Designing Decisions in the Unknown: A Generative Model

• Date: 01 June 2019
• DOI: http://doi.org/10.1111/emre.12289
• Author: Pascal Le Masson,Armand Hatchuel,Benoit Weil,Mario Le Glatin
• Published date: 01 June 2019

Designing Decisions in the Unknown: A

Generative Model

PASCAL LEMASSON

1

,ARMAND HATCHUEL

1

,MARIO LEGLATIN

1,2

and BENOIT WEIL

1

1

Center for ManagementScience, i3 UMR CNRS 9217, Mines ParisTech, PSL Research University, France

2

Zodiac Aerospace, Plaisir, France

This study examines how design theory enables to extend decision-making logic to the ‘unknown,’which often appears

as the strange territory beyond the rationality of the decision-maker. We contribute to the foundations of management by

making the unknown an actionable notion for the decision-maker. To this end, we build on the pioneering works in

‘managing in the unknown’and on design theory to systematically characterize rational forms of action in the unknown.

We show that action consists of designing decisions in the unknown and can be organized on the basis of the notion of a

‘decision-driven design path,’which is not yet a decision but helps to organize the generation of a betterdecision-making

situation. Our decision-design model allows us to identify four archetypes ofdecision-driven design paths. They enable us

to discuss the variety of known organizational forms that managers can rely on to explore the unknown.

Additional supporting information may be found in the online version of this article at the publisher’s website.

Keywords: unknown;design theory; decision theory; generativity

Introduction

In a paper recently publish in Science (Bonnefon et al.,

2016), the authors study how an algorithm should ‘decide’

when confronted with a question such as ‘If the brakes have

failed, should the driver system of the car kill the pedestrians

crossing the street or save the pedestrians by crashing the

car into a wall, thereby killing the occupants of the car?’

One can immediately understand the dilemma, and can be

tempted to find an alternative option that is unknown to date,

but would definitely surpass the two options presented.

This example underlines a basic issue in management

science: rational choice is often taken as a given, but there

are sometimes ‘unknowns’that are beyond rational choice

and could deeply influence the rational choice.Hence, the

general questionis can one extend decision-making to the

unknown to rationally support the creation of options?

This issue has largely been addressed by research in

strategic management and risk management (Wideman,

1992; McGrath and MacMillan, 1995, 2009; Pich et al.,

2002; Cunha et al., 2006; Loch et al., 2006, 2008; Mullins,

2007; Weick and Sutcliffe, 2007; Sommer et al., 2008;

Rerup, 2009; Feduzi and Runde, 2014; Feduzi et al.,

2016). The issue of the ‘unknown’is famous both in

professional circles (Wideman, 1992) and in the work of

decision-theory scholars (Miller, 2008). Studies have

contributed to clarifying what is ‘unknown’in relation to

decision-making: decision-makers are confronted with

‘the unknown’when they are confronted with alternatives

and events that were not imagined and taken into account

before and still might impact them to a considerable extent

by radically changing their decision situation.More

formally and more precisely, it has been shown that ‘the

unknown’corresponds to a type of situation that cannot

be handled by the theory of decision-making (Loch et al.,

2006). The issue is not related to decision bias (a

phenomenon that has largely been investigated), but to

generation bias (a phenomenon that is, formally speaking,

not included in decision theory).

As will be shown in the literature review, studies

have described and addressed the challenge of

managing the unknown: they have contributed to

clarifying the goal of generating an improved decision

situation and meeting the challenge of overcoming

generation bias by presenting multiple ways to generate

specific alternatives. However, they have failed to

develop a systematic approach to the unknown and a

structured map of the paths in the unknown that could

contribute to improving the decision situation. Without

such a formal framework, they tend to return, more or

Correspondence: P. Le Masson,Center for Management Science,i3 UMR

CNRS 9217, MINESParisTech –PSL Research University, 60 boulevard

Saint Michel, 75006 Paris, France. E-mail pascal.le_masson@mines-

paristech.fr

European Management Review, Vol. 16, 471–490, (2019)

DOI: 10.1111/emre.12289

©2018 European Academy of Management

less implicitly, to ‘decision-making in conditions of

uncertainty.’Typical examples can be found in

(Sommer and Loch , 2004; Loch et al., 2008): in these

papers, the authors explain that the issue stems from

the fact that, in a decis ion situation, the actors cannot

know all of the possible alternativ es and states of the

world, and explain that managing in the unknown

consists of discovering or generating new alternatives

and new states of the world. However, in the following

paragraphs of the papers, the model they use is actually

arestr iction of an ideal set of alternatives and events,

which is no longer a model of extension, but rather a

model of restri ction, which is well-known in decision

theory. This restrictive approach precludes an analysis

of all facets of the generation of alternatives and states

of the world.

