ORGANIZATIONAL IMPROVISATION: A CONSOLIDATING REVIEW AND FRAMEWORK Allègre L. HADIDA* University of Cambridge Judge Business School and Magdalene College Trumpington Street, Cambridge CB2 1AG United Kingdom Tel.: + 44 (0) 1223 339 612 E-mail: [email protected] William TARVAINEN E-mail: [email protected] June 2014 Forthcoming in International Journal of Management Reviews We gratefully acknowledge the very helpful comments of Tim Edwards, Katie Jones, Joe Lampel, Allan Macpherson, Jed Rose, Mark Thompson, Paul Tracey and two anonymous reviewers on earlier versions of this article. All errors and omissions remain ours. 1 Organizational Improvisation: A Consolidating Review and Framework Abstract Organizational improvisation is increasingly recognized as a relevant area of management research. However, the cumulativeness of research on improvisation in organizations remains low. This article organizes existing contributions on organizational improvisation within a new consolidating framework combining degrees (minor, bounded, and structural) and levels (individual, interpersonal, and organizational) of improvisation. The proposed degree/level framework allows for reviewing the existing literature on organizational improvisation in the management disciplines of strategy, organizational behavior, organizational theory, innovation and marketing in a systematic manner. It also exposes potential areas for future research across management disciplines, research areas, organizational settings and industries, and beyond existing metaphors, most notably of jazz and improvisational theatre. Keywords Organizational Improvisation, Metaphor, Jazz, Improvisational Theatre, Creativity 2 Organizational Improvisation: A Consolidating Review and Framework Introduction Accelerating globalization, trade liberalization and the increased interconnectedness permitted by advances in information systems and the Internet render long-term plans obsolete at a moment’s notice, and force organizations to nimbly and creatively navigate constantly evolving landscapes (D’Aveni 1994; Hamel and Breen 2007; Schrey gg and Sydow 2010). To illustrate their ability and willingness to improvise and adapt to changing circumstances, some multinationals compare themselves to “a jazz band, not a symphony orchestra” (Steinbock 2010: 107). Chief Executive and former professional comedian Dick Costolo also reports that he regularly applies lessons learned from improvisational theatre to running Twitter (Bilton 2012). As emergent strategies decreasingly conform to deliberate strategizing (Mintzberg and Waters 1985), organizational improvisation (OI) combines with heuristics1 to help managers understand and analyze organizational decisions and actions that display the following characteristics. They are complex and dynamic, cannot be understood a priori or managed using existing routines, and demand flexible and extemporaneous action (Ciborra 1999; Kamoche and Cunha 2001; Kirsch 1996). OI may thus enable companies to subdue more of the emergent part of their actions and environmental fortuities to their own will (Cunha et al. 1999), in particular in environments characterized by high velocity, heterogeneity in experiences, plentiful opportunities and high unpredictability (Bingham and Eisenhardt 2011, 2014). Yet, fifteen years after Cunha’s et al. (1999) first review, the cumulativeness of OI research remains low. None of the existing taxonomies takes full account of the different types of improvisation, and most are specific to a single metaphor (in particular, jazz or less frequently, improvisational theatre). The absence of a consolidating framework of OI may be justified under the assumption that improvisation may be inherently uncontained, and that discussions of improvisation may be inherently distorted by the need to force its manifestations into a model. Even so, this absence poses a threat to the future congruence of the field. The framework of OI introduced in this article addresses these concerns. It illustrates when and under which circumstances OI is inherently uncontained, and when it is not. It also allows for the continual testing of the boundaries of what is and is not legitimately included in discussions of OI.2 More generally, it synthesizes metaphorical, empirical and anecdotal 3 works to advance a typology that covers the different forms of understanding OI. The proposed framework allows for reviewing the existing literature on OI in strategy, organizational behavior, organizational theory, innovation, and marketing in a systematic manner, and exposes areas for future research across these five management disciplines and