An Agent-Based Model of Organizational Ambidexterity Decisions
Christine Chou
Department of International Business
National Dong Hwa University
No. 1, Sec. 2, Da Hsueh Rd.,
Shoufeng, Hualien 97401,Taiwan
Email:
Tel: 886-3-8633043
Fax: 886-3-8633040
Steven O. Kimbrough
University of Pennsylvania
The Wharton School
565 Jon M. Huntsman Hall
3730 Walnut Street
Philadelphia, PA 19104-6366 USA
Email:
Tel: 215-898-5133
Fax: 215-898-3664
Ambidexterity Decisions, 2009-10-01
ABSTRACT
Organizational ambidexterity, defined as the simultaneous pursuit of exploitation and exploration, has become a very important topic in the study of organizations. Yet, as a recent and comprehensive review notes, “Despite the rapidly expanding number of studies referring to organizational ambidexterity, empirical tests of the ambidexterity–performance relationship remain scarce.” The contribution of this paper is to introduce an agent-based model that captures many of the key aspects and tradeoffs, which have been identified in the literature, of the exploration–exploitation dilemma faced by firms. When data become available, the model can be calibrated to the data. Until then, the model serves as a useful tool for exploring hypotheses and revealing surprising aspects of the subject.
Key words: Organizational ambidexterity, Agent-based modeling, Exploitation-Exploration.
AN AGENT-BASED MODEL OF ORGANIZATIONAL AMBIDEXTERITY DECISIONS IN NEW PRODUCT DEVELOPEMENT
Introduction
Organizational ambidexterity, defined as the simultaneous pursuit by organizations of exploitation and exploration activities in product and service offerings, has become an important concept in organization and innovation management theory (See a literature review by Raisch and Birkinshaw, 2008). Previous studies offer conceptual or empirical evidence that organizational ambidexterity leads to firm performance with a focus more on internal organization structures or antecedents (Adler, Goldoftas & Levine, 1999; Beckman, 2006; Bradach, 1997; Gibson & Birkinshaw, 2004; Jansen et al., 2005, 2006; Jansen et al., 2009; Lubatkin et al., 2006; Smith & Tushman, 2005). However, as March’s (1991) influential study points out, returns to one organization are not strictly determined by the organization’s own performance but depend on the relative standings of the competitors. Some researchers incorporate the effect of the environment (Voss, Sirdeshmukh, and Voss, 2009; Cao, Gedajlovic, and Zhang, 2009), yet little has been known about whether the adoption of ambidexterity for a firm can always achieve competitive advantage despite the different strategies that the rival competitors adopt.
Consider, for example, the case in semiconductor industry where a standard product development roadmap, called International Technology Roadmap for Semiconductors (ITRS, 2008), is discussed and issued every year so that all firms in this industry can follow. As a manager we interviewed told us, ‘we often follow the roadmap and work close with our customers because there is a high consensus in this industry about what to produce and where we will go in at least 15 years.’ In such an environment, can a firm’s ambidextrous investment policy achieve competitive advantage? Or whether solely adopting an exploitative investment policy is already sufficient for a firm to keep profitable? Will the rapid change of customers’ needs moderate the relationship between ambidextrous decisions and firm performance? These questions are not fully investigated in previous studies but are important for theory and practice.
This study aims to examine these questions and extend the understanding of organizational ambidexterity in the new product development—a key factor to a firm’s performance and competitive advantage. To address these questions, specifically, this research echoes the call from Garcia (2005) that ‘ [L]ittle has been written in the new product development literature about the simulation technique agent-based modeling, which is a by-product of recent explorations into complex adaptive systems in other disciplines’ (2005: 380), and adopts an agent-based model as our analytical tool.
The contributions of this paper are three-fold. First, as Siggelkow & Levinthal (2003) note, the overall strategic emphasis of an organization is reflected in investments of resources in activities and processes that promote exploration or exploitation. Earlier studies suggest that due to conflicting administrative routines and managerial behavior, as well as limited organizational resources, firms need to make tradeoffs, and explicit or implicit decisions, between exploitation (for survival) and exploration (for prosperity) (Lubatkin, Simsek, Ling & Veiga, 2006; March, 1991; Wernerfelt & Montogomery, 1988). More recent literature, however, proposes that organizational ambidexterity is a more viable choice for a firm and this view has become established in the field of management and organizational studies (Gibson & Birkinshaw, 2004; He & Wong, 2004; Raisch & Birkinshaw, 2008; Tushman & O’Reilly, 1996). Although previous studies discussed many facets of organizational ambidexterity such as structural designs, contextual elements, as well as alliance formation. The issue of how managers make investment decisions regarding exploitative, exploratory, or even ambidextrous activities is seldom analyzed in details. For managers, the crucial issue of the daily work is to make decisions so as to achieve their performance goals. The analysis of organizational ambidexterity so far, however, has not specifically included this decision-making perspective. We analyze the tradeoffs under various decision-making scenarios. Our aim in this paper is to contribute to a more thorough understanding of this managerial decision problem.
