Arrangements For Information Technology Governance

Arrangements for Information Technology Governance: A Theory of Multiple
Contingencies
Author(s): V. Sambamurthy and Robert W. Zmud
Source: MIS Quarterly, Vol. 23, No. 2 (Jun., 1999), pp. 261-290
Published by: Management Information Systems Research Center, University of Minnesota
Stable URL: http://www.jstor.org/stable/249754
Accessed: 29-05-2017 23:40 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
Management Information Systems Research Center, University of Minnesota is

collaborating with JSTOR to digitize, preserve and extend access to MIS Quarterly
This content downloaded from 146.83.119.10 on Mon, 29 May 2017 23:40:24 UTC
All use subject to http://about.jstor.org/terms
Sambamurthy & Zmud/IT Governance
rfimmAmmm
ARRANGEMENTS FOR INFORMATION

TECHNOLOGY GOVERNANCE:
A THEORY OF MULTIPLE CONTINGENCIES1
By:
By: V.
V. Sambamurthy
Sambamurthy assumption
assumption underlying
underlying these
thesestudies
studiesisisasasthough
though
Decision
Decision and
and Information
InformationTechnologies
Technologies the organizational
organizational contingencies
contingenciesact
actininisolation
isolation
The Robert H. Smith School of Business in influencing
influencing the
the mode
modeof
ofIT
ITgovernance.
governance.
University of Maryland However,
However, inin reality,
reality, business
businessfirms
firmsare
aresubject
subjecttoto
College Park, MD 20742-1815 the pulls
pulls and
and pressures
pressures of
ofmultiple,
multiple,rather
ratherthan
than
U.S.A. singular,
singular, contingency
contingency forces.
forces.Therefore,
Therefore,toto
[email protected] acknowledge
acknowledge this
this reality,
reality,this
thisstudy
studyapplies
appliesthe
the
theory
theory of
of multiple
multiple contingencies
contingenciesto
toexamine
examinehow how
Robert W. Zmud contingency
contingency forces
forces influence
influencethe
themode
modeofofITITgov-
gov-
Michael F. Price College of Business ernance.
ernance. The
The theory
theory argues
arguesthat
thatcontingency
contingency
University of Oklahoma forces
forces interact
interact with
with each
eachother
otherby
byeither
eitheramplify-
amplify-
Norman, OK 73019 ing, dampening,
dampening, or
or overriding
overridingtheir
theirmutual
mutualinflu-
influ-
U.S.A. ences on
on the
the IT
IT governance
governancemode.
mode.Three
Threescenar-
scenar-
[email protected] ios of
of multiple,
multiple, interacting
interactingcontingencies
contingenciesareare
identified:
identified: reinforcing,
reinforcing,conflicting,
conflicting,and
anddominat-
dominat-
ing. Each
Each of
of these
these scenarios
scenariosofofmultiple
multiplecontin-
contin-
Abstract
gencies
gencies is
is hypothesized
hypothesized to toinfluence
influencea aparticular
particular
IT governance arrangements refers to the patterns mode ofof IT
IT governance.
governance. Utilizing
Utilizingrich
richdata
datafrom
from
of authority for key IT activities in business firms, case studies
studies of
of eight
eight firms,
firms,empirical
empiricalevidence
evidenceisis
including IT infrastructure, IT use, and project presented to support these hypotheses.
management. During the last 20 years, three pri- Implications of the multiple contingencies theory
mary modes of IT governance have become for research and for practice are presented.
prevalent: centralized, decentralized, and the
Keywords: IS organization design, multiple con-
federal mode. These modes vary in the extent to
tingencies, federal governance, qualitative
which corporate IS, divisional IS, and line man- research
agement are vested with authority for the key IT
activities. While a significant volume of research ISRL Categories: DA03, EG01, EG03
has examined the influence of contingency fac-
tors on the choice of a specific mode of IT gov-
ernance, most of this research has examined the
singular effects of the contingency factors. The Introduction
Today, information technologies (IT) present

