Four key organizational paradoxes of AMbidexterity
Parsons (1971) argues that any enduring social system—including organizations—must satisfy four functional imperatives: Latency, Integration, Goal attainment, and Adaptation. He argues further that in complex social systems, these functions are typically distributed across differentiated subsystems. Within organizations, the key subsystems addressing the four function are, respectively: the fiduciary systems that foster internalized values shared across the organization; the interactional norms that set behavioral expectations for relationships among people playing specialized roles; the authority structure of leadership and reporting relationships that allows it to formulate and pursue goals; and the capabilities that allow the organization to adapt to the external environment (Heckscher, 2009).
When there is strong alignment within and across these four subsystems, the behaviors that people expect of each other are supported by shared values and norms, as well as by the authority structure and capabilities mobilized by the organization. Under such conditions, activity can be effectively coordinated and the organizational form is sustained. In contrast, where there is misalignment, coordination is impaired and the organizational form is unstable and cannot reproduce itself. We can use this four-system framework to synthesize the results of prior research on the organizational paradoxes of ambidexterity.
(We should note that Parsons’ is not the only framework we could use for the purposes of the theory we aim to develop in this article; but it has the advantages of theoretical depth, generality, and parsimony. Moreover, its key constructs overlap with others in the organization theory tool-kit: Latency/values corresponds to what much organizational research calls “culture;” Integration/norms overlaps much of the conceptual territory of “organizational climate;” Goal-attainment/authority and Adaptation/capabilities are widely acknowledged as key dimensions of organization design (Galbraith, 2002; Jones, 2012). Andriopoulos and Lewis (2009) offer an alternative to this interacting-subsystems model with their inductively derived nested-levels model, which differentiates strategic intent, customer relations, and personal drivers.)
Let us rapidly review prior research on each subsystem and identify the key organizational paradoxes associated with ambidexterity and the challenges facing the DA synthesis.
Values: Creativity versus efficiency. Much of the discussion of ambidexterity focuses on the challenge of reconciling the distinctive values associated with exploration and exploitation. Andriopoulos and Lewis (2009) describe it as the tension between the values of creativity and discipline at the personal level and between breakthrough and profitability goals at the strategic level. Gotsi et al. (2010) articulate nicely the paradoxical challenges posed for designers, who must attend simultaneously to the creativity and the cost of their new product designs, and the associated challenges for the organization, which must tailor its selection and socialization efforts to instill the requisite paradoxical values and identities.
Norms: Organic versus mechanistic. The paradox of combining organic and mechanistic relational norms is a common theme in prior research on ambidexterity, where exploration is associated with organic norms and exploitation with mechanistic ones. Among the various dimensions of the organic/mechanistic contrast (Burns & Stalker, 1961), research on ambidexterity has focused on formalization and communication patterns. In terms of formalization, Tushman and O’Reilly (1996) contrast the informality required for exploration and the formality that supports exploitation, and Jansen et al. (2006), Mom et al. (2009), and Rogan and Mors (2014) argue that formalization impedes exploration and facilitates exploitation. Regarding communication patterns, Mom et al. (2009) hypothesize that a broad network of personal connections (a key feature of the organic type) would support ambidexterity, in contrast with the more balkanized networks of the mechanistic type that would better support exploitation. Rogan and Mors (2014) make a similar argument concerning the intra-organizational ties likely to support managers’ ambidextrous behavior. Mom et al. (2007) argue that bottom-up and horizontal knowledge flows (characteristic of organic systems) are associated with exploration and that top-down flows (mechanistic) are associated with exploitation. Andriopoulos and Lewis (2009) argue that loose coupling communication patterns are required for exploration and tight coupling patterns for exploitation.
Authority: Individual autonomy versus centralized control. The literature on innovation has frequently noted that creativity is higher where centralized control is weaker (Damanpour, 1991; McGrath, 2001) and where individual autonomy and the associated intrinsic motivation is stronger (Amabile, 1983). Conversely, exploitation and efficiency are best supported by high levels of centralized control (McGrath, 2001; Miles, Snow, Meyer, & Coleman Jr, 1978) and by low levels of autonomy (Ostroff & Schmitt, 1993). In the scholarship on ambidexterity, Jansen et al. (2006) point to the tension created by centralized control which supports exploitation but suppresses exploration.
Capabilities: Broad versus deep. Much of the literature on exploration and exploitation has noted the contrasting types of skills involved: exploration benefits from broad skills while exploitation benefits from deep specialization (Bierly & Daly, 2007; Jehn, Northcraft, & Neale, 1999; Leiponen & Helfat, 2010; Schildt, Maula, & Keil, 2005; Sidhu, Volberda, & Commandeur, 2004; Smith & Tushman, 2005). Capabilities are not only a matter of skills: these skills must be mobilized in order to activate their potentiality. This, in turn, implicates the compensation system, where there are important tensions between compensation policies that support the development and deployment of broad skills versus deeply specialized ones (Kretschmer & Puranam, 2008).