Chapter 13 - Main contributions and results synthesis

Our PhD journey had started from a literature review questioning models of ambidexterity at the light of the unknown. Innovation management puts an emphasis on generating novelties, practicing innovative design, and using methods who actively embrace the unknown. Since implementing an ambidextrous organization has become rather common - and it was rooted on the assumptions of (March 1991) model - academics have recently notified the risk that the separation between exploration and exploitation could bear numerous complexities requiring further qualitative studies (Birkinshaw and Gupta 2013; Gupta, Smith, and Shalley 2006; Benner and Tushman 2015; O’Reilly and Tushman 2013). Over the years and with developments in innovation management, the unknown has conquered the grounds of ambidexterity and its problem-based foundations.

We have addressed these open statements and calls for research by revealing several projects in a conglomerate of SMEs, Zodiac Aerospace. These cases were somehow at odds with predictions of ambidexterity models. Indeed, exploration projects, which actively engage with the unknown, stretch beyond what the non-mutual conditioning between exploration/exploitation is capable of explaining. We have shown that ambidexterity can effectively kill innovation as it enforces organization design conformity and a reductive adaptation to exploit explorations. In the case of ZA and its engineering and marketing conditions, we explain such phenomenon by product/organization design fixations and interdependencies that are partially ignored and left unmanaged.

Consequently, we have proposed a model revisiting the foundations of exploration/exploitation. With the help of design theories staging generative processes, we were able to address the design of decisions, leading to decision-making and coordination of collective action. Decisional ambidexterity relies on the mutual conditioning between exploration and exploitation to generate alternatives and states of nature laying the ground work for interactions on: the nature of decisions, and models of thought ruling decisions, engineering and organization design. Working out organization/product design fixations and interdependencies, with the dynamic mirroring of engineering, provides additional knowledge to elaborate on the potential decisions framing collective action.

The ambidexterity re-developed in our model is placed back at a behavioural level with decision-design seen as a naturally managed and interactive process perhaps even more salient than actual decision-making (Christiansen and Varnes 2007). We were able to reinterpret the management anomalies of ZA with a better understanding of organization design fixations and interdependencies that were more or less managed and targeted in the management of exploration project.

Finally, two other projects and the study of a management device (Multi-BU Committee) enabled testing the model. We thus confirmed the value of previously hidden variables and offered new predictions. We then had opportunities to test in vivo decisional ambidexterity with several heuristics and practices with an exploration project. Moreover, we formalized some of them in the quality management system procedures for Innovation and Development. These practices were also validated and complemented by interactions with the community of R&T Managers.

All in all, the last stretch of the PhD journey allowed to touch upon the organizational metabolism and how decisional ambidexterity could sustain the associated generative interactions around decisions and organizational change.

Decisional ambidexterity for exploration project management

Our modelling effort on ambidexterity an its organization started anew with the foundational work of James March (March 1991). His original intent for organizational learning and adaptation was based on problem-solving and the bounded rationality of agents exploiting routines or exploring new ones. The balance of both regimes was a starting point for a rich literature on ambidexterity in organizational studies, strategic management as well as innovation.

Over time, the construct was transposed to different research fields in management and prescribed in practice. However, innovation management kept developing whilst showcasing generative processes and a keen interest to managing the unknown. The valuable literature review of (Wilden et al. 2018) gives an overview of the trajectory of the ambidexterity construct. Several key academics have also made their own assessment on their contributions to the field (Birkinshaw and Gupta 2013; Gupta, Smith, and Shalley 2006; O’Reilly and Tushman 2013; Benner and Tushman 2015). They all come to the conclusion that the relationship between exploration and exploitation requires deeper qualitative studies to better understand the implementation of ambidexterity. Its decision-making and managerial capabilities have remained blurry. Placing exploration and exploitation on a same continuum or not could be understood in different ways depending on units of analysis (Gupta, Smith, and Shalley 2006). Furthermore, innovation was seen as a potential driver shifting the performance of exploration and/or exploitation.

At the light of these calls for research, we have studied how the models of ambidexterity relying on the assumption of the non-mutual conditioning between exploration and exploitation would behave in the unknown. The idea was to test whether or not this strong hypothesis based on problem-solving would hold true in the unknown. Several units of analysis were based on adaptation or interactionism provided frameworks for management.