'Lynda Applegate was the accepting senior editor for many opportunities for firms to enhance or trans-
this paper. form their products, services, markets, work
MIS Quarterly Vol. 23 No. 2, pp. 261-290/June 1999 261
processes,
processes, and
andbusiness
businessrelationships.
relationships.Such
Such
ini-
ini- governance
governance mode,
mode,
divisional
divisional
IS andISline
andman-
line man-
tiatives,
tiatives, however,
however,demand
demandthe
theintertwined
intertwined
efforts
efforts agement
agementassume
assume
authority
authority
for allfor
IT activities.
all IT activities.
of technology
technologyand andbusiness
businessspecialists
specialists(Boynton
(Boynton However,
However, as as
illustrated
illustrated
in Table
in Table
1, five1,
different
five different
et al.
al. 1994).
1994). Therefore,
Therefore,research
researchattention
attentionhashas variations
variations in in
decentralized
decentralized
IT governance
IT governance
are are
begun
begun toto focus
focusononthe
thenature
natureofofITITcapabilities
capabilities observable
observable in in
practice.
practice.
TheseThese
variations
variations
depict depict
and
and organization
organizationdesigns
designsthat
thatwill
will
enable
enablefirms
firmsto to different
different patterns
patterns
of allocation
of allocation
of authority
of authority
for IT for IT
exploit
exploit the
the business
businesspotential
potentialofofITIT(Broadbent
(Broadbent infrastructure
infrastructure management,
management,
IT useIT
management,
use management,
and Weill 1997; Henderson and Venkatraman and
andproject
project management
management decisions
decisions
among among
the the
1992; Rockart et al. 1996; Ross et al. 1996; divisional
divisional IS IS
staff
staff
and and
line management.
line management.
Finally, Finally,
Sambamurthy and Zmud, 1994). with
withthethe federal
federalgovernance
governance mode,mode,
both corpo-
both corpo-
rate
rateISISand
andthethe
business
business
unitsunits
(either(either
the division-
the division-
Prior IS research suggests that three distinct
al
al IS
ISor
orline
line
management)
management)assume
assume
authority
authority
for for
spheres of IT activities underlie IT-based innova-
specific spheres of IT activities (Brown and
tion: IT infrastructure management, IT use man-
Magill 1994; Dixon and John 1989; LaBelle and
agement, and project management (Clark et al.
Nyce 1987; LaPlante 1991; Rockart 1988; von
1997; Cross et al. 1997; Sambamurthy and Zmud
Simson 1990; Zmud, 1988). As Table 1 illus-
1994; Weill and Broadbent 1998). IT infrastruc-
trates, three primary variations of federal gover-
ture management consists of decisions that
nance mode are observable in practice. Further,
address the nature of hardware and software plat-
within each of these primary variations, there
forms, annual enhancements to these platforms,
might be even more variations depending upon
the nature of network and data architectures, and
whether divisional IS or line management is the
the corporate standards for procurement and
primary authority. For example, Table 1 shows
deployment of IT assets (LaBelle and Nyce 1987;
the first federal governance pattern as allocating
von Simson 1990; Weill and Broadbent 1998;
project management authority to either divisional
Wilder 1990). IT use management consists of
IS or line management. Similar variations can be
decisions that address applications prioritization
observed in the other federal governance patterns.
and (short-term and long-term) planning, budget-
ing, and the day-to-day delivery of operations Observations of practice indicate that consider-
and services (LaBelle and Nyce 1987; von able diversity exists in the patterns of IT gover-
Simson 1990; Wilder 1990). Project manage- nance arrangements across contemporary firms.
ment involves blending knowledge of IT infra- Understanding how and why a firm has adopted
structure capabilities and capacities with knowl- a specific IT governance arrangement is impor-
edge associated for the conceptualization, acqui- tant in order to advance knowledge about the
sition, development and deployment of informa- effectiveness of alternative governance arrange-
tion systems applications (Curtis et al., 1988; ments for sustaining IT-based innovation. Prior
Slevin and Pinto 1987; Walz et al. 1993). research has examined the influence of a variety
of factors: industry (Ahituv et al. 1989; Clark
The above three spheres of activities are directed,
1992), firm size (Ahituv et al. 1989; Brown and
controlled, and coordinated through IT gover-
Magill 1994; Clark 1992; Tavakolian 1989), cor-
nance arrangements. Governance arrangements,
porate strategy (Brown and Magill 1994;
hence, represent an organization's IT-related
Tavakolian 1989), and corporate structure
authority patterns. Typically, three sets of stake-
(Ahituv et al. 1982; Applegate et al. 1996; Brown
holders are provided varying degrees of decision
and Magill 1994; Tavakolian 1989). Although
authority in designing IT governance arrange- such research has increased our collective
ments: corporate IS, divisional IS, and line man-
understanding of the factors influencing a firm's
agement (Brown and Magill 1994; Sambamurthy
choice of IT governance arrangements, the vast
and Zmud 1993).
majority of prior studies have focused on singu-
Over the last 20 years, three primary modes of IT lar impacts of specific contingency factors. For
governance arrangements have emerged (Table example, Olson and Chervany (1980) examined
1). With the centralized governance mode, cor- the influence of corporate structure (centraliza-
porate IS has the authority for all three of the tion vs. decentralization), whereas Tavakolian
spheres of IT activities. With the decentralized (1989) examined the impact of business strategy
262 MIS Quarterly Vol. 23 No. 2/June 1999
Centralized Governance Mode
Spheres of IT Activities Patterns in the Locus of Authority

