates result not only from resource constraints but also from a blurred vision of success should be relevant to both practice and research.
The environmental context of organizations has been widely acknowledged as an important determinant of business success (Kotey & Meredith, 1997;Rogoff et al., 2004;Ruffo et al., 2018;Wang et al., 2017). Environmental context has also been identified as an important external antecedent of hybridity (Battilana et al., 2017). Institutional environments are not static, and regulatory or cultural changes may create pressures for organizations to develop models that combine elements of multiple identities. For example, the market logic has permeated several domains in Western societies (Davis & Marquis, 2005), such as medicine (Reay & Hinings, 2005), that were historically dominated by the professional logic. Hybrid microorganizations operating within a contested market category face further limitations because they experience a legitimacy vacuum due to the lack of a "socially familiar categorical type" (Dobrev & Gotsopoulos, 2010, p. 1153. Legitimacy vacuum is understood as an "environmental deficiency" (Dobrev & Gotsopoulos, 2010, p. 1157. Stakeholders are more reluctant to support organizations with deficient legitimacy, because they struggle to assess their potential for success due to limited institutional consent (Dobrev & Gotsopoulos, 2010;Ruffo et al., 2018;Wang et al., 2017).
A summary of the literature review is compiled in Table 1.

Context of study
CAM hybrid microorganizations constitute the context of the study. They position themselves as "complementary" and/or "alternative" medicine offering therapies embedded in "holistic" and "whole person" values, patient empowerment, and use of natural remedies (Frass et al., 2012;Hirschkorn, 2006;Keshet, 2010). CAM spans unconventional, alternative, or unorthodox therapies designed to address health problems and heal patients. CAM services include (1) alternative medical therapies (e.g., homeopathy, acupuncture, and naturopathy); (2) mind-body approaches (e.g., hypnosis and biofeedback); (3) biologically based treatments (e.g., iridology and aromatherapy); (4) body-manipulative methods (e.g., osteopathy, massage, and reflexology); and (5) energy rehabilitation (e.g., healing and reiki) (Frass et al., 2012;Keshet, 2010). CAM organizations in this study are very small, with self-employed therapists or up to 10 employees and a balance sheet total that does not exceed 2 million euros, which classifies them as microorganizations (European Commission, 2005). CAM organizations target a market niche-customers who look for alternative or complementary services to conventional "scientific" medicine (Markman & Waldron, 2014). CAM establishments are good examples of contested hybrid organizations. First, CAM's legitimacy is publicly questioned; it has been described as nonscientific quackery (Almeida, 2016;Almeida & Gabe, 2016;Winnick, 2005). Certain CAM practices are considered an inappropriate option for health care since they challenge some basic assumptions of orthodox medicine (Mizrachi et al., 2005). CAM is often surrounded by controversy and scrutiny, caused to a large extent by the inappropriate safeguards for minimizing potential harm for CAM users (Wardle & Adams, 2014). Several risks associated with the use of CAM include financial exploitation of patients, unnecessary treatments, and patient harm (Bodeker & Kronenberg, 2002;Wardle & Adams, 2014). Many CAM treatments have not been tested following the standard methods of biomedicine (Almeida, 2016;Wardle & Adams, 2014). Hence, one of the main reasons for the contestation of CAM is related to the rigid boundaries between conventional and unconventional medicine. CAM is largely defined by exclusion from conventional medicine, which impedes the legitimization of organizations within the category (Almeida & Gabe, 2016;Wardle & Adams, 2014). It has been seen as  Gorgievski et al., 2011;Kotlar et al., 2018;Maltz et al., 2003;Reijonen & Komppula, 2007;Singh et al., 2016;Wang et al., 2017.
Success as a subjective measure -Managers have their own perceptions of the meaning of success. -Subjective measures have been used in past research mostly applied to performance in organizational and entrepreneurial contexts. -Subjective success is an individual understanding and assessment of the achievement of criteria that are personally important to the manager, such as the business's performance, growth and status. Gorgievski et al., 2011;Reijonen & Komppula, 2007;Rogoff et al., 2004;Wach et al., 2016;Wang et al., 2017.

