The value-rational organizational form as the context for distributed ambidexterity By

The Challenges Of Ambidexterity

Ambidexterity is the ability to combine exploration and exploitation. These terms can be understood as referring to goals, activities, abilities, or outcomes, depending on the context and the author’s emphasis. Synthesizing a large body of organizational research, March (1991) posited that exploration and exploitation represent two key tasks facing organizations. Exploration involves search, variation, experimentation, and discovery, while exploitation involves refinement, efficiency, selection, and implementation. Some scholars have interpreted exploitation as the absence of learning and innovation: we agree with Gupta et al. (2006) that this view is less fruitful than seeing exploitation as a distinct type of learning and innovation. The term ambidexterity has been used to designate the ability to combine different stages of the innovation process (Duncan, 1976), incremental and discontinuous innovation (Tushman & O'Reilly, 1996), efficiency and flexibility (Adler, Goldoftas, & Levine, 1999), alignment and adaptability (Gibson & Birkinshaw, 2004), and the ability to compete in mature and emerging markets (He & Wong, 2004). (For recent reviews of this literature, see Lavie et al., 2010; Nosella, Cantarello, & Filippini, 2012; O'Reilly & Tushman, 2013; Turner, Swart, & Maylor, 2013.) The common element here is the combination of qualitatively and radically different goals and activities (Lavie et al., 2010).

The organizational paradox of ambidexterity: An initial characterization

The idea of and evidence for ambidexterity force us to reexamine a long-established proposition in organizational theory asserting that exploration and exploitation are deeply incompatible goals that must “trade-off” against each other, and that ambidexterity is therefore a paradoxical notion. It is obvious that investments in exploration and exploitation activities will be in a trade-off relation insofar as they must compete for scarce organizational resources; beyond that, however, there is considerable debate as to whether and how organizations can mitigate that trade-off and can successfully combine efforts towards both goals.

In the strongest version of the trade-off view, ambidexterity is a dangerous chimera, because exploration and exploitation require such radically contrasting behavior patterns and organizational forms—adhocracies versus bureaucracies (Mintzberg, 1979). In this view, organizations should focus on one goal or the other: if they attempt both, they will achieve only mediocrity in both and risk getting “stuck in the middle” (Abernathy, 1978; Holmqvist, 2009; Porter, 1985; Thompson, 1967).

An accumulating body of empirical research challenges this view. As noted in the Introduction above, a growing body of research finds that the ambidextrous pursuit of both goals can, at least in some industry contexts, improve business performance. Moreover, most of these studies find a positive rather than negative correlation between exploration and exploitation intensity observed across firms, suggesting that many organizations are able to mitigate the trade-off to at least some extent (Bierly & Daly, 2007; Jansen, Van Den Bosch, & Volberda, 2006; Katila & Ahuja, 2002; Knott, 2002; Kristal, Huang, & Roth, 2010; Lin et al., 2013; Rothhaermel & Deeds, 2004).

These empirical findings are consistent with March’s argument. March acknowledged tensions between the two goals, but did not see the trade-off as so severe as to necessitate a dichotomous strategic choice. On the contrary, he argued that the firm’s overall performance would be improved if it resisted the tendency to drift toward exploitation and instead pursued both types of activity.

The DA approach to ambidexterity

Ambidexterity can be pursued in several different ways, and to understand DA’s distinctive organizational challenges, it is helpful to situate DA alongside these other approaches. The various approaches reflect the three generic strategies through which paradoxes can be resolved rather than merely accepted (as presented by Poole and Van de Ven (1989)): temporal separation, spatial separation, and synthesis.

Organizations can leverage temporal separation by switching, or alternating over time, between the two types of activity (Chen & Katila, 2008; Duncan, 1976; Gupta et al., 2006; Lavie & Rosenkopf, 2006). This might happen more regularly and deliberately for a whole organization (Siggelkow & Levinthal, 2003) or for a product development project team as it progresses through project phases (Andriopoulos & Lewis, 2009; Brown & Eisenhardt, 1997; Venkatraman, Lee, & Iyer, 2007). It may also happen in the more intermittent and disruptive form of punctuated equilibrium (Lant & Mezias, 1992; Tushman & Romanelli, 1985).