Second, as mentioned earlier, the aspect of how a firm reacts to its competitors’ moves is often neglected in this stream of literature. While many studies acknowledged the importance of environment or industry, they either control the environmental effects or measure it using an aggregate index to represent the environment such as environmental threats (Voss, Sirdeshmukh, and Voss, 2009) or environmental munificence (Cao, Gedajlovic, and Zhang, 2009). What is not clear is how a firm actually makes its decision as a response to its competitors’ moves. That is, whether a firm’s competitor adopts different strategies affects the focal firm’s decision is not understood fully. This study aims to fill this knowledge gap to see how a firm behaves differently as a response to its rival competitors.
Third, this paper introduces and discusses an agent-based model (or ABM; North & Macal, 2007), Ambidextrous.nlogo, that captures many of the key aspects and tradeoffs, which have been identified in the literature, of the exploration–exploitation dilemma faced by firms in the context of organizational ambidexterity. Our model is public, it is extensible, and it can be used to gain insight into the general domain and into specific hypotheses. Because it is a model, its assumptions are explicit and accessible. In virtue of being a computational model, the assumptions are readily modified. The model, or rather its implementation in NetLogo as Ambidextrous.nlogo, is available as an applet. Through this simulation method we can estimate good policies for how and when the choice of ambidexterity can be made on favorable terms. This is an advantage of our methodology because simulation can provide multiple runs of experiments so that tests of different levels and combinations of organizational ambidexterity can be executed. This model can be used as a decision-making tool.
We turn now to a review of the essential background on organizational ambidexterity with a focus on new product development. Based on this background knowledge and interview insights, we describe a model that deals with some of the central issues in organizational ambidexterity, in particular investment decision issues under different conditions. Then, we describe the results of our model and discuss insights from the model and related literature. Finally, we provide discussion, implications and conclusions.
Organizational ambidexterity and performance
Duncan et al. (1976) first introduced the concept of organizational ambidexterity and asserted that a dual organizational structure is needed to initiate innovation and implement innovation. As a further step, March’s (1991) influential paper proposed exploitation and exploration as two different learning activities which should be pursued in balanced way but, at the same time, may compete for organizationally scarce resources. Based on March’s work, many scholars applied exploitation and exploration concepts into their research in different contexts such as organizational learning, technological innovation, organizational adaptation, strategic management, and organizational design (Ancona et al., 2001; Atuahene-Gima, 2005; Benner & Tushman, 2003; Burgelman, 2002; Gupta, et al., 2006; He & Wong, 2004; Katila & Ahuja, 2002; Levinthal & March, 1993; Raisch & Birkinshaw, 2008; Tushman & O'Reilly, 1996; Tushman & Smith, 2002). In spite of conflicting, or at least varying, definitions, the literature agrees that exploitation and exploration refer to very different learning activities within an organization. Exploitation includes activities associated with refinement, extension or improvement of current components, competences, and technologies; exploration includes activities including experimentation, innovation or a shift to different technological trajectory (Benner & Tushman, 2003; March, 1991). A combination or simultaneous pursuit of exploitation and exploration in an organization is defined as ambidexterity (He & Wong, 2004; Lubatkin et al., 2006; Tushman & O’Reilly, 1996). Studies including concepts such as ‘reconciling exploitation and exploration’, the ‘simultaneity of induced and autonomous strategy processes’, ‘synchronizing incremental and discontinuous innovation’, and ‘balancing search and stability’ belong to the same stream of literature (Raisch & Birkinshaw, 2008).