Hybrid microorganizations and contested categories
Microorganizations and success -Survival rates are low due to barriers such as resource constraints, market inexperience and lack of legitimacy, vulnerability, and risk and uncertainty. -Decision-making process is mostly singleperson centered and thus success is embedded in the manager's personal and subjective business goals. a "residual category," since it has grown not because of the existence of consistent and clear categorical boundaries but as a result of being excluded by conventional medicine (Wolpe, 2002). Because they do not follow the scientific method, CAM organizations are ostracized and marginalized by biomedicine and societal institutions, situating CAM organizations in a legitimacy vacuum (Dobrev & Gotsopoulos, 2010). A second reason for contestation of CAM practices is the lack of regulation, which is an important issue for CAM managers/practitioners (Almeida & Gabe, 2016;Battilana & Lee, 2014;Hirschkorn, 2006;Wardle & Adams, 2014). Wardle and Adams (2014, p. 412) state that "failure to regulate a CAM profession can lead to a steady decline in training standards, particularly when combined with professional fragmentation or lack of a defined standard for CAM practitioners." As managers have to deal with the lack or insufficiency of training of CAM practitioners, regulatory and legal mechanisms of CAM practice, and an unclear demarcation of health/business categories, such managers might find themselves in a legal limbo. Third, being for-profit-healthservices providers spans the category boundaries and logics of "business/ profit" and "health/care," creating ethical dilemmas for managers and practitioners. Depending on the predominant logic in their background, they may feel uncomfortable running a for-profit establishment or making decisions whether to treat a patient who cannot afford a treatment. Such clashes of logics would reflect on their notion of success (Wardle & Adams, 2014). Fourth, CAM hybridity extends beyond the health/business boundaries as CAM combines practices and techniques from the West, such as medicine, and the East, such as healing (Keshet, 2010). Its hybrid nature mixes core aspects that convention would not allow to be together by drawing on different sectoral paradigms (Doherty et al., 2014) from two separate knowledge categories (Keshet, 2010)-biomedicine (based on the scientific method) and "complementary or alternative" medicine (not based on the scientific method). CAM practitioners attempt to respond to the demands from biomedicine, such as following medical-like procedures (e.g., collecting

Main concepts Authors (year) Contested categories
-Environmental context of organizations is an important determinant of success and an external antecedent of hybridity. -Hybrid microorganizations in contested categories lack "socially familiar categorical type" and have "environmental deficiency." -Organizations in contested categories struggle to assess their potential for success due to limited institutional consent. Battilana et al., 2017;Dobrev & Gotsopoulos, 2010;Kotey & Meredith, 1997;Rogoff et al., 2004;Ruffo et al., 2018;Wang et al., 2017. detailed clinical information, keeping records of patient medical history, and identifying patient symptoms) (Keshet, 2010), and from "complementary or alternative" medicine, which focuses on treating the cause instead of symptoms and healing through natural procedures (Wardle & Adams, 2014). In sum, CAM practices are in a market category without clear boundaries and their legitimacy is contested. The complex hybridity of the business logic, the scientific logic, and the healing logic creates an extremely difficult terrain for CAM microorganizations. Such conditions of fluid boundaries, competing logics, slack regulations, and scarce resources are bound to influence decision-making by CAM managers and their vision of success.

Methods
By combining cognitive mapping and MCDA, the current study aims to address the complexity of CAM organizational hybridity and micro size together with adverse environmental conditions in revealing managers' concept of success. The development of a collective cognitive map allows the decision problem to be structured, while the application of MCDA allows weights to be assigned to perceived success factors, so that an assessment mechanism can be developed for CAM organizations' profile analysis.

Cognitive mapping
Cognitive mapping is a simple, interactive, and extremely versatile problem-structuring tool, which facilitates and encourages discussion among decision-makers (Eden, 2004). Although subjective and strongly dependent on participants' willingness and availability, this constructivist approach allows for increased transparency and a reduction of omitted criteria in the decision-making framework, leading to a better understanding of the decision problem at hand (Marques et al., 2013). Cognitive maps are graphic representations of nodes and links that can assume diverse visual and interactive forms and they are well accepted in the decision-making field (Eden, 2004;Eden & Ackermann, 2004). At the top of the hierarchy is the goal of the decision problem. Then, the maps follow a network of nodes and arrows as links, where the direction of an arrow implies a cause-and-effect relationship between two linked concepts (Eden & Ackermann, 2004;Tegarden & Sheetz, 2003).