The prospects of exploration project management (Lenfle 2008) was a good candidate to reconnect with innovation management and the increasing projectification. Managing the unknown is one of its main drivers and given an exploitation regime ruled by a project-based organization, we identified in the literature potential shortcomings for organizational learning (Lenfle 2016) and change (Pollack 2017; Hornstein 2015).

Given the heavily constrained environment of Zodiac Aerospace, this conglomerate of SMEs offered a rich research field to question the validity domain of ambidexterity in the unknown. Several exploration projects and practices were studied revealing anomalies and opportunities to experiment with a reconciliating model of ambidexterity.

Avoiding killing innovation in a constrained environment

We have shown that ambidexterity, with its underlying model of non-mutual conditioning between exploration and exploitation, can actually kill innovation. The cases reported by Zodiac Aerospace present a wide range of exploration projects, relying on different generative processes. Despite having the a priori required elements to transition these projects, go to market or have business units take over initiatives, they would be considered as failures with no tomorrow for exploitation, learning, adaptation, competitive advantage. Some would only score a few partial points.

These anomalies of ambidexterity can only be understood by reconsidering the frontier of problems and decisions in the unknown. Design theories and reasoning bring a new language for decision-making, problem-formulation and problem-solving. The model decisional ambidexterity provides a new understanding of generative processes, environment cognition, organization design, as well as collective action. It allows tracking decisions and their design along a project’s trajectory.

We have illustrated that an exploration project can succeed in developing a new product, in a new market, and have several business units taking a stake in the innovation process to regenerate their organizations. A careful work of designing decisions and understanding what makes the decision come true is required through design of the states of nature: market, environmental constraints, norms, standards, users’ acceptance, etc. These heuristics of decision-designers provide them with means of action enabling the materialization of their innovation effort.

Hacking philosophy: the decision-designer

The philosophy of designing decisions can be related to hacking. Indeed, applying design theories to decision theories offers new degrees of freedom on a highly normative model of action. The decision-designer does what everyone does: we choose from readily accessible alternatives induced by path-dependency and fixation effects, and if we have or imagine the consequences, we may be willing to generate new alternatives and scenarios where we would be as satisfied or even more satisfied.

Of course, this satisfactory level overcomes and twists bounded rationality (Simon 1955). We would, in fact, extend our rationality by engaging in actions embodying what is necessary to realize designed decisions. Instead of choosing A or B given C or D, one can optimize consequences given a certain belief, but one can also find a generic solution fitting all outcomes and beliefs. Or one could even totally aim for different comparable alternatives in another scenario. One could finally aim for less comparable alternatives as we mix up other value networks in different scenarios.

Optimization, genericity, and wishful decision by hacking one or all states of nature sets different courses of action for project management. The hacking is performed given parameters identified in the originally given exploitative situation. It helps rooting, nurturing and making sense of the exploration stretch. Conditioning is no longer seen as reductive, it is instead expansive when designing.

These statements were tested and validated by practices, discussions, and accepted in process management procedures. But, they still raise numerous questions on a synthetic technology of organizing to channel interactions on the design of decisions and commitment. The need is perfectly natural: we are tricking probability theory with a conditioning that in reality implies hacking. Uncertainty is increased but we counter it with means of action reconfiguring relationships and knowledge.

It is rather stimulating to think of our modelling effort as a means to reconnect the rational augmentation hoped by Herbert Simon with the foolishness and Romantic views of leadership by James March. A conscious and rational model of thought led by design theory and reasoning could occupy this grey zone. In other words, our research results targeted the relationship between rationality and foolishness in (March 2006, 209):

It is argued that the link between rationality and conventional knowledge keeps rational technologies reliable but inhibits creative imagination. This characterization seems plausible, but it probably underestimates the potential contribution of rational technologies to foolishness and radical visions.

In this PhD thesis, we developed this bone of contention with the mutual conditioning between exploration and exploitation. Decisional ambidexterity, with its decision-design and decision-making in the unknown, offers a means to rationalize and sustain foolishness in a way. It cools down a romantic view of decision-making and collective action in organizations, but we offer in return manageable risks and opportunities sustaining imagination and creativity in markets (Bronk 2009).