IT infrastructure Corporate IS
IT use Corporate IS
Project management Corporate IS

Decentralized Governance Mode

Pattern D1 Pattern D2 Pattern D3 Pattern
IT infrastructure Divisional IS Divisional IS Divisional IS Divisional

IT use Divisional IS Divisional IS Line management Line manageme
Project management Divisional IS Line management Divisional IS Line manage
* This governance scenario would be very rare and would occur only when an organizati
or divisional levels. Further, it is practically not feasible to conceptualize a governance
Cl) manage the infrastructure, but divisional IS staff would manage other IT activities.
I
Federal Governance Mode
cb
Pattern F1 Pattern F2 Pattern
r\)
IT infrastructure Corporate IS Corporate IS Corpor
IT use Corporate IS Divisional IS or line management Divisional IS

"I
0
Project management Divisional IS or line management Corporate IS Divisiona
co
(AD
Note: Other governance scenarios where corporate IS has the authority for IT use or project management, but divisional IS has the
rc3
authority for IT infrastructure, are not practically feasible in contemporary firms.
orientation
orientation (prospectors,
(prospectors,defenders,
defenders,oror
analyz-
analyz- key
keysubunits
subunitsin order
in order
to lower
to lower
their organization-
their organization-
ers). al coordination costs (Blau and Schoenherr
1971; Galbraith 1967; Lawrence and Lorsch
However, in reality, the evolution of firms' IT
1969). IS researchers have found that the mode
governance arrangements depends on the con-
of corporate governance significantly influences
fluence of many contingency factors. To add
the mode of IT governance: firms that have cen-
even more complexity, these contingency factors
tralized their corporate governance also tend to
(in practice) interact with one another in shaping
centralize their IT governance, whereas firms that
firms' IT governance arrangements. The advance-
have decentralized their corporate governance
ment of our knowledge of why and how firms
tend to decentralize their IT governance
exhibit particular IT governance arrangements
(Applegate et al. 1996; Brown and Magill 1994;
requires accounting for the effects of complex
Ein-dor and Segev 1982; King 1983; Leifer 1988;
interactions among the multiple contingency
Olson and Chervany 1980; Tavakolian 1989).
forces that are likely to simultaneously prevail in
practice. This research draws on the theory of Further, the size of the firm influences the mode
multiple contingencies (Gresov 1989) to provide of IT governance through its influence on the
explanations of how several contingency forces mode of corporate governance. Smaller firms
actively influence firms' IT governance arrange- tend to be functionally structured (i.e., the H-
ments. Through conceptual argumentation and form), with individual subunits responsible for
evidence across eight case sites, the research key functions, such as production, marketing,
demonstrates that firms' IT governance arrange- finance, and accounting. Such organizations uti-
ments are the outcome of specific interaction lize centralized corporate governance arrange-
patterns. ments in order to facilitate coordination of inter-
This paper is organized as follows. The next two dependencies among the activities of the indi-
vidual subunits. Further, such firms utilize inte-
sections present a literature review of IT manage-
ment governance modes and their determinants. grated information systems in order to manage
Next, the paper presents its theoretical model the subunit interdependencies. As a result, small
based on the application of the theory of multi- firms are more likely to adopt a centralized gov-
ple contingencies and articulates three hypothe- ernance arrangement. Growth in the size of a
ses. The subsequent section details the research firm, however, typically leads to the divisional-
methodology. Then, results are presented. ization of its operations and a resultant decen-
Finally, the paper discusses implications for tralization of the corporate governance (Blau
research and practice. 1970; Blau and Schoenherr 1971; Grinyer et al.
1980). Divisionalization occurs in the form of
segmentation along lines-of-business or geo-
graphic regions. In these larger firms, a signifi-
Determinants of the Modes of IT cant amount of the information systems require-
ments tend to be customized to the needs of indi-
Governance Arrangements vidual business or geographical units. As a result,
Prior studies of IT governance arrangements large firms are likely to decentralize aspects of
have identified a number of contingency factors their IT governance arrangements in order to be
(see Table 2). These earlier findings are built more responsive to individual subunits' IS needs
upon by categorizing them into three sets of (Ein-Dor and Segev 1982).
forces-corporate governance, economies of
scope, and absorptive capacity-as determi- Economies of Scope
nants of a firm's adoption of a particular IT gov-
Economies of scope are benefits accruing to a
ernance arrangement.
firm when it is able to share its resources across
multiple products/services. As a result, the syner-
Corporate Governance gistic costs are less than the sum of the costs of
Business firms seek to mirror their corporate gov- providing each product or service by itself
ernance arrangements in the activities of their (Panzar and Willig 1981; Teece 1980).
I 0 4 K, f I 117 ro M-Mil I I-N i
I I
Category of Locus of
Forces Contingency Factors Decision Rights Supp
Factors Levels
Corporate Overall Centralized Corporate IS Applegate et al. 1996; Brown and M