MCDA and MACBETH
The advantage of applying MCDA and measurement attractiveness by a categorical-based evaluation technique (MACBETH) to microorganizations relates to a characteristic of our study-the development of an evaluation framework of key success factors from the point of view of managers of contested microorganizations. As discussed, microorganizations often struggle to survive and do not have a clear vision of success. The integrated use of cognitive mapping and MCDA helps the group of decision-makers discuss the factors that in their vision best assure organizational success.
MCDA is described as "a collection of formal approaches which seek to take explicit account of multiple criteria in helping individuals or groups explore decisions that matter" (Belton & Stewart, 2002, p. 2). It is applied when there is a desire for a formal procedure to assist with decision-making (Ferreira & Santos, 2018;Montbelier & Belton, 2006) and when there is an established and well-supported approach in decision science (Ferreira & Santos, 2018;. As an MCDA technique, MACBETH quantifies differences of attractiveness among elements of a certain set based on semantic-value judgments (Bana E Costa & Vansnick, 1997). Through a constructive learning process, MACBETH uses simple qualitative question-answer procedures that allow decision-makers to enter the domain of cardinal measurement. Due to its intrinsic characteristics as a humanistic decision-making tool based on mathematics, the technique has been successfully applied in different decisionmaking contexts (Bana E Costa et al., 2012).

Participants
The high level of interaction between the decision-makers and the nature of the problem requires a panel of relevant and experienced participants, since they are the ones responsible for the development of the evaluation framework. The role of the researchers is to facilitate the application of the methodologies and conduct the negotiation process Belton & Stewart, 2002).
This study was conducted in Portugal where CAM is still emerging as it witnesses the slow legalization of some of its practices (Almeida & Gabe, 2016) but remains marginalized and contested in the Portuguese health care system (Almeida, 2016;Almeida & Gabe, 2016). In setting up the panel of decision-makers, the aim was to ensure participants who: (1) had started or managed legally registered CAM organizations; (2) had worked in the field for more than 10 years; (3) have less than 10 employees and; (4) were available for three face-to-face group meetings with an average duration of four hours. To ensure that CAM organizations were legally recognized, a more systematic approach to the participants' selection was used. The sampling frame was the AMADEUS database that identifies registered CAM organizations under the economic activity code "86906 -other human health activities." After applying the three selection criteria above, the list resulted in 48 organizations. The time commitment necessary for participation was a major challenge that reduced the final group to seven decision-makers.
Two of the participants (alpha 6 and alpha 7) were unable to attend the last two sessions. This situation has happened in other studies using the MACBETH method (Ferreira, 2011). Nevertheless, the input of these two participants was taken into account in the collective cognitive map . The profiles of the participants and their organizations appear in Table 2. The academic background of the panel members varies; the backgrounds include physiotherapy, public relations and advertising, and nuclear medicine. The participants' CAM specializations are mostly based in osteopathy, Chinese medicine, ayurvedic medicine, and reiki.
While a group of seven managers is not large, it should be noted that (1) it falls within the recommended guidelines-between 5 and 12 participantsfor this type of study ); (2) other studies applying cognitive mapping and MACBETH have also addressed the respective decision problems with smaller groups Filipe et al., 2015); (3) this study is process oriented (Bell & Morse, 2013) (i.e., although the output reflects the ideas and experience of this particular group of participants, due to the constructivist stance of this research, the procedures followed can be replicated in other contexts and/or with other participants). Two of the authors-both with practical experience as group facilitators in different contexts-conducted the meetings. Anonymity was promised to the participants; thus their organizations were labeled "Alphas."

Application and results
The multidimensional framework in this study developed through three main phases: (1) the structuring phase; (2) the evaluation phase, and (3) the recommendations phase. The procedures followed in each of these phases are presented in Figure 1.
During the structuring phase, the factors underlying the decision problem were specified. In this case, the aim was to identify the success factors of CAM organizations by applying cognitive mapping techniques. In the evaluation phase, MACBETH was used to obtain value functions and calculate trade-offs between evaluation criteria. In the recommendations phase, recommendations and improvement suggestions were formulated based on the obtained results.