Hence, the aim of this study is to follow the program

outlined by Loch et al. (2006, (2006, 2008) and (Feduzi

and Runde, 2014) to develop normative models that can

provide ‘the standards for comparison and evaluation that

are fundamental to the progress of both descriptive and

prescriptive work’(Feduzi and Runde, 2014: 269). We

are seeking a model for the generation of new states of

the world and new decision alternatives. That is, we

propose a formal model of the extension of decision-

making theory to the unknown, or simply a model of

‘decision design’in the unknown. The requirements for

such a model can be listed: this extension should be

formally consistent, it should contain the decision logic,

it should help characterize and understand critical

phenomena that occur when actors are confronted by the

unknown, and it should lead to a discussion of a new

organizationallogic related to the unknown, makingsense

of the multipleforms and notions that have beenidentified

in contemporary management of innovation and could

actually be related to different types of ‘management in

the unknown.’

As will be described in the literature review, one of the

key issues in such a research program is to develop a

model of generativity that is adapted to decision-making.

This is possible because of the great advances in recent

years in the field of innovation management, wherein

researchers must analyze situations where collective

actions, organizations, and strategiesconsist of addressing

the issue of previously unknown products, services,

business models, and competences. Hence, the findings

of recent studies on innovation management, and more

precisely those on design theory for innovation

management provide us with a model of generativity.

Can it be applied to decision-making? In this study, we

show how models of generativity developed for

innovation management can actually be used fordeci sion

design in the unknown, that is, the generation of ‘better’

decision-making situations, and thus can enrich the field

of decision-making in the unknown.

This paper follows a classical construction: literature

review, methodological approach, construction of the

model, results of the model, and discussion. Hence in

the next part, our literature review identifies a twofold

gap that should be bridged by a formal model

extending decision theory to the unknown: (1) the

model should formally (systematically) account for the

various ways of ‘broadening’a decision space; a nd

(2) the model should help characterize the performance

of this process in terms of ‘comprehensiveness’(Feduzi

et al., 2016) and ‘offsetting cognitive biases’(Feduzi

and Runde, 2014). As we will show, while decision

theory helped characterize ‘selection bias,’our model

should help characterize ‘generation bias.’In the third

part, we present our method and, in particular, explain

why it appears fruitful to rely on design theory to

model the extension of the decision-making framework

to the unknown. Research has enabled the development

of a basic science, design theory, that accounts for the

unique phenomenon of design, namely generativity,

and is comparable in its rigor, foundations, and

potential impact to decision theory, optimization, and

game theory (Hatchuel et al., 2018). As a consequence,

today, design theory appears to provide a promising

way to model the generation of a better decision space

from a given one. In the fourth part, we construct a

formal model that extends decision theory to the

unknown and present its main implications. In the fifth

section, we present the results, i.e. we show how this

model bridges our twofold gap. In the final section

we discuss the results and present our conclusions.

Literature review

The unknown as a limitation to classic decision theory

As noted in (Buchanan and O’Connell, 2006), the

history of decision-making could be considered to begin

with prehistory. However, it was only after World War

II that models of decision-making were progressively

formalized and integrated into a general framework.

Recent historians’studies have enabled us to understand

the ‘rational choice’movement that unfolded at the end

of World War II and during the Cold War (Erickson

et al., 2013).

One of the greatest achievements was the formulation

of a general theory of statistical decision-making under

uncertainty, first by Wald (1939, 1950a, 1950b), which

was then extended to the so-called subjective expected

utility theory (SEUT) by Savage (1951, 1972), and also

codified in management science by Raïffa and Schlaifer

(1961) (see in particular the in-depth analysis of ‘how

homo economicus became Bayesian decision-maker’in

Giocoli, 2013).

472 P. Le Masson et al.

©2018 European Academy of Management