Governance governance mode Segev 1982; King 1983; Leifer 1988; O
Decentralized Division 1980; Tavakolian 1989
Firm size Small Corporate IS Ein-Dor and Segev 1982
Large Division
Economies Diversification Internal growth Corporate IS Brown and Magill 1994; Zmu
of Scope mode
Acquisition growth Division
Diversification Related markets Corporate IS Boynton et al. 1992; Brown and

breadth
Unrelated markets Division
Exploitation Enterprise-wide Corporate IS Henderson 1990; Keen 1991; L

strategy for scope consolidation of La Plante 1991; Maglitta and Meh
"I economies assets 1988; von Simson 1990; Wilder 1990
.0
Enterprise-wide Division
Cr\
p line/IS partnerships
Absorptive Line IT Low Corporate IS Boynton et al. 1992, 1994; Brown

Capacities knowledge
C-
High Division
Cb
Qo
cAD
0)
Economies
Economies of
ofscope
scopeoften
oftenarise
arisefrom
from
the
the
sharing
sharing nizational relationships
relationships and
and governance
governance struc-
struc-
of knowledge
knowledgeandanddistinctive
distinctivemanagerial
managerial
compe-
compe- tures might exceed
exceed the
the achievable
achievable economies
economies of of
tencies
tencies (Ginsberg
(Ginsberg1990;
1990;Hamel
Hameland
and
Prahalad
Prahalad scope. First, the desired
desired organization
organization design
design con-
con-
1994;
1994; Nelson
Nelson and
andWinter
Winter1982;
1982;Winter
Winter 1987).
1987). flicts so much with
with that
that existing
existing in
in the
the acquired
acquired
However,
However, for
foraafirm
firmtotorealize
realizeeconomies
economiesofof businesses that attempts
attempts to
to implement
implement the the "new"
"new"
scope,
scope, administrative
administrativemechanisms
mechanisms oror
operating
operating design are simply not
not undertaken
undertaken (Ghoshal
(Ghoshal 1987;
1987;
relationships
relationships must
mustbebeestablished
establishedamong
amongitsits
units
units Kazanjian and Drazin
Drazin 1987).
1987). Second,
Second, organiza-
organiza-
to keep
keep track
trackof
ofopportunities
opportunitiesfor
for
synergies
synergiesandand tional inertia in the
the behaviors
behaviors of
of the
the senior
senior man-
man-
to ensure exploitation of the associated agement of the acquired
acquired units
units may
may forestall
forestall
economies of scope (Hoskisson and Hitt 1988; efforts to transition
transition toward
toward aa unified
unified organization
organization
Reed and Luffman 1986; Teece 1980). In the design (Bastien 1987;
1987; Hannan
Hannan and
and Freeman
Freeman
present study, economies of scope are the bene- 1984; Napier 1989).
1989). Third,
Third, circumstances
circumstances sur-
sur-
fits arising from sharing of appropriate IT-related rounding the acquisition
acquisition itself
itself often
often create
create aa lack
lack
expertise, know-how, and investments across of harmony, goal incongruence,
incongruence, andand competitive,
competitive,
the enterprise. rather than cooperative,
cooperative, behaviors
behaviors on
on the
the part
part of of
the managers within
within the
the acquiring
acquiring and
and acquired
acquired
Research in strategic management suggests that units of a business (Jones and Hill 1988;
the feasibility of leveraging economies of scope
Williamson 1975). In the IT governance context,
is largely determined by diversification mode
firms that grow through acquisitions might dis-
and diversification breadth. Diversification mode
cover that the acquired units have disparate IT
refers to whether the enacted growth strategy of
infrastructures, different histories and cultures of
a business is driven by internal expansion or
interactions and relationships among the line and
external acquisitions (Simmonds 1990). During
IS managers, and different levels of reliance on IT
internal expansion, a firm "grows its business"
in their business strategies and value chain oper-
through internal capabilities and product-market ations. Further, the circumstances of the merger
expansions. Two considerations guide the expan-
of acquired units could raise concerns among the
sion of such firms into areas that leverage their