The structuring phase
The structuring phase involved the construction of the cognitive map and the development of a tree of fundamental points of view (FPVs) and the development of the descriptors and respective performance impact levels Filipe et al., 2015). Two workshop sessions of about 4 hours each were conducted with this aim. Building the cognitive map This phase consists of identifying the criteria and building the collective cognitive map. Cognitive mapping usually begins with a "trigger question" to elicit participants' perceptions. For the current study, it was, "Based on your own values and professional experience, what should be the goals and characteristics of a CAM organization so that it can be considered successful?" Methodologically, the strategic options development and analysis (SODA II) approach was followed, where the decision-makers are jointly involved in a workshop (Belton & Stewart, 2010;Eden, 1995). The first session proceeded with the application of the "post-its technique" , where participants were encouraged to brainstorm on the subject through active discussion until a saturation point was reached. During that process, participants wrote relevant criteria for a successful organization on post-its and stuck them on a whiteboard for easy visualization . The facilitators instructed the participants to write only one main idea per postit. The next step was to identify key areas of concern and build clusters from the post-its (Belton & Stewart, 2002). Subsequently, the clusters are given a name, which should capture the unifying concept of the cluster and should be Source. Ensslin et al. (2000, adapted) positioned hierarchically, wherein the most general concepts are at the top of the cluster and the more specific ones at the bottom (Belton & Stewart, 2002).
To conclude the first stage of the structuring phase, all the decision-makers were asked to agree on the form and content of the cognitive map. The final version of the cognitive map contains 187 nodes, above the minimum of 100 nodes suggested by Eden (2004), which were grouped into seven clusters (i.e., infrastructure, management, marketing, professional development, training, external factors, and organizational aspects). This final version was analyzed and validated by the group in a collective discussion (see Figure 2). The map was constructed in the Decision Explorer software-a popular software for cognitive mapping (Belton & Stewart, 2002;Eden, 2004). The collective cognitive map has been considered a valuable tool to both structuring and understanding of complex decision problems Tegarden & Sheetz, 2003). It allowed the panel members to share their perspectives and experiences, significantly reducing the rate of omitted criteria in the decision-making process (Montbelier & Belton, 2006;Tegarden & Sheetz, 2003).

Tree of fundamental points of view (FPVs)
The second group session started with a review of the cognitive map. The decision-makers were asked again to agree on the content and form of the cognitive map (they suggested one minor change, which was incorporated in the final version). After they all agreed that the map contained the most relevant aspects of the goals and characteristics of a successful CAM organization, the study proceeded to the next stage-the creation of a tree of FPVs (Bana E Costa et al., 2012). Following Keeney's (1992) methodological guidelines, the M-MACBETH software (www.m-macbeth. com) was used to pass from the cognitive map to the tree of FPVs. The value tree was built from the branches of the cognitive map, and each FPV was composed of relevant criteria chosen by the panel of decision-makers (Montbelier & Belton, 2006). To ensure the value tree's properties, mutual preferential independence tests among FPVs were carried out until reaching a nonredundant set of FPVs (Bana E Costa et al., 2012). The seven FPVs composing the value tree and their meaning are presented in Table 3. The decision-makers considered the tree of criteria containing these FPVs complete and representative of the group consensus.
Descriptors and impact levels Descriptors were carefully defined by the decision-makers and served as the basis for the evaluation phase. Impact levels are performance levels that allow for local evaluations in each descriptor (e.g., L i with i = 1, …, n). For instance, the panel of decision-makers collectively decided to define the descriptor as numerical intervals for each of the impact levels previously defined (e.g., L1: [16-24];  . The creation of descriptors in this study resulted from an adaptation of Fiedler's (1967) scale, considered a very consistent psychometric tool in the field of MCDA Filipe et al., 2015). Previous research suggests that to prevent cognitive fatigue, it is important to set "good" and "neutral" reference levels (Bana E Costa et al., 2012). As can be seen in Figure 3, FPV 2 becomes operational through a professional development (PD) index that contains six ordered impact levels (L i = 1, …, 6) based on the decision-makers' value preferences. For instance, their preferences within organizational aspects are (1) between 10 and 20 years of active experience; (2) continuous training perfectly suited to professional skills; (3) technical mastery inherent to the   profession; (4) total availability in the clinical monitoring; (5) effective and regular monitoring of the results of users; and (5) excellent perception of users and colleagues of interpersonal qualities of professionals. As shown in Figure 3, L 1 represents the best performance possible, while L 6 represents the worst performance level. This procedure was repeated for all FPVs, as each descriptor represents a different dimension and can present a different number of impact levels. According to the literature, the construction of descriptors and impact levels constitutes the final stage of the structuring phase (Montbelier & Belton, 2006). After sorting the descriptors and impact levels for the FPVs, a value function for each FPV was obtained, reflecting the decisionmakers' preferences (Belton & Stewart, 2002). This allowed the start of the evaluation phase.