beyond specific improvisation metaphors. The next section details our research methodology, and outlines the evolution of OI research before and after Cunha’s et al. (1999) first consolidating review. Section 3 introduces the definition of OI adopted in the article, and examines the “why” (rationale) and “how” (namely, metaphorical and naturalistic studies) of OI research to date. Section 4 lays the foundations of the proposed framework of OI. It discusses six taxonomies most commonly used in the OI literature, and defines degrees (minor, bounded and structural) and levels (individual, interpersonal and organizational) of OI. The combined degree/level framework of OI is introduced in Section 5. A discussion and conclusions section summarizes our findings, highlights the contributions and limitations of the degree/level framework and suggests directions for further research. Methodology Although organizations and their members have always improvised to some degree, research into OI is relatively recent. The academic field of management was slow to transcend the historical principles of managerial planning (Taylor 1911) and the definition of improvisation as a dysfunction in planning (March and Simon 1958) or in organizational design (MacKenzie 1986) put forward by the formal strategizing approach (Chandler 1962). New management perspectives emerged in the 1980s to respond to an increasingly dynamic business environment. They challenged conventional wisdom by noting that the environment may change before elaborate plans can be implemented, and by advising managers to create organizations that can flexibly respond to such changing circumstances (Mintzberg 1990; Mintzberg and McHugh 1985; Mintzberg and Waters 1985). Although initially contentious (Ansoff 1991; Mintzberg 1991), these developments led management research to shift its focus towards addressing change and OI (e.g., Jackson and Philip 2010). This article offers a systematic and comprehensive account of research published in English and investigating OI in strategy, organizational behavior, organizational theory, innovation, and marketing up to June 2014. These five disciplines of management were chosen on the basis of the importance of OI research carried out within their boundaries. We undertook data collection in three complementary steps. 4 We first carried out systematic manual and electronic searches for all articles published on OI in the following journals. For strategy, organizational theory and organizational behavior: Academy of Management Journal, Academy of Management Review, Administrative Science Quarterly, Organization Science, Strategic Management Journal, Journal of Management Studies, Journal of Management, Organization Studies and Organization; for innovation: R&D Management and Research Policy; and for marketing: Journal of Marketing, Journal of Marketing Research and Marketing Science. These journals were selected on the basis of their relevance to academic research in the five disciplines under study and of their high impact factors. We undertook this first step to make sure that we did not inadvertently overlook OI articles published in top journals. We then expanded and complemented this first collection with systematic reviews of OI articles in all other publications listed on EBSCO, JSTOR and ScienceDirect. The keywords “improvisation”, “organizational learning”, “spontaneous action”, “spontaneous process”, “chaos”, “order”, “structured”, “planning”, “jazz” and “theatre” were used to refine the electronic searches. This second step led to the identification of articles dealing with OI in a number of other journals, including (yet not limited to): British Journal of Management, International Journal of Management, Sloan Management Review, and Journal of Organizational Change Management. Third, we visited three libraries to search for additional publications, in particular books and monographs: the London Business School Library (London, UK), the University of Cambridge Central Library (Cambridge, UK), and the Sibelius Academy Library (Helsinki, Finland). As Figure 1 attests, studies of OI published before 1990 were few and far between. Following a very successful symposium at the 1995 Academy of Management Conference in Vancouver (Frost 1998, Meyer et al. 1998), their number peaked in 1998 with the publication of a special issue of Organization Science. A review article in the International Journal of Management Reviews followed in 1999 (Cunha et al. 1999). Our final sample, give or take involuntary oversights, constitutes the whole population of OI research in the selected five disciplines of management up to