managers of the acquired units that they might
existing core capabilities and knowledge. First,
lose the responsiveness and service quality of
these firms possess an inherent familiarity with
their IS units, particularly if efforts are initiated
the new entry markets and adopted technologies
to tightly integrate the IS activities across the
and are thus able to extend their current admin-
acquired units. Overall, these arguments sug-
istrative and operating systems to the new busi-
gest that firms that grow through acquisition
ness venture (Jemison and Sitkin 1986; Lubatkin
might judge the costs of integrating the differ-
1986). Second, the pace of internal expansion ent IT units into a centralized governance
tends to be gradual in nature, thereby enabling
arrangement to outweigh the potential
the firm to experiment with governance mecha-
economies of scope. Therefore, such firms may
nisms that permit exploitation of synergies and
very well adopt decentralized IT governance
economies of scope (Pitts 1976; Simmonds arrangements.
1990). In the IT governance context, firms pur-
suing internal growth would expect to leverage Diversification breadth refers to the degree of
existing technology investments and existing product/market relatedness of a multi-business
technology competencies throughout their firm. Firms that compete in related product-mar-
enterprise. This would be easiest to accomplish kets tend to posses considerable similarity in
when IT resources and decisions are centrally
their production and logistics processes, con-
orchestrated.
sumers, distribution channels, and marketing
On the other hand, firms that grow through strategies. As a consequence, such firms tend to
acquisitions often experience problems in their exhibit numerous commonalties across their IT
efforts to integrate newly acquired units with infrastructures and applications (Boynton et al.
their existing businesses. The strategy literature 1992; Brown and Magill 1994; Cash et. al.,
advances three reasons why acquiring firms 1992). A centralized locus for IT decision making
might conclude that the costs of integrating orga- makes it possible to exploit the IT-related com-
monalties
monalties and
and
leverage
leverage
considerable
considerable
economies As
As aa point
economies pointofofclarification,
clarification,
these
these
exploitation
exploitation
of
of scope.
scope.However,
However, when when
a firm's
a firm's
operating
operating
units strategies
strategiesfor
units foreconomies
economies
of scope
of scope
are only
are only
feasi-feasi-
are
arecharacterized
characterized by unrelated
by unrelated
product-markets,
product-markets, ble
ble when
whencorporate
corporatediversification
diversificationmodemode
or or
these
theseoperating
operating unitsunits
oftenoften
possesspossess
distinctive
distinctive
IT breadth permits the leveraging of scope
IT
needs.
needs.AsAsa result,
a result,locallocal
IS staff
IS staff
and lineand
managers economies.
line managers
are
aregenerally
generallymuch
much
better
better
positioned
positioned
than thethan the
corporate
corporate IS IS
staff
staff
to effectively
to effectively
apply ITapply
in sup-
IT in sup-
porting
portingthe theoperating
operating
units'units'
work processes
work processes
and Absorptive
and Capacity
business
business strategies.
strategies.
As information technologies become a key ele-
As
As discussed
discussedabove,
above,
diversification
diversification
mode andmode andment of the business strategies or core operating
breadth
breadthinfluence
influence
the feasibility
the feasibility
of attaining
of attaining
processes, firms are seeking to stimulate IT-based
economies
economies of of
scope
scope
through
through
IT governance
IT governanceinnovation by assigning more authority for IT
arrangements.
arrangements. In addition
In addition
to these,
to these,
a third a
impor-
third impor-
decisions to line managers (Rockart 1988).
tant
tantinfluence
influence on the
on the
modemode
of IT governance
of IT governance
is is
However, the level of IT-related knowledge pos-
the
theexploitation
exploitationstrategy
