Prediction and control: An agent‐based simulation of search processes in the entrepreneurial problem space

• Author: René Mauer,Malte Brettel,Robert Wuebker,Jan Schlüter
• Date: 01 June 2018
• Published date: 01 June 2018
• DOI: http://doi.org/10.1002/sej.1271

RESEARCH ARTICLE

Prediction and control: An agent-based simulation

of search processes in the entrepreneurial

problem space

René Mauer

1

| Robert Wuebker

2

| Jan Schlüter

3

| Malte Brettel

3

1

Chair of Entrepreneurship and Innovation,

ESCP Europe, Berlin, Germany

2

Department of Entrepreneurship and

Strategy, University of Utah, Salt Lake

City, Utah

3

TIME Research Group, RWTH Aachen

University, Aachen, Germany

Correspondence

René Mauer, ESCP Europe, Chair of

Entrepreneurship and Innovation,

Heubnerweg 8-10, 14059 Berlin, Germany.

Email: rmauer@escpeurope.eu

Research Summary: This study seeks to understand the boundary

conditions and successful application of search processes in the

entrepreneurial problem space. To this end, we employ an agent-

based simulation approach to formally investigate the influence of

environmental isotropy, unpredictability, and goal ambiguity on

two distinct search processes, one prediction-based and one

control-based. Specifically, we investigate the performance of

effectuation as an example of non-predictive, control-based search

and causation as an example of prediction-based search. Results

enhance theory by revealing a more nuanced relationship between

the environment and entrepreneurial search than previous concep-

tual and empirical work has suggested.

Managerial Summary: Historically, entrepreneurship research has

tended to view the external environment as given. The guidance

for entrepreneurs given an immutable landscape sits on a contin-

uum between “learn to plan better”and “learn to react faster.”

These two approaches differ only in their suggestions about what

entrepreneurs should do assuming the environment is given. A

growing body of theoretical work questions this environment-as-

given assumption and suggests an alternative entrepreneurial

search process that seeks active control of environmental ele-

ments. To understand the successful applications as well as limita-

tions of these entrepreneurial search processes, this study runs a

simulation in which agents engage in different search processes.

Practitioners receive more realistic guidance on how search for

value needs a match between the idea, the context, and the search

process.

KEYWORDS

agent-based simulation, control, effectuation, entrepreneurial

problem space, prediction

Received: 11 June 2014 Revised: 25 August 2016 Accepted: 30 January 2017 Published on: 13 November 2017

DOI: 10.1002/sej.1271

Copyright © 2017 Strategic Management Society

Strategic Entrepreneurship Journal. 2018;12:237–260. wileyonlinelibrary.com/journal/sej 237

1|INTRODUCTION

An important question in entrepreneurial strategy concerns the conditions under which search processes that

emphasize information gathering and planning are superior to strategies that emphasize action and learning-by-

doing (e.g., Atuahene-Gima & Li, 2004; Brown & Eisenhardt, 1995; Burgelman, 1983; Corbett, 2005; Delmar &

Shane, 2003; Fiet, 2002; Fredrickson & Mitchell, 1984; Zott & Amit, 2007). In uncertain, rapidly changing contexts,

perspectives on search processes that emphasize low-cost experimentation and rapid adaptation (e.g., Blank, 2013;

Dew, Read, Sarasvathy, & Wiltbank, 2009; Eisenhardt, 1989; McGrath, 1999; Mosakowski, 1997; Mullins & Komi-

sar, 2009; Ries, 2011) have supplanted those that privilege research and planning (Brinckmann, Grichnik, &

Kapsa, 2010).

However, conceptual work underpinning both of these streams of work tends to take the environment as given

and immutable. The two approaches differ only in their suggestions about effective search given an exogenous,

unchanging environment. However, a growing body of scholars spanning strategy, entrepreneurship, and economics

challenges this environment-as-given assumption (Alvarez & Barney, 2007; Dew, Read, Sarasvathy, & Wiltbank,

2011; Felin, Kauffman, Koppl, & Longo, 2014; Sarasvathy, 2001b). This view argues that the heart of value creation

is the realization of a previously non-existent—rather than difficult to predict—future and suggests there are impor-

tant boundary conditions on the efficacy of both planning and learning processes in these settings (Alvarez, Barney,

McBride, & Wuebker, 2017; Welter, Mauer, & Wuebker, 2016; Wiltbank, Dew, Read, & Sarasvathy, 2006). While

the literature admits to a variety of approaches to entrepreneurial search, much work remains to be done to under-

stand the boundary conditions and successful application of search processes in the entrepreneurial context (Arend,

Sarooghi, & Burkemper, 2015; Read, Sarasvathy, Dew, & Wiltbank, 2016).

This article moves toward answering these questions. We begin by creating a simulated world—a laboratory to

examine the entrepreneurial process. We then populate this world with two types of agents: economic agents

attempting to build valuable economic artifacts and consumer agents representing the market for those artifacts. The

economic agents in our simulated world use one of two available search processes to create their economic arti-

facts. The first process focuses on gathering information from consumer agents. Using the first process, economic

agents learn how to tailor their original idea to better correspond to customer preferences from these information-

gathering activities. The second process focuses on gathering information through interacting with other economic

agents rather than consumer agents. Using this second process, focal economic agents focus on enrolling other

makers and builders in a set of resource commitments to co-create their product or service. We use environmental

context as a shift parameter to explore the boundary conditions of these two stylized processes by varying the level

of uncertainty (parameterized as environmental isotropy, environmental unpredictability, and goal ambiguity) across

simulation runs, and we observe the relative performance of these two approaches to entrepreneurial search. Our

results reveal a more nuanced relationship between the environment and entrepreneurial search than previous con-

ceptual and empirical work has suggested.

Our study advances the conversation in strategic entrepreneurship in several ways. First, this article builds on a

body of work using simulation-based approaches to contribute to theory development in entrepreneurship and

strategy (e.g., Adner & Zemsky, 2006; Butler & Grahovac, 2012; Davis, Eisenhardt, & Bingham, 2007, 2009; Graho-

vac & Miller, 2009; Keyhani, Lévesque, & Madhok, 2015). Simulation avoids the inherent limitations of scope, data,

and complexity in empirical investigations and provides a valuable window into the entrepreneurial process, notori-

ously difficult to study in the wild (Davis et al., 2007; Harrison, Lin, Carroll, & Carley, 2007; McKelvey, 2004).

Second, the particular approach to simulation we take in this article reveals novel insights about the contextual

drivers of entrepreneurial search. Though some conceptual work in entrepreneurship and strategy considers the role

of uncertainty as an influence on search outcomes, little work directly tests this conjecture. Where it has been

addressed, uncertainty has often been parameterized using a single construct or proxy variable. Our approach adds

additional specificity and depth to representations of uncertainty in models of search processes.

238 MAUER ET AL.