Chapter 14 - Discussion and perspectives for decisional ambidexterity

In this final chapter, we open up our conclusion for discussions on our model of decisional ambidexterity and research perspectives. We have so far reworked the conditioning between exploration and exploitation. This modulation allows reconnecting models of ambidexterity with managing the unknown. It holds true that exploration and exploitation should be mutually conditioned in a constrained environment with rules, norms, regulations, path dependency sensitive to potential threats of disruption. It avoids deteriorating both regimes. It actually supports a sustainable generativity, organizational change and design.

The developed model of decisional ambidexterity tries to reconnect decision theories with design theories. We have shown how the unknown could be integrated. With the rising projectification of firms and the economy, exploration projects heavily discuss organizational learning and adaptation. Organizational ambidexterity and design implemented by top management should then be reviewed at the light of meso-level analyses at work in projects driven by decision-design and the unknown.

However, designing decisions, from the standpoint of cognitive psychology and even behavioural economics, can be seen as paradoxical. For decision-making on one hand, research agendas have studied and modelled the influence of biases and heuristics (Tversky and Kahneman 1974; Kahneman and Tversky 1979; Kahneman 2011). Moreover, different dispositional variables, such as stress or positive emotions and other conditioning settings can be added to measure their influence (Cassotti et al. 2012).

On the other hand, psychology of creativity has been driven by how individuals and groups generate ideas. Instead of decision biases, they would discuss fixation effects and its inhibitory control (Cassotti et al. 2016) on generativity.

These two research streams stand on different grounds. They rely on different perspectives of rationality and use different experimental protocols. How could we provide a general model of generation of alternatives, their selection and commitment? How should experiments be conducted to grasp design during decision tasks? Like decisional ambidexterity providing a “unifying” framework, we would wonder about a unified theory of generativity and choice. Or at least, we are curious about pushing both models towards a common asymptote where uncertainty becomes unknown and where the unknown becomes uncertain.

We have considered hacking practices for decision-designers. All silent designers in organizations (Dumas and Mintzberg 1989, 1991) could then offer new means of collective action for corporate entrepreneurship and radical innovation (O’Connor 2016). Not only is institutionalizing innovation necessary to recognize the phenomena and avoiding ineffective risk taking, but it can also be sustained with technologies of organizing. The importance of valuation tools and methods when designing decisions would be critical to manage exploration projects but also for organizations.

The idea of an Organizational Design Thinking would be quite stimulating. Common design thinking methodology will be fixated by users and a given ontology of operations in the ecosystem. Tackling users pains and building empathy helps generating concepts where users can easily adopt, interact with and open new perspectives.

Could we imagine a decision-design methodology where the purpose is to promote concepts that systematically address what will federate organization around a common purpose? It would be a sort of positive no-man’s land that drives learning and change, instead of fighting for primacy or against inertia.

The valuation methods would be critical to target the cracks between organizations and to generate such concepts targeting organizational constraints and drivers. The Lower Deck project would appear as a suitable example by contrast with the Connected Cabin (see chapter).

Finally, our concern for valuation tools and organizational design thinking brings us to discuss a renewed organization design: generative organization design. In the literature, we have pointed at several authors inviting to rethink this research topic.

The modelling effort concretized into decisional ambidexterity took organization design as knowledge to generate new concepts and keep an eye on mirroring dynamics. There would not be any target organization design since we focus on engineering design and decisions. For open innovation, problem formulation and solving will be driven through a decision-design practice. Partnering with users, suppliers, integrators and even competitors would have a different twist. Separate and interconnected agendas could support different decision categories engaged with the environment.

If designed decisions drive organization re-design, what would be the managerial action in such context? What is the governance of such distributed and generative innovation function?

---

A bridge between decision and generativity

When facing a decision task, alternatives are given to the decision-maker. However, we have brought forward that in the course of action, rational choice of theory has several limitations (Le Glatin, Le Masson, and Weil 2017). One could be irrational with respect to several paradoxes that can be embraced in prospect theory for instance (Kahneman and Tversky 1979). Reversing preferences in time or willingly complexifies the understanding of decision-making (Slovic 1995; Camerer, Loewenstein, and Rabin 2004; Machina 1989; Chabris, Laibson, and Schuldt 2006).

For decision theories stemming and extending the works of Abraham Wald (Wald 1949; Savage 1954; Giocoli 2013; Fourcade and Khurana 2013), inconsistency or intransitivity are violations of rationality. A simple way of accepting such violation is the integration of new information in time. Otherwise, considering that the decision-maker generates new decisions and actions, to gain information given a bounded problem, unsettles rational theories of choice.