strategy
for economies
for economies
of scope.of scope.
sessed by business managers across the enter-
Specifically,
Specifically, firms
firms
can seek
can to
seek
exploit
to exploit
the theprise and their willingness to accept IT decision-
economies
economies of of
scope
scope
either
either
through
through
development
development
making responsibilities can significantly influ-
of
of enterprise-wide
enterprise-widepartnerships
partnerships
or through
or through
con- con-
ence design choices about the location of author-
solidation of enterprise-wide IT assets. ity for IT management decisions. The conceptual
Exploitation of economies of scope through part- notion of absorptive capacity (Cohen and
nerships reflects the realization that relevant Levinthal 1990), which refers to the ability of a
knowledge and political authority for IT-based firm's employees to develop relevant knowledge
innovation is dispersed across the enterprise. bases, recognize valuable external information,
Partnerships are required to integrate these dis- make appropriate decisions, and implement
persed pockets of knowledge and authority. Such effective work processes and structures, nicely
partnerships enable the formation of trust among captures these notions.
business and IS managers so that they are willing
to jointly explore prospects for IT innovation in
Absorptive capacity is reflected in the line man-
the firm, assume risk for IT innovation, and
agers' IT management experience in a firm. If
undertake leadership for the innovation projects
line's IT management experience is limited, little
(Cooprider and Victor 1993; Henderson 1990).
IT-related absorptive capacity will be present out-
However, such partnership relations are more
side of the knowledge possessed by the firm's IT
likely to prosper when the decision rights for sig-
specialists. A firm's capability to effectively
nificant IT activities are located closer to the
implement and sustain a decentralized locus for
business units, for example, through decentral-
IT decisions is largely predicated upon the extent
ization at the divisional level (Rockart 1988).
to which line managers throughout its operating
On the contrary, when firms seek to exploit units possess the requisite level of IT-related
economies of scope through IT assets, the prima- absorptive capacity. Such an absorptive capacity
is normally built over time through these busi-
ry logic is that the dispersed IT assets across the
enterprise must be consolidated and managed ness managers' interaction with IS staff and
through a centralized entity, viz. corporate IS. their participation in significant IT initiatives.
The dispersed IT assets are consolidated through Firms whose line managers lack an adequate
unified systems and procedures, planningunderstanding about IT and IT management
approaches, and a focus on leveraging the ITpractices are, stated simply, unprepared for
assets across the enterprise's information needsand
( unwilling to support significant IT decen-
(Keen 1991; LaBelle and Nyce 1987; Maglitta tralization (Boynton et al. 1992; Brown and
and Mehler 1992; Rockart 1988; VenkatramanMagill 1998). On the other hand, when firms'
1994; von Simson 1990; Wilder 1990). Efforts line
to managers are knowledgeable about IT,
achieve such intentions generally benefit from they might seek greater authority for certain IT
centralized governance arrangements. management decisions. Therefore, in such
MIS Quarterly Vol. 23 No. 2/June 1999 267
firms,
firms, there
there might
mightbe
begreater
greaterpressures
pressurestoward
toward contingency
contingencyforces
forcesthat
thatapply
apply
to to
a given
a given
con-
con-
decentralization. textual situation. Based on Gresov's (1989)
work, three mutually exclusive scenarios-rein-
forcing, conflicting, and dominating-are iden-
tified to characterize this interplay and provide
The Study's Theoretical Model - a basis for positing the influence of multiple
contingencies on the location of IT decision
The previous section has identified three major
rights.
sets of forces influencing the assignment of