The evaluation phase
The third and last session consisted in the application of MACBETH, aiming at obtaining (1) value judgments and local value scales; (2) trade-offs among criteria; and (3) overall scores for evaluated Alphas.

Value judgments and local value scales
The session started with the creation of local value scales. This step consists of filling matrices of value judgment for each of the descriptors Filipe et al., 2015) using the semantic categories of difference of attractiveness proposed by the MACBETH approach-namely, C 0null; C 1very weak; C 2weak; C 3moderate; C 4strong; C 5very strong; and C 6extreme (Bana E Costa et al., 2012). The value judgment is facilitated by nonnumerical pairwise comparisons of difference of attractiveness between the impact levels (Bana E Costa et al., 2012;Ferreira et al., 2015). This means that the decision-makers are asked to make comparisons between the impact levels of each FPV and then attribute a semantic category to this comparison (Bana E Costa et al., 2012). Figure 4 shows the value judgments provided for professional development (FPV 2 ).
As can be seen, the decision-makers attributed a semantic category of "weak" (C2) to the difference of attractiveness between L 2 and L 1 . By applying linear programming to the value judgments projected, a partial-value function (or cardinal-value function) is obtained for each FPV (Bana E Costa & Vansnick, 1997). The value judgments expressed by the decision-makers for each descriptor were then entered in the M-MACBETH software, and incompatibility between semantic judgments was automatically verified. In case of inconsistencies, the judgments were reanalyzed and the inconsistencies resolved (Bana E Costa et al., 2012;Filipe et al., 2015). As exemplified in Figure 4, the value scale obtained for professional development (FPV 2 ) attributed a partial score of 300 points to L 1 (highest level), whereas the lowest level (L 6 ) was assigned a negative score of −125 points. It should be noted that the allocation of 100 points to the good level and 0 to the neutral level is a standard procedure carried out in all descriptors, which facilitates the cognitive comparisons made by the decision-makers (Bana E Costa et al., 2012;Filipe et al., 2015).

Trade-off procedures
After obtaining a cardinal scale for each descriptor, the next step entailed obtaining the trade-offs (i.e., weights or substitution rates) between the FPVs (Bana E Costa et al., 2012). First, the decision-makers fill in an ordering matrix of pairwise comparisons to rank the seven FPVs. Whenever an FPV was considered more attractive than another, it was assigned a value of 1 and a value of 0 otherwise (Filipe et al., 2015). The matrix of overall attractiveness ranked "professional development" (FPV 2 ) in first place, while external factors (FPV 5 ) were ranked last (see Table 4).
Once the FPVs were ordered, the next step was filling in a matrix of differences of attractiveness between FPVs, based on the MACBETH semantic categories previously defined. This process is interactive and allows the tradeoffs among FPVs and the respective value functions to be obtained, discussed, and approved by the panel members. Figure 5 represents the results.
The results indicate that professional development (FPV 2 ) has the highest weight, with 23.24 percent, followed by training (FPV 1 ) with 22.22 percent. The lowest weight (1.01 percent) belongs to external factors (FPV 5 ). It is interesting to note the low importance assigned to infrastructure (FPV 6 ) and external factors (FPV 5 ). These results contrast with prior findings on the importance of external factors for the success of a business (Rogoff et al., 2004). Because these are factors derived from managers' perceptions, it can be conjectured that due to their limited resources, hybrid microorganizations do not pay enough attention to environmental factors. Due to the inherent subjectivity in the process, the defined weights should have sufficient flexibility (e.g., confidence intervals) so that the weights do not miss the  consistency of judgments made by the decision-makers in case of variation within certain parameters. Next, the evaluation framework was tested.
Measuring perceived success: A practical application of the evaluation framework Using information provided by the panel members, partial and overall scores for each of the CAM organizations (Alphas) were obtained. Table 5 presents the levels of partial performance of each Alpha. For example, the partial values show that Alpha 3 is the worst performer on management (FPV 4 ) and that Alpha 2 is the best performer on marketing (FPV 3 ). The scores obtained in each descriptor also show that, except Alpha 3, all Alphas score above the good reference level in terms of professional development (FPV 2 ). The next step was to aggregate the partial performances and obtain an overall score for each Alpha (see Table 6). For the purpose, a simple additive aggregation model was used, where Good and Neutral stand for two fictitious CAM organizations that were defined by the panel members to serve as "anchors" and facilitate cognitive comparisons (for details, see Bana E Costa et al., 2012). Good represents a CAM organization that performs at a good level in all FPVs, while Neutral is a CAM organization that performs at a neutral level in all FPVs.
As can be seen in Table 6, Alpha 2 presents the highest overall score (155.57), while Alpha 3 shows the worst performance (1.01). More important than the Alphas' ranking, however, is the profile analysis that can be carried out for each Alpha, which allows well-focused intervention priorities to be immediately detected. Figure 6 presents the cases of Alpha 2 and Alpha 3.  The analysis emphasizes in which FPVs the Alphas need to improve in order to increase their chances of success. Even though the participants spent a total of more than 12 hours in group sessions, they appreciated the results and thought it was time well spent.