Bounding and unbounding problems

Rational theories of choice, operational research and refinements brought by cognitive psychology have taken a rationally bounded perspective on agents. A problem would be predefined, the agent will struggle to oversee the full landscape and will follow the path of least resistance and be biased. However, the perspectives from creativity theory or generativity theory (Epstein et al. 1984; Epstein 1990, 1999) tend to unbound problems or at least reformulate them. New decisions can be designed and engaged given problem parameters; it was the case for pigeons solving their problem. In a more ill-defined setting, design can not only optimize, find generic solutions or recombine, it can also revamp altogether the problem at stake.

The question for us would be to understand what are the parameters that would trigger the need to design instead of simply deciding. Is it comparable to risk aversion/seeking behaviours? The major difference is when designing, an active interaction is engaged with the environment instead of a passive one expecting for the states of nature to reveal. The decision-designer would hope to generate new knowledge and tie new relationships.

The issue is with the very construction of probabilities. Imprecision and fuzzy sets could be a first patch. Unfortunately, they do not model the reconfiguration and complete revaluation of alternatives and states of nature. Even if a theoretical effort is required for theories of behaviour, extending for instance prospect theory with design reasoning, we could also turn towards experimentation’s means.

Differences in experimental protocols

Our interest would be in the inhibitors and triggers for designing decisions. Some situations could be intolerable for decision-making, perhaps they could support the generation of other alternatives countering a priori consequences. Moral judgement and ethical behaviour could be an exciting approach to unlock the deliberate will to design decisions instead of deciding.

For instance, the trolley problem extensively studied by Judith Thomson (Thomson 1976, 1985, 2008) could be an interesting baseline to discuss the design of new alternatives and what could trigger it. This problem also concerns autonomous driving (Bonnefon, Shariff, and Rahwan 2016). The matter is given in an iso-environment for the car, whereas it could be perhaps even more stimulating to think of autonomous cars actively protecting pedestrians (i.e. not just with shock absorbers, sensors to detect the variety of obstacles and pedestrians and deciding).

The difficulty in approaching experimentation at the frontier of decision and creativity is the protocol and briefs given to the decider/designer. It may totally inhibit one or the other whereas we would be more curious about the comings and goings between the two. Since the Centre de Gestion Scientifique has a partnership with the LabPsyDé of Descartes University, we already had several discussions on these topics with Anaëlle Camarda and Matthieu Cassotti. We had some ideas for moral judgement, studying the anchoring effects of Kahneman and Tversky with respect to design, but also biases for negative conditioning and asking for information like in Wason’s experiment (Wason 1960; Houdé and Moutier 1996).

Reexplaining behaviours: biases, heuristics and hacking

Furthermore, designing decisions in the unknown encourages the decider to interact and interfere with his beliefs and states of nature. Biases and heuristics could contribute to the dominant design and fixation effects. However, decision hacking with genericity and wishful decisions reveal another behaviour providing a more positive and pro-active feature of rationality.

With the increasing number of practices promoted in start-up environment, we wonder how we could study in more detail the behaviours associated with effectuation (Sarasvathy 2001; Agogué, Lundqvist, and Middleton 2015), lean start-up (Ries 2011; Blank 2013), growth hacking, incompleteness by design (Garud, Jain, and Tuertscher 2008) and platform strategies (Gawer and Cusumano 2014). The practices could easily be explained with decision-design and they could also benefit from research on cognitive psychology around the heuristics developed to make and design decisions in entrepreneurship and early-stage decision-making.

These comments are also valid for finance in venture capitalism and private equity. From early-stage to A/B/C/D rounds, investments are also promising grounds to understand the logics of decisions, information gathering and influence on strategy through corporate governance bodies. Hopefully, some answers will be found in the PhD results of Laure-Anne Parpaleix (Centre Gestion Scientifique, Mines ParisTech, and CIFRE contract with BPI France).

---

Organizational Design Thinking and management devices

The model of decisional ambidexterity offers a valuable basis to track and explain projects. It can also be used to derive means of actions and valuation practices. However, it would be interesting to reconnect with existing methodologies and management tools providing valuation for decision-making and action.