The literature has identified three main approaches that leverage spatial separation. In the functional approach, the output of some functional sub-units specializing in exploration (e.g. R&D, product engineering, or marketing) becomes input for other, exploitation-oriented functional sub-units (e.g. the manufacturing or operations) (Adler, 1995; Galbraith, 1973; Jansen, Tempelaar, Van den Bosch, & Volberda, 2009; Jelinek & Schoonhoven, 1990; Lawrence & Lorsch, 1967). In the network approach, the focal organization specializes in either exploration or exploitation, and relies on network partners for access to the other type of capability; alternatively, the focal firm orchestrates a network of partners some of which focus on exploration while others focus on exploitation (Arikan, 2009; Boudreau & Lakhani, 2009; Grabher, 2002; Hagel & Brown, 2005; Kauppila, 2010; Kristal et al., 2010; Lavie & Rosenkopf, 2006; Miles, Snow, Mathews, Miles, & Coleman Jr, 1997; Rothaermel & Alexandre, 2009; Simsek, 2009). In the structural approach, sub-units are dedicated to business lines that focus on either exploration or exploitation types of goals; each of these business-line sub-units is largely self-sufficient in its array of dedicated functional resources; these sub-units collaborate directly only in targeted, tactical activities; and the top management team plays the integrative role (Benner & Tushman, 2003; Tushman & O'Reilly, 2002; Tushman et al., 2010). In all these approaches, the top-management team plays a key role (Cao, Simsek, & Zhang, 2010; Jansen et al., 2009; Lubatkin et al., 2006).

In contrast with these separation-based approaches, we focus on a synthesis path to ambidexterity, where individuals, teams and subunits each continuously combine exploration and exploitation rather than allocating each of the two activities to specialized roles or to different time periods (Andriopoulos & Lewis, 2009; Gibson & Birkinshaw, 2004). This distributed ambidexterity (DA) approach relies an intra-organizational context that encourages and facilitates such a dual focus in individual behaviors and enables continual synthesis at all levels; as a result, this was originally called the “contextual” approach by Gibson and Birkinshaw. We propose to re-name the approach “distributed” so as to focus on the distinctive locus of ambidexterity in this approach: the label “contextual” is not very evocative and the concept as originally developed confuses the focal behavior with the context designed to induce it (Birkinshaw, 2014). The studies cited in the Introduction above provide compelling evidence that DA can lead to significant performance benefits, at least in some settings. This distributed approach to ambidexterity has been found in supply chain dyads (Im & Rai, 2008), manufacturing operations (Adler et al., 1999), R&D functions (Güttel & Konlechner, 2009; McCarthy & Gordon, 2011), and product design groups (Andriopoulos & Lewis, 2009; Gotsi et al., 2010).

What organizational form can support the DA synthesis of exploration and exploitation? The first effort to respond to this question was Gibson and Birkinshaw (2004), who argued that the organizational challenges of DA can be met by creating a climate that combines trust and social support with discipline and stretch (see also Allen, 2013). In positing these two dimensions, Gibson and Birkinshaw were building on Ghoshal and Barlett’s (1997) account of the organizational climate most likely to support innovation. And in doing so, they were reprising the Ohio State “consideration” and “initiating structure” characterization of the two dimensions of effective leadership (Stogdill & Bass, 1981), and restating Blake and Mouton’s “managerial grid” model of this same theory (Blake & Mouton, 1964). (This two-factor model fell out of favor many years ago, but more recent research shows that it holds up well to meta-analysis — see Judge, Piccolo, & Ilies, 2004.)

We aim to push the causal chain further back, to explore the organizational form that could foster and support such a climate. To date, such research has proceeded inductively (most notably Andriopoulos & Lewis, 2009; Gotsi et al., 2010; Güttel & Konlechner, 2009; Heracleous & Wirtz, 2014; Volery et al., 2013). The first step is to identify the specific organizational paradoxes that are posed by ambidexterity: we do this in the following section, by mobilizing Parsons’ theory. The sections after that will build on Weber’s theory of value-rationality to show how these paradoxes can be resolved to achieve DA.