Analyzing and validating the results
Exploring the sensitivity and robustness of the evaluation system provides a deeper understanding of the decision problem (Bana E Costa et al., 2012). Both analyses intend to investigate the model outputs in light of some type of data uncertainty (Bana E Costa et al., 2012;Belton & Stewart, 2002). Sensitivity analysis aims to explore the impact of changes in a criterion's weight on the system's output. For example, the sensitivity analysis of professional development (FPV 2 ) (the weight attributed is 23.24) reveals that its weight can vary significantly (approximately to the boundary level of 14 and 25) without violating the decision-makers' value preferences nor the position of the ranking of each Alpha, as shown in Figure 7.
Robustness analysis, in turn, works with effects in the model outputs caused by simultaneous changes in different criteria . Two types of dominance were analyzed: (1) absolute or classical (▲), where alternative a is globally better than alternative b and partially better or equal to b in all identified FPVs and (2) additive (✚), where alternative a is globally better than alternative b but is not partially better than b in at least one FPV (Bana E Costa et al., 2012;Ferreira et al., 2015;Filipe et al., 2015). A battery of robustness analyses was carried out and, as exemplified in Figure 8, the model created can be considered robust, since significant simultaneous variations (e.g., of +5%, +10%, +5%, +2% and +1% in the weights of FPV 1 , FPV 3 , FPV 4 , and FPV 5 , respectively) are required to produce changes in the Alphas' ranking. In other words, we can say that the analyses carried out confirmed the robustness of the evaluation framework, because the stability is verified under simultaneous changes in the weighting coefficients.
Such results reveal a high degree of consistency . The results of the sensitivity and robustness analyses indicate that the evaluation framework is adequate and the ordering of the importance of the major factors would not change despite the underlying subjective manner of their derivation.
In a follow-up stage, the participants had to answer a short survey to help the researchers gain further insight into the extent to which the evaluation system was preferred to other practices in place. This exercise aimed to assess the perception of the panel members regarding the potential practical adoption of the proposed framework and followed the four stages of the performance measurement process (i.e., design, measurement, analysis, and improvement) discussed in Filipe et al. (2015). A five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree) was used to collect the participants' perceptions (see Table 7). The participants showed strong agreement regarding the value of the evaluation system proposed to assist them in successfully carrying out the four stages of the performance measurement process.

The recommendations phase
One of the main advantages of the adopted approach is its constructive nature. The framework was built together with and validated by the panel members, where their experiences and knowledge contributed from the beginning until the final solution. In addition, the participants received assessment of their organizations based on the constructed framework for success. The evaluation system presented in this study should not be considered an optimal and final solution. Even though this might be seen as a limitation of the methodology, one must keep in mind that the evaluation framework is process oriented and reflects the agreement of the decisionmakers throughout the process (Bell & Morse, 2013).