Methods like platform engineering and Design Thinking were referred to in two case studies. The first has a strong logic on economic evaluation with modularity benefits (Sanchez and Mahoney 1996). The second builds on user empathy and generative practices incorporating such knowledge. We have revealed that both were minored by outdated models of ambidexterity. However, it would encourage to engineer heuristics and valuation methods in order to steer their generativity countering the limitations of organizational ambidexterity. In the philosophy of hacking, we would prefer thinking of Judo or Aikido moves where light imbalances are turned into an advantage. It echoes the phronesis (Nonaka, Hirose, and Takeda 2016), and sensing, seizing and reconfiguring of (Teece 2007). This economic vigilance would use organization design fixation effects and valuation tools to guide Design Thinking and platform architecture. This vigilance would consequently guide exploration project management.

From user pains and needs to recombination of organizational routines

The divergent and convergent phases, with multiple feedback loops of Design Thinking could also be driven by organizational design. We have shown that their fixations can be hard to overcome because of external biases, so it is worth endogenizing such constraints to generate new alternatives and concepts.

For instance, the Design Thinking cases, by comparison with the later Lower Deck project, show the importance of problem formulation, concept positioning, decision enablers and value space. These elements were part of the innovation potential of attraction descriptor. The Lower Deck project formulated a concept that would overcome several limitations:

• BUs will struggle to see the value of working around a common purpose

• BUs will takeover the full project for primacy

• BUs engineering and marketing departments may not able to value learning and adaptation to a whole new environment

• BUs will prefer waiting for the clients to manifest new use cases and new ecosystems interdependencies, instead of proposing new ones

The Lower Deck built upon the lessons and attempts of previous projects, where it was a design feature or enabler for other functional objectives. However, by managing its value temporarily in a empty space unoccupied by BUs, the concept managed to create sufficient attraction for business units to federate along a common purpose of learning and recombining their engineering routines. Indeed, constraining certification and market demand were shaped in a way that organizational barriers can be lowered.

In other words, an Organizational Design Thinking would specifically target what organizations can’t do on their own. However, it is not a simple technology to be outsourced and integrated. The Connected Cabin project highlighted these limitations. Instead, the concepts aim for reconfiguring interdependencies between BUs, engineering rules and interfaces, as well as the ecosystem. It is not driven by organization pains, but it would focus on existing organization design and routines. One could then search for cracks left by interdependencies and design systematically what BUs can’t do by themselves. We would be thrilled to study these areas with Lisa Carlgren (Chalmers University) and Katharina Hölzle.

Valuation methods: innovation potential

A means to drive the generation of concepts and potential decisions could also around valuation methods. The mutual-conditioning between exploration and exploitation would inspire us to look simply on existing tools and devices. Relying on net present value calculations, business plans and real options could offer numerous opportunities for designing new alternatives and concepts corresponding to hacked valuation methods.

For instance, if several scenarios are presented for a business plan (e.g. pessimistic, nominal and optimistic), one could design alternatives offering them complementary background and comparison. But, the designed decisions could help to think of what should be done for the most favourable scenario. This design practice forces to interact differently with the formalism expected from valuation methods. Problems are inverted. In that perspective, we hope the PhD works of Agathe Gillain (Centre de Gestion Scientifique, CIFRE contract at Airbus) will provide some answers on the economic performance of R&D investments with respect to learning strategies for example.

One of the members of the R&T community had reported developing a selection tools for concepts. Not only would it filter them based on criteria agreed with stakeholders, but they also planned in having a derived scoring that would warn decision-makers to reformulate some concept’s acceptance. In other words, besides filtering, some concepts would trigger to imagine of different states of nature and criteria to accept them. Consequently, it forces stakeholders to review their department’s capabilities, knowledge and relationships across interfaces.

All in all, our perspective on valuation is that decisional ambidexterity and the hacking philosophy promoted with decision-design in the unknown, could support exploration project management. What is even more exciting is that we would use exploitation frameworks to articulate the exploration regime and generative processes. The distance between the two regimes could be measured through this clever conditioning with performance criteria traditionally used. It would avoid isolation between the two as they would use a common language and management tools but have a different relationship to these artefacts.

The intervention for Hypoxia Protection project revealed the usefulness and generativity of such heuristics applied to product conformity matrices (e.g. house of quality) and technical feasibility and performance (e.g. physical measures). In the sociology and ethnography literature, we would also look for echoes of this dual relationship with management devices for exploitation nurturing exploration and preparing its acceptance. It could offer new actions means for marketing as the level unknow heavily contests existing relationships. For instance, the works of Liliana Doganova (Centre for the Sociology of Innovation, Mines ParisTech) on the performativity of valuation devices could be revisited with generativity practices. This concern for exploration project valuation could also be refined in the future with the ongoing works of Svenja Sommer, Kathrin Möslein and Sylvain Lenfle.