June 2014. It consists of 197 studies: 149 articles and working papers, 13 book and conference proceedings chapters, and 35 books and monographs. The overwhelming importance of peer-reviewed empirical articles within this population primarily reflects their precedence in disseminating knowledge in the five selected disciplines. Insert Figure 1 about here 5 Research published before Cunha et al. (1999) mostly developed along two streams. The first proposes arts-based metaphors, in particular jazz, to illustrate and shed light on improvisation in organizations (e.g., Hatch 1998). This first stream relies heavily on secondary data from musicology monographs (Berliner 1994; Kernfeld 1995; Schuller 1968) and builds on specific authors’ experiences of jazz as audiences or – on rare occasions – musicians (Peplowski 1998; Barrett and Peplowski 1998). The second stream of research uses empirical, naturalistic-based illustrations and anecdotal evidence to define improvisation and its causes and effects within organizations (e.g., Crossan and Sorrenti 1997). Both research streams have succeeded in creating interest in OI as a research topic and organizational phenomenon (e.g., Weick 1998). Noticeable shifts in focus of OI research after 1999 call for a new review. New metaphors of OI have emerged (e.g., Kamoche et al. 2000, 2003), and the improvisational theatre metaphor has gained in strength (e.g., Moss Kanter 2002; Vitug and Kleiner 2007). Post-1999, a larger number of contributions set out to investigate empirical implications of OI, particularly in new product development (e.g., Akgun and Lynn 2002; Samra, Lynn and Reilly 2008), in new ventures (Hmieleski and Corbett 2008; Evers and O’Gorman 2011), and under conditions of change and turbulence (Charles and Dawson 2011; Leybourne 2006). The methodology adopted in OI research has also shifted in recent years from a focus on detached theorizing in metaphorical works to conceptually rigorous in-depth case studies and interviews carried out in organizations (e.g., Bigley and Roberts 2001; Plowman et al. 2007; Sonenshein forthcoming) or about specific events (e.g., Mendonca and Wallace 2004; Brady 2011) during which improvisation took place. Taking Stock: Definition, Rationale for and Existing Studies of OI Organizational Improvisation: A Definition As a young, interdisciplinary and occasionally uncontained concept, OI has struggled to develop a comprehensive definition. There is, however, a high level of agreement on many of its properties (Vera and Crossan 1999). The word improvise comes from the Latin providere, “make preparation for”, and its derivative improvisus, “unforeseen” (Oxford Dictionaries 2014). Improvisation thus involves dealing with the unforeseen without the benefit of preparation. Improvisation is organizational when it is done by the organization or its members. It therefore occurs at various levels, and different dynamics apply to it depending on whether 6 improvisation happens within one (individual), between two or a few (interpersonal) or amongst many (organizational) individual actors. Table 1 provides an overview of central OI research, which it lists by primary level. As Table 1 illustrates, OI research has mainly focused so far on the individual and organizational levels. There also seems to be no major trend in how contributions differ in their description of improvisation. Since improvisation is a fairly commonly used concept in everyday discourse and given the challenges associated with trying to contain it, some research even eschews all explicit definitions of OI. Taking the term for granted, however, may lead to confusion and misunderstandings. Insert Table 1 about here The overarching theme in the literature featured in Table 1 is spontaneous action without preparation, described mainly as convergence of composition and performance (nine papers in Table 1, including Baker et al. 2003; Crossan et al. 2005), unfolding (five papers, including Barrett 2000; McKnight and Bontis 2002), and emergence (four papers, including Hutchins 1991; Lockford and Pelias 2004). Some authors also define spontaneous action without preparation as extemporaneousness (e.g., Moorman and Miner 1995; Schuller 1968), immediacy (e.g., Weick 1993b, 2001; Lockford and Pelias 2004); quickness (e.g., Meyer 1998; Holbrook 2007); and real-time formulation and implementation (e.g., Perry 1991; Pasmore 1998). Others see individual reflection in or on action as a key element of responsive action (Schön 1987; Yanow and Tsoukas 2009). Closely related recurring concepts are spontaneity (twelve papers in Table 1), bricolage (six), intuition (four), and “adhockery” (two). In what follows, we define OI as the conception of unhindered action as it unfolds, by an organization or its members, often (yet not exclusively) in response to an unexpected interruption or change of activity. This definition foregoes additional qualifiers, since its austerity already obliges organizations and organizational actor(s) to act extemporaneously, spontaneously, intuitively and ad hoc in an emergent manner. The convergence of planning and action is also not used in this definition, since it seems to imply that any rapid decision- making, due to a degree of convergence, is improvisational. Yet, improvisation is not deciding just before acting but whilst acting. OI is generally assumed to take place first and foremost in organizations that tolerate failure (Cunha et al. 2009; Sonenshein forthcoming), have a working environment that supports improvised work (Leybourne 2010b), and have minimal resistance to change (Leybourne 2006). In some instances, improvisation may also 7 act as a political statement and serve a political agenda within the organization (Janos and Rich 1994). Rationale for Organizational Improvisation Weick notes that: “even if organizations are capable of improvising, it is not clear they need to do it” (2001: 301). Unforeseen circumstances are not a necessary condition for improvisation: jazz bands, for instance, or newly formed start-up teams fully expect improvisation. Even so, OI is often triggered by some unexpected event that requires immediate action and cannot be addressed using pre-approved, “safe” routines and solutions (Hatch 1997; Moorman and Miner 1998b; Weick 1993a). Dealing with the unforeseen may involve removing barriers to instinctual ideas, or building new ideas out of nothing. The unforeseen may come either from outside or within the organization. From outside the organization, the external environment may become so complex as to render planning unfeasible (Cunha et al. 1999) or counterproductive (Mintzberg 1994). The organization may also face market and technological turbulence (Akgun et al. 2007; Nunez and Lynn 2012; Pavlou and El Sawy 2010), or an unpredictable environmental shock (Chelariu et al. 2002; Crossan et al. 1996) or crisis (ten Brinke et al. 2010; Mendonca and Wallace 2004; Webb and Chevreau 2006). Such disruptions may compel the organization to improvise (Crossan and Sorrenti 2002) and to proactively train itself to improvise crisis procedures – for instance, in the context of mental simulations of initial responses to aircraft accidents (Barreto and Ribeiro 2012) or in response to the situated practices and problems of users implementing large-scale Enterprise Resource Planning (ERP) change (Charles and Dawson 2011). OI scholars generally agree that external triggers for improvisation proliferate as markets become more dynamic. From within the organization, the impulse to improvise may come about when the enactment of a new vision requires emergent changes (Mintzberg and McHugh 1985) that may be addressed by improvisation (Crossan et al. 1996; Perry 1991). As the organization deliberately develops safer contexts for improvisational actions, skunk-works teams may also work on specific projects in unconventional ways with an aim to developing them fast and with minimal managerial interference (Janos and Rich 1994; Moorman and Miner 1998a). By improvising, an organization seeks to gain longer-term benefits beyond the situation at hand. It may also break free from flawed mental models of itself and its environment susceptible of preventing it from predicting otherwise foreseeable changes (Cunha et al. 1999; Senge 1990). Improvisation is commonly seen to promote greater organizational flexibility (Cunha et al. 8 1999). Even so, Barrett’s (1998) discussion of “provocative competence” mostly concerns top-down provocation, and Berniker states that: “The task of the managers' jazz combo is to make music. The role of employees is to dance” (1998: 583). Improvisation brings high autonomy in the context of clear rules (Cunha et al. 2003). It may help the organization to learn to improvise better (Crossan et al. 1996; Chelariu et al. 2002), to innovate (Vera and Crossan 2005), to explore new solutions (March 1991), or to perform certain activities better through routinizing successful improvisations (Miner et al. 1997; Ferriani et al. 2011). Team improvisation also positively impacts the effectiveness of new product development processes (Akgun and Lynn 2002). Coupled with unlearning and in conditions of environmental turbulence, it facilitates team flexibility, learning and new knowledge development (Chelariu et al. 2002). Ultimately, team