Main findings and discussion
This study set out to explore how contested hybrid microorganizations conceptualize success. For the purpose, an expert panel developed an evaluation framework of perceived drivers affecting organizational success. Cognitive mapping and MACBETH were combined to identify and quantify perceived factors leading to the success of CAM hybrid microorganizations.  The expert panel identified 187 success criteria, which were grouped into seven clusters: external factors, infrastructure, management, marketing, professional development, organizational aspects, and training. The final perceived success-factors framework shows that professional development and training are the most important indicators and represent 46 percent of the success framework. Management and marketing factors account for 36 percent. Organizational aspects represent 11 percent and infrastructure, 6 percent. External factors represent only 1 percent of the overall success framework (see Figure 5). The conclusion is that the participants emphasize human capital as the related three factors taken together represent more than 50 percent of the perceived success factors: professional development (i.e., professional skills and know-how required from human capital and their continuous professional development); training (i.e., academic, scientific, and technical qualifications), and organizational aspects (i.e., appropriateness of human capital to the organizational needs). These findings are in line with previous research (Gorgievski et al., 2011;Maltz et al., 2003;Rogoff et al., 2004;Staniewski, 2016;Wach et al., 2016). The organizational perspective advocates that adequate human capital and organizational design are linked to the survival and success of organizations (Birley & Westhead, 1990;Combs et al., 2005;Hopp & Sonderegger, 2015). Hybrid establishments may be particularly constrained in choosing an adequate workforce as it would need to address multiple institutional logics demands. Consequently, specialized training and development appear to be a major instrument of "hybridizing" personnel. It is not surprising that CAM microorganizations are concerned with delivering a good service since CAM seeks to respond to society's search for morepersonalized services and a "whole person" approach (Winnick, 2005). For example, organizations in a legitimacy vacuum have to offer a set of recognizable practices that will increase the organization's ability to decrease its environmental deficiency (Dobrev & Gotsopoulos, 2010;Wang et al., 2017). Confirming prior research, education, and professional experience are perceived to contribute to increased legitimacy as they provide credibility (Ruffo et al., 2018;Tornikoski & Newbert, 2007;Wang et al., 2017). For example, educational and professional certifications are a way of complying with established regulations and standards (Wang et al., 2017). Accordingly, human capital and professional image are seen as crucial for success for hybrid organizations in the contested CAM category. The next most important success factors are management (i.e. the organization's financial condition, managers' skills and leadership abilities, and their moral and ethical conduct), marketing (i.e. strategic (market research and positioning), and tactical actions (product/service, communication, distribution and price)). The management indicators identified by the panel are both financial (e.g., annual turnover) and nonfinancial measures (management, marketing, leadership, and ethics); most of the measures are operational (see Figure 2-Cognitive Mapping). The only financial measures mentioned by the managers were financial sustainability and annual turnover. This is surprising, as financial measures such as sales growth and return on equity (ROE) are commonly used and considered to be the most important measures in organizational success (Birley & Westhead, 1990;Maltz et al., 2003;Reijonen & Komppula, 2007;Singh et al., 2016;Staniewski, 2016;Staniewski & Awruk, 2019). The small business literature indicates growth and profitability as the two most often used performance measures (Gorgievski et al., 2011;Lonial & Carter, 2015). The microorganizations in this study do not consider profit and growth as the main factors in their success. While many quantitative studies use financial measures as dependent variables (Birley & Westhead, 1990;Combs et al., 2005;Lonial & Carter, 2015), the current findings agree with Gorgievski et al. (2011) who state that many organizations are not preoccupied with maximizing financial performance but are rather more concerned with indicators such as "contributing back to society, personal satisfaction, satisfied stakeholders (employees and customers), work-life balance, public recognition, and utility or usefulness" (Gorgievski et al., 2011, p. 212). These findings extend the belief that business success is related to organizational goals (Kotlar et al., 2018). While the legitimacy literature argues that stakeholders appreciate organizations that are financially sustainable for the long run (Ruffo et al., 2018), the participants do not consider financial indicators to be their primary mechanism for measuring success. In line with the literature, managers' skills and abilities are other important factors increasing legitimacy (Tornikoski & Newbert, 2007;Überbacher, 2014;Ruffo et al., 2018;Wang, Stewart, & Castro 2017). This is particularly relevant in a contested-category context, as the competence and credibility of managers may be crucial in promoting and creating a favorable image of the organization (Galvin et al., 2004;Überbacher, 2014).
Regarding marketing, previous research in small-business success suggests that marketing is fundamental, because it aims to understand the market and promote the business (Rogoff et al., 2004). The findings confirm these claims as, for example, participation in events, promotion, and communication initiatives may be used to endorse the legitimacy of the organization (Rogoff et al., 2004;Tornikoski & Newbert, 2007). Surprisingly, the least important factors for success are infrastructure (facilities and its physical surrounding) and external factors (social, political, economic, legal and competitive factors). The participants in the study perceive the facilities' conditions and location as secondary to the business success of CAM microorganizations. This contradicts previous studies, which report that infrastructure significantly affects success and the performance of organizations (Birley & Westhead, 1990;Reijonen & Komppula, 2007). Finally, external factors are perceived to be the least important factor for success (1 percent of the overall evaluation framework). The expectation was that external factors such as economic and competitive context, public beliefs about CAM practices, and regulations would be of particular relevance due to the hostile environment constricting CAM organizations' legitimacy. One tentative explanation is that CAM managers pragmatically focus on what is under their direct control, which excludes most external factors. Also, short-term survival might imply a forced focus on internal ability rather than long-term and strategic decisions as would be the case in larger organizations with more resources and capabilities. It is important to note that the managers did not have access to the success indicators identified in the literature, and the indicators that appear in the evaluation framework resulted from the discussion among the decision-makers and represent their collective view of success of a hybrid microorganization in a contested category.

