Product and Organizational Modularity: A Contingent View of the Mirroring Hypothesis

• Author: Metehan Feridun Sorkun,Andrea Furlan
• Published date: 01 June 2017
• Date: 01 June 2017
• DOI: http://doi.org/10.1111/emre.12101

Product and Organizational Modularity: A

Contingent View of the Mirroring Hypothesis

METEHAN FERİDUN SORKUN

1

and ANDREA FURLAN

2

1

Department of Business Administration, İzmir University of Economics, İzmir, Turkey

2

Department of Economics and Management, The University of Padova, Padova, Italy

This study aims to advance a contingent view of the mirroring hypothesis, focusing on the causaleffect of product

modularity on organizational modularity. It compares the group of papers (‘orthodox’) that confirms the mirroring

hypothesis with the group of papers (‘critique’) that assumes a critical position. The citation network analysis

performedon the ‘critique’group searchesfor contingentfactors that explainunder which circumstancesthe mirroring

hypothesis does not hold. Results show that the mirroring hypothesis is contingent on a set of six distinct, although

interdependent, factors:(i) component technologicalchange and diversity; (ii)innovativeness of productarchitecture;

(iii) complexityof product architecture; (iv)capability dispersionalong the supply network; (v) rivalryamong leading

firms & among suppliers; and (vi) logistics costs. This study discusses each contingent factor in turn, highlighting

theories or approaches that ought to be used in conjunction (not in contrast) with the modularity theory to explain

the boundariesof firms and industries.A contingent view of themirroring hypothesisreconciles the twoopposite views

on the mirroring hypothesis,enhances its ramificationsinto the theory of the firm, and givesinsights to practitioners.

Keywords: productmodularity; organizational modularity;network analysis; literaturereview; mirroring hypothesis

Introduction

The relationship between product modularity (PM) and

organizational modularity (OM) lies at the heart of the

modularity theory of the firm that aims to explain the

positions of the boundaries of firms and, as a

consequence, the vertical contracting structure of

industries. The relevance of this relationship dates back

to Simon’s (1962) ground-breaking study on the

architecture of complexity based on hierarchy and

near-decomposability. Thanks to the author’s intuitions,

it is possible to explicitly integrate technology (a variable

often considered exogenous in organizational economics)

into the analysisof the evolution of organizationaldesign.

The seminal paper by Sanchez and Mahoney (1996) has

crystalized the effect of PM on OM indicating that the

architecture of organizations mirrors the corresponding

product design. In other words, modular organizations

come from modular products, and integral organizations

come from integral products.

1

A direct and positiverelationship between PM and OM

has been framed as the mirroring hypothesis (Colfer and

Baldwin, 2010), which has aroused the interest of many

studies so far. The mirroring hypothesis implies a

bi-directional positive relationship between PM and OM.

One causal relationship that the mirroring hypothesis

embeds is from organizations’architectures to product

architectures (Henderson and Clark, 1990). This

hypothesizedcausation rests on the argumentthat problem

solvingand communication patternsof organizations (OM

level) draw the solution space for product design (PM

level) (MacCormack et al., 2012). Therefore, modular

organizational forms in which loosely coupled

organizational units specialize in distinct knowledge

domains are more likely to design modular products

because no mechanism is available to integrate the

dispersed knowledge domains to maximize component

synergies that is necessary for the developmentof integral

products.

This review focuses on the mirroring effect from

products’architectures to organizations’architectures. In

other words, it focuses on the effects that the architecture

of a product has on the organization of the development

and production activities of that product. Thus,in the rest

of the paper, the term ‘mirroring hypothesis’refers to the

causal relationship from product modularity (PM) to

organizational modularity (OM).

Correspondence: Metehan Feridun Sorkun, İzmir University of

Economics, Faculty of Business, Department of Business

Administration, Sakarya Caddesi, No. 156, Balçova, İzmir (35330),

Turkey. Tel:+90 5069350029. E-mail metehan.sorkun@ieu.edu.tr

1

Although there is an emergent literature on service modularity (Voss and

Hsuan, 2009; Bask et al., 2010; Cabigiosu et al., 2015), the present study

focuses on productmodularity.