improvisation influences new product success (Akgun et al. 2007). Among employees, OI may lead to higher motivation (Eisenberg 1990), to feelings of success (Eisenberg 1990) or to stronger teams (Powers 1981). In increasingly inter-connected organizational ecosystems, acting in inter- firm networks that do not have a single leader also often involves OI (Pavlovich 2003). OI however has a dark side, and may not be significantly associated with satisfactory project outcomes (Leybourne and Sadler-Smith 2006). Thus, entrepreneurs who are avid improvisers and confident in their ability to succeed may develop new ventures with record sales growth, but they also seem to be the least satisfied with their work (Hmieleski and Corbett 2008). This result may be due to a tendency to spread themselves too thinly (Baker and Nelson 2005), and points to the risk of stress and psychological burnout associated with excessive improvisation (Hatch 1999). By going against the idea that improvisers find themselves in the enhanced mental state of “flow” (Csikszentmihalyi 1990) or “groove” (Barrett 1998, 2000) when improvising, this result may also reveal limitations of the theatre and jazz metaphors of OI (Hmieleski and Corbett 2008). The effects of improvisation on financial performance may be delayed in time (Bergh and Lim 2008), and improvisation may even have negative organizational consequences (Aram and Walochick 1996). Organizations may over-eagerly generalize successfully improvised solutions to wrong contexts (Kamoche and Cunha 2001), and neglect planning and preparation by over-legitimizing OI (Eisenberg 1990). Excessive improvisation and improvisation without clear rules and boundaries may lead to lack of focus and unwanted variation in the delivery of products or services (Cunha et al. 2009: 666; Hatch 1999). Tackling every challenge with an ad hoc improvisational task force may also hinder the development of experience-based teams (Kamoche and Cunha 2001; Weick 1998). 9 Ultimately, organizations consistently struggle to achieve a balance between countervailing forces of organizational efficiency (control) and adaptability (improvisation) (Ciborra 1999; eybourne 2010a; unez and ynn 2012; Schrey gg and Sydow 2010). Large-scale distributed systems specifically deal with the improvisational paradoxes of learning (learned improvisation and reflective spontaneity), organizing (planned agility and structured chaos), and belonging (collective individuality and anxious confidence) through the “enacted emergence” of collective agility (Zheng et al. 2011). On balance, researchers tend to emphasize positive over negative outcomes (Magni et al. 2008; Vera and Crossan 2005), and OI may show positive effects in the right context (Vera and Crossan 2005). Yet, improvisation is inherently neither positive nor negative (Baker, Miner and Eesley 2003; Crossan et al. 2005; Magni et al. 2009; Miner et al. 2001). Metaphors and Empirical Studies of Organizational Improvisation Since improvisation in organizations involves a number of human actors dependent on complex local circumstances, individual instances of OI are hard to model. Improvisation is generally more openly discussed and documented in the arts (including music, theatre and dance) than in business, and using metaphors derived from these sectors allows researchers to draw meaning from alternative sources. A metaphor is a “comparative figure of speech […] through which humans create meaning by using one element of experience to understand another” (Morgan 1980, 1998: 4). As an “invitation to see the world” (Barrett and Cooperrider 1990: 222), a metaphor presents an alternative social reality (Tsoukas 1993), and connects the lay and scientific discourses (Tsoukas 1991). It also offers “an epistemologically valid approach to making sense of organizations” (McCourt 1997: 511). Metaphors used to explain improvisation include conversation (e.g., Berliner 1994; Hatch 1998; Ramos 1978; Weick 1998), problem solving (Bernstein 2000; Ramalho and Ganascia 1994), games (Hudak and Berger 1995), stories (Tanenbaum and Tanenbaum 2008), and role theory, Indian music and music therapy (Kamoche et al. 2000; 2003). Improvisational theatre is the second-most common metaphor in OI (Crossan 1997, 1998; Crossan et al. 1996; Gagnon et al. 2012; Gibb 2004; Koppett 2002; McKnight and Bontis 2002; Meyer 2005; Singh and Sonnenburg 2012; Vera and Crossan 2004, 2005; Vitug and Kleiner 2007; Weick 1993a). As theatrical interaction happens primarily with words, its lower level of abstraction generally makes theatrical improvisation more intelligible to people than musical improvisation. 10