Conclusions
Crossover research between microorganizations and organizational struggle for legitimacy is still in its infancy. This study addresses the lack of research related to the success of hybrid microorganizations in contested market categories. It also responds to a wider scarcity of research on subjective perceived organizational success factors. Most prior studies are based on measures selected by the researcher rather than on the managers' point of view. To explore how managers perceive the success of their hybrid microorganizations in a market category that lacks legitimacy, this study uses the context of complementary and alternative medicine (CAM).
The main contribution is the documentation of the process of a successframework construction from the point of view of managers of hybrid microorganizations. In addition, the study applies cognitive mapping and MACBETH to construct the framework-a methodological contribution to the small business literature. As opposed to survey methods, this constructivist approach: (1) identifies subjective and objective components of success (FPVs) along with cause-and-effect linkages between criteria, which are hard to identify exclusively by statistical approaches; (2) calculates the trade-offs between these components based on managers' practical experience and collective perceptions to achieve negotiated rankings of the different components of success; (3) provides study participants the opportunity to reflect on the assessments and suggest focused improvements; and (4) develops an easy-to-replicate process-oriented framework that can be used in other contexts. The MCDA approach leads to a justifiable evaluation framework from the managers' point of view, because the decisionmakers exchange ideas and learn in an open and interactive environment.
The findings also add some valuable insight into the small business literature. First, hybridity, size, and scale disadvantages mark to a great extent the decision-making process of managers. The results show that hybrid microorganizations, especially ones in a contested category, assess success in a different way compared to traditional businesses; namely, they prefer nonoperational indicators to financial measures. While growth is particularly important as a measure of success for larger companies, the fragile nature of the microorganizations prompts them to be more tactical and short-term oriented. Second, managers of microorganizations with legitimacy deficit do not perceive external factors to be major determinants of success. This result disagrees with the notion that environmental conditions are of utmost importance for the success of organizations.
In practical terms, this study allows managers of hybrid microorganizations to use the resulting framework to reflect on their own perceptions of success as it provides a reference to collective perceived success in organizations with similar characteristics. This research brings to light what possible success criteria and factors can be used to individually or collectively understand success.

Limitations and future research
The major limitation of this research is the lack of generalizability. The final evaluation framework cannot be considered a definitive and universal solution as it is context dependent. It reflects the specific values of the participants and it uses the CAM hybrid microorganizations as a single illustrative example of a contested category. The focus of this study is a multifaceted problem-managers' subjective understanding of organizational successwith multiple influences (e.g., internal and external to the organization). Further investigation of other hybrid microorganizations' perceptions of success in a nonsupportive environmental context would allow comparability among findings and could aid the delivery of a more generalizable framework. Future longitudinal research could better capture its dynamicsnamely, whether the vision of CAM managers' success factors changes as they develop their business or as it becomes more legitimized. Although the sample of decision-makers was within the sample limits proposed in studies of the topic, larger samples could produce additional insights.
The sampled managers perceive that external factors are not among their primary concerns when thinking about success raises the question, to be further explored in future research, of whether it is possible that hybrid microorganizations are prone to failure because they have alternative visions of success that are not fully compliant with expectations from their environment. This research brings to light what possible success criteria and factors can be used to individually or collectively understand success. Future research could test the hypothesis that limited internal and external resources prevent managers of microorganizations from having a clear sense of the path to long-term organizational success.