Because of the complex nature of organizational ambidexterity, the linkage between organizational ambidexterity and performance remains controversial. On one hand, following March's assertion that firms run the risk of being mediocre at both exploitation and exploration resulting from the inherent challenge of doing both, some scholars propose to pursue only one direction instead of both (Barney, 1991; Ghemawat & Ricart i Costa, 1993). Firms simultaneously pursuing both activities are likely becoming internally inconsistent so as to lead to inferior performance (Wernerfelt & Montgomery, 1988). The knowledge processes of exploitation and exploration are thought to be contradictory because they tap different administrative routines and managerial behaviors. As suggested by Lubatkin, et al. (2006), exploitation primarily involves learning from top-down process because managers are used to those organizational routines, while exploration involves a bottom-up process which managers are persuaded to abandon their old routines and make a change or innovative action.
On the other hand, Tushman & O’Reilly (1996) point out that firms capable of simultaneously pursuing exploitation and exploration are more likely to achieve superior performance than firms only emphasizing one. Levinthal & March (1993) also argue that ‘an organization that engages exclusively in exploration will ordinarily suffer from the fact it never gains the return of its knowledge …an organization that engages exclusively in exploitation will ordinarily suffer from obsolescence.’ In other words, the exclusive pursuit of exploration may end up with endless search efforts or R&D expenditure without appropriate return (Raisch & Birkinshaw, 2008). In contrast, the exclusive pursuit of exploitation may enhance short-term return but may lead the organization to being incapable of adapting to a new environment (Levitt et al., 1988). Firms should exploit existing competencies and explore new ones and these two activities are inseparable (Floyd & Lane, 2000). Besides these conceptual works, there are a few empirical studies regarding this topic. Knott (2002) found that exploitation and exploration coexist in Toyota’s product development, and concluded that these two activities are likely to be complementary. Katila & Ahuja (2002) found a positive interaction between exploitation and exploration on new product development but did not test their effects on firm performance. In the empirical work of He & Wong (2004), they found that the interaction between explorative and exploitive innovation strategies is positively related to sales growth rate, and the relative imbalance between explorative and exploitative innovation strategies is negatively related to sales growth rate at the firm level. Gibson & Birkinshaw (2004), focusing on the business unit level, found that the capacity to simultaneously achieve alignment and adaptability positively affects performance. Lubatkin et al. (2006) suggest that the joint pursuit of an exploitative and exploratory orientation positively affects performance in small- and medium-sized enterprises. Other empirical work, however, did not support the ambidexterity-performance linkage. Rather, it found that temporal cycling between exploitation and exploration has a positive effect on firm performance (Venkatraman, Lee, & Iyer, 2007). Proposing two dimensions of ambidexterity—balanced and combined and using Chinese high-tech firms, Cao et al. (2009) found the performance of adopting either dimension of ambidexterity depends on a firm’s size and environmental munificence. In addition, Rothaermel and Alexandre (2009) investigate the moderating role of absorptive capacity between internal-external technology mix and firm performance.
As for the methods adopted, previous literature exploring formal models pertaining to organization ambidexterity is thin. Siggelkow and co-workers have produced some excellent related work, based on implementations of Kauffman’s NK model framework with a focus on organizational structural designs (e.g., Rivkin, 2000; Rivkin & Siggelkow, 2003; Siggelkow & Levinthal, 2003; Siggelkow & Rivkin, 2006). See also Levinthal (1997) for an early treatment of NK modeling in the context of organizational studies. The aims of these papers are quite distinct from those of the present work, focusing instead on such matters as comparing regimes of information flow within an organization. The Tay & Lusch (2007) work is perhaps closest in outlook to the model reported here. It offers a complementary perspective, assessing how a population of firms will behave and imputing to them fuzzy-logic-based learning. Finally, Garcia (2005) offers some high-level observations on the promise of agent-based modeling in new product development and an introductory example of firm’s exploitative and exploratory activities. The result of the study found that organizations can still keep profitable even they do not balance exploitation and exploration activities and respond slowly to the threats and opportunities in the environment.
Despite the rapid increase in studies concerning ambidexterity, the empirical data for the ambidexterity-performance (Raisch & Birkinshaw, 2008), and the studies that examine the conditions under which ambidexterity leads to success are relatively scare (Raisch, Birkinshaw, Probst, and Tushman, 2009). Specifically, we did not find a model or study that discusses and includes the investment decisions made by managers as a main topic in previous literature. This may be due to the constraint of the methodology these studies adopted. In fact, an agent-based model can enhance our understanding of the process or patterns in an organization and, therefore, the question of how the investment decisions are made can be analyzed by this model. Furthermore, we make this model public so that managers can use this tool to assist their decisions. We now describe our model.