---

Generative organization design and governance

For this final section, we will address organization design and governance based on the previous discussions. The cognition, interaction and valuation conveyed in decisional ambidexterity opened stimulating areas to reflect on. Earlier in the manuscript, we had taken a keen interest on organization design and its practice. The conglomerate of SMEs, Zodiac Aerospace, had stressed this dimension for engineering, marketing, market arrangements and group strategic decision-making. Since, the field is willing to renew itself (Barry 2011; Van de Ven, Ganco, and Hinings 2013), we propose to discuss its relationship with generative processes at work for the Innovation function.

The influence of managing as design (Boland and Collopy 2004), as discussed in our literature review, can be bring a different viewpoint on organization design (Romme 2003). The place given to generative processes scattered across the firm, like the existence of silent designers (Dumas and Mintzberg 1989; Dumas 1994), can easily allow us to think of design-driven organizations. It doesn’t necessarily mean that the firm focuses on design services, nor that design methodologies are institutionalized such as firms having embraced Design Thinking, C-K Theory.

However, we would rather be leaning towards a weaker form of design reasoning distributed across the firm, cultivated by the mutual conditioning between exploration and exploitation. In other words, we would not organize collective action for a target organization design. We would instead prefer offering the management tools and managerial philosophy through a specific governance for innovation. The organization will design itself through clearly managed devices, decision-design and a overarching governance body overlooking this fractal organization (Nonaka et al. 2014).

Governance of Innovation function - problem re-formulation

Having a simple rule and heuristic constantly challenging and designing decisions could be used as a pattern to generate organization designs in non-deterministic fashion. Problems are constantly reformulated and solved but without relying specifically on leadership or top management as suggested by (Lakhani, Lifshitz-Assaf, and Tushman 2013).

A culture of design and decision would be formalized through the dual relationship of mutual-conditioning of exploration/exploitation regimes. Re-articulating the decisions and problems vertically, among internal stakeholders but also with the exterior, raises questions on the nature of innovation function’s governance. The writings for the ZAOS Innovate process addressed such heuristics around the overly famous innovation funnel. We realized that beyond the institutionalization of the innovation function (O’Connor 2016) and associated practices reported by R&T managers and in projects, having a governance body for innovation could emphasize the strategic importance of concepts and their associated decision-design scoping. The Multi-BU committee could have been a corporate example. Problems being reformulated and solved by managing the unknown would require to re-engineer interdependencies echoing its organization design mirroring.

We would encourage to continue these discussions with Dominique Laousse, Head of Foresight and Innovative Group at SNCF. He recently defended his PhD at the Centre de Gestion Scientifique under the supervision of Sophie Hooge and Armand Hatchuel. Looking back on the deployment of the innovation function at SNCF, he proposes a staged analysis: from institutionalization, to industrialization (with innovative design projects derived from C-K theory - DKCP) and up to organization and governance. This governance topic reflects also the works of Blanche Segrestin, Kevin Levillain and Armand Hatchuel on purpose-driven companies. Some teachings could possibly be drawn from the ongoing PhD thesis of Jeremy Levêque on firms’ mission drift.

Governance of Innovation function - design-oriented organization

This extreme reconfigurability organized around a new relationship to decision-making by decision-design could perhaps preliminary perspectives to the multi-dimensional organizations (Galbraith 2010). However, there is no target organization design managed and prescribed. Local action through projects, designed decisions and a governance body would sustain new relationships with the environment. This generativity theory of organizing (Van de Ven, Ganco, and Hinings 2013), clearly endogenizes change management in project management (Pollack 2017).

Furthermore, it places the unknown at the heart of management. Managers transfer, solve, and formulate problems that those who report to them can’t solve given their knowledge and relationships. A governance body, dedicated to the unknown through decision-design, would manage its cognition of the environment and strategic health check-up. These perspectives were foreshadowed in the works of Henri Fayol (Fayol 1916). It is worth pointing out that many would only retain the control and regulation roles of Fayol’s management principles (Turner and Keegan 2001; Hodgson 2004; Olin and Wickenberg 2001), whereas he would also insist on strategic knowledge management, relationships with the ecosystem and scientific work (Hatchuel and Segrestin 2018; Hatchuel, Le Masson, and Weil 2002). These practices and heuristics would enact a design-oriented organization (Hatchuel, Weil, and Le Masson 2006) giving a different flavour to the generative fit (Avital and Te’eni 2009). Finally, these concerns would encourage to reconsider the place of entrepreneurship and innovation within the organizations with emergent strategies synonymous with design (Drucker 1985).

---

Chapter synthesis: decisional ambidexterity, a model contributing to perspectives opened in literature

---

References