Decisional ambidexterity for organizational ambidexterity

Our research made a specific focus on the encapsulated fashion of project management. It was a means to discuss managing the unknown with a clear view of how collective action was coordinated and the assumptions made from an organizational standpoint. Project management was demonstrated to be historically linked to innovation management (Davies, Manning, and Söderlund 2018; Lenfle and Loch 2010, 2017). It gave us a baseline to discuss the relationship between project-based and temporary organization (Turner 2009; Söderlund and Müller 2014; Sydow, Lindkvist, and Defillippi 2004; Bakker et al. 2016) and exploration projects (Lenfle 2016).

The “deconstruction” path taken from the model of exploration/exploitation (March 1991), who took us to discussions on organizational studies, leadership and strategic management, is now taken backwards to reconnect the encapsulated way of separating exploration and exploitation with organizational learning and adaptation. The new perspectives on organization design in (Van de Ven, Ganco, and Hinings 2013; Puranam 2012; Avital and Te’eni 2009) and change management (Pollack 2017; Hornstein 2015) bring additional perspectives on where projects could interfere. The cruciality of epistemic interdependencies is flagged as they allow reconnecting with engineering design and its mirroring hypothesis (Colfer and Baldwin 2016).

We were interested in organizational learning and change prospects that tend to be regenerative. This specific feature is given by the heavily constrained environment of Zodiac Aerospace, with numerous design rules, standards, norms, and business units mirroring market segments with associated engineering for certified products. The ecosystem has rather settled after numerous acquisitions and consolidations. Business Units have their specific niches and playfields, and are grouped in divisions and branches for an Executive Committee. Creating spin-offs appeared out of the equation. It would force (re-)generative learning (Senge 1990) and mirrored organizational change to overcome core rigidities and incompetencies (Leonard-Barton 1992; Dougherty 1995).

Organizational learning articulated by exploration/exploitation mutual conditioning

We have shown cases where learning was kept isolated from exploitation because of non-mutual conditioning assumptions. Exploration projects had however played an active role in cultivating learning and opening new innovation fields. Some projects even relied on a clear hacking philosophy of exploitation regime with preferences reversal for instance.

Learning was also inhibited because of generative processes and project management not systematically taking on board organization design and environmental dimensions. Exploration would be conducted for an iso-exploitation regime and it would expect that operational organizations would learn from exploration during hand-over. Despite clever and prepared interactions for transitions, most projects would not articulate (enough) the exploration-exploitation mutual conditioning.

Two cases and an intervention allowed however to better identify learning mechanisms around problem-formulation, value space and open discussions on future decisions. Decisional ambidexterity model played a crucial role in forcing learning and shaping the unknown to enable decisions-making. Means of actions can be derived and a sense of direction is given for learning. Learning of engineering design rules and absorption of new knowledge engage decision-designers in the reconfiguration of organization’s knowledge management. This generative learning around decision-design would then require new relationships to be tied.

Organizational adaptation and change driven by unlocking fixations and interdependencies

The studied exploration projects were all potential emergent strategies from intra or inter BU perspectives. They addressed potential competitive advantage and adaptation to the environment. However, we have seen that organizational ambidexterity could potentially inhibit their prospects. Decisional ambidexterity model surfaces the criticality of organization design fixations and interdependencies to steer exploration projects. Otherwise, projects may float and disconnect from host organizations.

Two contrasted cases revealed the importance of problem formulation and managing value around what business units are capable of deciding given their engineering capabilities. Conceptualizing the mirroring hypothesis (Colfer and Baldwin 2016), through the design of decisions and organization design knowledge, forces to merge change management with project management. It encourages also to find means (e.g. prototypes, management tools, decisions committees, business models, etc.) bearing the reconfiguration of relationships and knowledge through novel designed decisions.

As we have shown in the researcher’s intervention and other contributions to Zodiac Aerospace, the categorization effort in concepts and designed decisions offer handles on interdependencies and organization design fixations. These channelled interactions with an effort on engineering models and mirrored organizations. They triggered side-projects and valued positively projects that would otherwise be simply considered deviant. We insist again on the fact that exploration is richer and more performant by referring to exploitation. It sustains an elasticity and plasticity in the conditioning between exploration/exploitation offering room for organizational metabolism to thrive.

Chapter synthesis: decisional ambidexterity, a model for strategic innovation management

References