European Management Review, Vol. 14, 205–224, (2017)

DOI: 10.1111/emre.12101

©2016 European Academy of Management

Several studies have empirically supported the

mirroring effect from PM to OM at different levels of

analyses and in several industries (Schilling and

Steensma, 2001; Sturgeon, 2002; Fixson and Park, 2008;

Cabigiosu and Camuffo, 2012). Modules are interpreted

as market-supporting institutions providing technical

design rules that standardize the interfaces between

different product components or stages of the production

process (Sabel and Zeitlin, 2004). The modularity in

product servesas the functional equivalentof coordination

mechanisms between firms in a supply chain or among

organizational units of the same firm.

As the number of empirical studies have increased and

the scope of enquiry has broadened, a considerable

number of studies revealed contradictory findings. It is

argued that the presumed positive effect of PM on OM

conjectured by the mirroring hypothesis is inhibited by

inextricable performance interactions among the

components of the products (Langlois, 2002; Brusoni

and Prencipe, 2006; Zirpoli and Becker, 2011; Whitford

and Zirpoli, 2014).Even in the case of modular products,

there can be residual interdependencies which force

organizations to adopt complex and expansive

coordination mechanisms in order to assure the

performance of the final product.

The first part of this study carries out a systematic

analysis to compare the group of papers (‘orthodox’)that

confirms the mirroring hypothesis with the group of

papers (‘critique’) that assumes a critical position. At the

best of authors’knowledge, this study is the first attempt

to compare these two groups. Although the majority of

the papers belong to the second group, the contradiction

between the two groupsis fairly persistent over time. This

paper searches for any specific pattern related to the

research designs of papers that can explain their different

positions.Results show that the differences in theresearch

designs (such as industries, adopted measures, and levels

of analyses for PM and OM) are not responsible for the

positions of papers in one of the two groups.

Having ruled out the possibility that different research

settings might be responsible for the contrasting views

on the mirroring hypothesis, the authors argue that the

discrepancy between product and organizational

architectures derives from contingent factors ‘that

ultimately impinge on the ability of modular design rules

to bring about modular organizations’.(D’Adderio and

Pollock, 2014: 1814). However, there is a lack of studies

that systematically analyze and discuss these contingent

factors.

Therefore, the second part of this study performs a

citation networkanalysis on the ‘critique’group searching

for contingent factors that explain under which

circumstances the mirroring does not hold. Results show

that the mirroring hypothesis is contingent on a set of six

distinct, although interdependent, factors: (i) component

technological change and diversity; (ii) innovativeness

of product architecture; (iii) complexity of product

architecture; (iv) capability dispersion along the supply

network; (v) rivalry among leading firms & among

suppliers; and (vi) logistics costs. This study discusses

each contingent factor in turn, highlighting theories or

approaches that ought to be used in conjunction (not in

contrast) with the modularity theory to explain the

boundaries of firms and industries. The authors maintain

that a contingent view of the mirroring hypothesis

reconciles the two opposite views on the mirroring

hypothesis, enhances its ramifications into the theory of

the firm, and gives insights to practitioners by helping

them to assess the contingencies under which the

mirroring hypothesis does not hold true.

PM, OM, and the mirroring hypothesis:

conceptual definitions

While a product is composedof different components and

sub-systems, an organization can be partitioned into

departments, separate firms and vertical industrial layers.

Therefore, it is possible to identify both products and

organizations as systems whose constituents have

interdependencies to implement their system functions.

For example, while technical, volumetric and aesthetic

interdependencies exist among components and

subsystems of products, there are informational,

governmental and resource interdependencies among

organizational units (Worren, 2012).

Product modularity (PM) refers to the degree of

technical interdependencies between the constituents of a

product (Schilling, 2000). This review identifies three

different levels of PM: the final product level that

indicates the modularity degree of a final product; the

sub-system level that indicates the modularity degree of

a single sub-system; and the component level that

indicates the modularity degree of a single component.

The term ‘PM’is used to cover all these three levels of

analyses. Regarding the modularity degree of a final

product, Ulrich (1995) proposes two criteria: (a) their

constituents are coordinated through standardized

interfaces and (b) each constituent performs only one

product function. The subsystems of a final product can

also be defined aseither modular or non-modular (Simon,

1962) because they can contain complex components’

interactions such as in the aircraft engine control system

(Brusoni et al., 2001), the exhaust system (Persson and

Åhlström, 2013), and the air-conditioning system in

automobiles (Cabigiosu et al., 2013). Finally, there are

studies (Hsuan, 1999; Furlan et al., 2014) assessing

whether components are modular or not. The degree of

component modularity depends on how many functions

206 M.F. Sorkun and A. Furlan