Organizing for knowledge generation: internal knowledge networks and the contingent effect of external knowledge sourcing

• Published date: 01 February 2017
• Author: Frank T. Rothaermel,Konstantinos Grigoriou
• Date: 01 February 2017
• DOI: http://doi.org/10.1002/smj.2489

Strategic Management Journal

Strat. Mgmt. J.,38: 395–414 (2017)

Published online EarlyView 22 February 2016 in WileyOnline Library (wileyonlinelibrary.com) DOI: 10.1002/smj.2489

Received 26 October 2011;Final revision received30 October 2015

ORGANIZING FOR KNOWLEDGE GENERATION:

INTERNAL KNOWLEDGE NETWORKS AND THE

CONTINGENT EFFECT OF EXTERNAL KNOWLEDGE

SOURCING

KONSTANTINOS GRIGORIOU1*and FRANK T. ROTHAERMEL2

1College of Business, Florida International University, Miami, Florida, U.S.A.

2Scheller College of Business, Georgia Institute of Technology, Atlanta, Georgia,

U.S.A.

Research summary: When faced with a new technological paradigm, incumbent rms can

opt for internal development and/or external sourcing to obtain the necessary new knowledge.

We explain how the effectiveness of external knowledge sourcing depends on the properties

of internal knowledge production. We apply a social network lens to delineate interpersonal,

intra-rm knowledge networks and capture the emergenceof two important rm-level properties:

the incumbent’s internal potential for knowledge recombination and the level of knowledge

coordination costs. We rely on rm-level internal knowledge networks to dynamically track the

emergence of these properties across 106 global pharmaceutical companies over a 25-year time

period. We nd that a rm’s success in developing knowledge in a new technological paradigm

using external knowledge sourcing is contingent on these internal knowledge properties.

Managerial summary: Incumbent rms in high-tech industries often face competence-destroying

technological change. In their effort to adapt and develop new knowledge in a novel paradigm,

incumbent rms have several corporate strategy options available to them: internal knowledge

development and a wide array of external knowledge sourcing strategies, including alliances

and acquisitions. In this study, we make an effort to address a critical question: How effective

is external knowledge sourcing under different internal knowledge generation regimes? We nd

that external sourcing strategiesare less effective when rms can already internally generate new

knowledge or if they have high internal coordination costs. Therefore, when considering external

sourcing, managers must carefullyweigh the benets of it vis-à-vis its commensurate costs as the

benets of external sourcing may be overstated. Copyright © 2015 John Wiley & Sons, Ltd.

INTRODUCTION

Competence-destroying technological change

poses signicant challenges for an industry’s

Keywords: organizing for adaptation; internal knowledge

networks; innovation strategies; complementarities; exter-

nal knowledge sourcing

*Correspondence to: Konstantinos Grigoriou, 11200 SW 8th

Street, MANGO 443, Miami, FL, 33199, U.S.A.

E-mail: kgrigori@u.edu

Copyright © 2015 John Wiley & Sons, Ltd.

incumbent rms (Tushman and Anderson, 1986).

To build knowledge within a new technological

paradigm, incumbents invest in internal knowl-

edge development (Tripsas, 1997), human capital

(Zucker and Darby, 1997), strategic alliances

(Rothaermel, 2001), acquisitions of new entrants

(Higgins and Rodriguez, 2006), or in combinations

of those strategies (Rothaermel and Hess, 2007).

There is a signicant degree of consensus in the

strategy literature that successful renewal depends

on incumbents developing skills in both internal

396 K. Grigoriou and F. T. Rothaermel

knowledge development and external knowledge

sourcing (Helfat et al., 2007).

Clearly, both internal development and external

sourcing have received considerable attention in the

literature as more or less isolated strategic choices.

We have a somewhat limited understanding, how-

ever, about the conditions that favor one sourcing

mode over the other. Capron and Mitchell (2009)

echo this statement when arguing that a rm’sselec-

tion capability, dened as the ability to choose

among modes of knowledge sourcing, is an under-

emphasized form of capability. Further, Nickerson

and Zenger (2004: 1) suggest that the key question

is not “how to organize to exploit already developed

knowledge or capability, but rather how to organize

to efciently generate knowledge and capability.”

We contribute to this line of burgeoning research

by applying social network concepts to address

the question of how to organize and choose among

knowledge sourcing efforts to efciently generate

new knowledge. We predict the effectiveness of

knowledge sourcing strategies for incumbent rms

trying to develop new knowledge within an emerg-

ing technological paradigm. We track their internal

and external sourcing activities in a ne-grained

manner as they co-evolve over time. As the

incumbent rms generate new knowledge, we use

collaborative network analysis to capture important

properties of their internal knowledge production

process. In particular, we look deep inside the rm

to track the interpersonal networks of knowledge

workers to assess not only the value created by

knowledge generation, but also its production

costs. To track external sourcing, we examine

the incumbents’ activity with knowledge-related

alliances and acquisitions.

Our core argument is that the effectiveness of

a rm’s external knowledge sourcing strategies to

build new knowledgeis contingent on the properties

of that rm’s internal knowledge production. We

focus on two such properties: the rm’s potential

for subsequent internal knowledge recombination

and the knowledge coordination costs associated

with its existing knowledge production system.

We explain why external sourcing is bound to be

less effective for new knowledge generation for

rms that either already possess strong potential

for further knowledge recombination internally

or already incur high coordination costs in their

existing knowledge production process.

To proxy for these rm-level properties, we look

at ne-grained internal knowledge networks and we

rely on the concept of emergence (Kozlowski etal.,

2013). We explain how the structure and evolution

of interpersonal knowledge networks inside incum-

bent rms result in the emergence of rm-level

knowledge properties such as recombinative poten-

tial and coordination costs. In essence, we describe

how rm-level network properties emerge from

internal individual-level network structures and

interactions (Moliterno and Mahony, 2011). In

short, we argue that differences across rms in these

knowledge network properties drive the contingent

effect of their external knowledge sourcing choices

on their knowledge generation within a new techno-

logical paradigm.

The novel contribution in our study results

from the application of social network concepts to

develop a ne-grained picture of the state of a rm’s

internal knowledge production process, and to cap-

ture the potential for knowledge recombination and

accompanying coordination costs. Moreover, this

fresh approach enables us to document the role

of the rm’s internal knowledge network as an

important but overlooked contingency factor when

evaluating the effectiveness of external knowledge

sourcing choices. As a result, we also offer a

signicant contribution to the growing literature on

knowledge networks and their importance for rm

performance (Phelps, Heidl, and Wadhwa, 2012).

Empirically, we examine the phenomenon using

large but detailed longitudinal panel datasets of

both rms and individual inventors within these

rms. When estimating the contingent effects of

external knowledge sourcing choices, we dynami-

cally update the rms’ internal knowledge networks

to address the co-evolution of internal knowledge

and knowledge sourcing choices. We test the

theoretical framework in the global pharmaceutical

industry, where we track the innovative activities of

106 incumbent rms in their effort to adapt to the

new biotechnology paradigm over a 25-year period

(1974– 1998). We rely on the rms’ patenting port-

folios of more than 267,000 individual inventors

to build internal networks for biotech knowledge

generation. We apply network theory to extract

information from intra-rm networks about the

individual inventors’ network positions, the overall

network structure, and subsequently, the rms’

knowledge properties of knowledge recombinative

potential and coordination costs in its knowledge

production process.

To foreshadow our conclusions, we submit

that selecting a specic path to develop new

Copyright © 2015 John Wiley & Sons, Ltd. Strat. Mgmt. J.,38: 395–414 (2017)

DOI: 10.1002/smj