decision rights for IT activities. However, the Reinforcing contingencies describe situations
current study recognizes the reality that organi- where the salient contingency forces produce
zations are subject to the pulls and pressures of similar influences on the locus of IT decision-
multiple, rather than singular, contingency making responsibilities. These salient contin-
forces. A pragmatic theoretical perspective must gency forces combine to reinforce a singular
acknowledge the simultaneous influence of effect either toward a centralized or a decentral-
these forces. ized IT governance mode. While countervailing
dispositions might be produced by less salient
The theory of multiple contingencies provides a
contingency forces, the effects of these sec-
sound theoretical lens for viewing the influence
ondary forces are overwhelmed by the reinforc-
of several contingency forces on the locus of IT
ing contingencies. In other words, in such orga-
decision rights. This theory is implicit in theories
nizations, the most salient forces all tend to rein-
of organization design (Lawrence and Lorsch
force each other to produce a strong tendency
1969; Thompson 1967) and has recently been toward either centralization or decentralization.
used to study the design of work units (Gresov
This leads to the study's first hypothesis:
1989). Researchers have recently drawn upon
this theory to examine interdependencies Hypothesis 1: Organizations facing reinforcing
between organizational contexts and the nature contingencies regarding corporate gover-
of the IS role and function (Fisher 1994; Rathnam nance, scope, and absorptive capacity forces
et al. 1995; Sabherwal and Kirs 1994; Zerfane are likely to exhibit a centralized or decen-
and Gul 1993). Essentially, the theory recognizes tralized mode of IT governance.
work unit design as being simultaneously influ-
Conflicting contingencies refer to situations
enced by numerous contingency forces, whose
where multiple contingency forces are salient,
effects might complement or counteract one
but they have conflicting influence on the loca-
another. Thus, an organization's design is an
tion of IT decision-making responsibilities. Some
evolving consequence of the interplay among
of these salient contingency forces would be
different forces. Conceptually, individual contin-
expected to favor a disposition toward the cen-
gency forces influence the choice of particular
tralization of IT decision making, whereas other
organization designs.
forces would be expected to favor a disposition
Multiple contingencies theory directs our atten- toward decentralization. With this scenario,
tion toward two fundamental issues. First, in a these conflicting effects resolve themselves
given set of situations, the individual contin- through the emergence of a federal mode of IT
gency forces will likely differ in their salience, governance, where certain IT decisions are cen-
i.e., relative strength, regarding work unit tralized and others are decentralized.
design. Organization designers or managers in
Hypothesis 2: Organizations facing conflicting
one contextual situation will invariably perceive
contingencies regarding corporate gover-
certain contingency factors to be highly salient,
nance, scope, and absorptive capacity
while organization designers or managers in
forces are likely to exhibit a federal mode of
other contextual situations are likely to perceive
IT governance.
these same factors as being less salient. Second,
the effects of individual contingency forces are Dominating contingencies reflect the situation
amplified, dampened, or overridden depending where, despite the presence of numerous contin-
on the nature and magnitude of the full set of gency factors, one or two contingencies impose
prohibitively
prohibitivelyhigh
high
opportunity
opportunity
costs
costs
on firms
on firms
that that nance, scope, and absorptive
absorptive capacity
capacity forces
forces
ignore
ignoretheir
theireffects.
effects.
As As
a result,
a result,
firms
firms
facing
facing
such such are likely to exhibit
exhibit centralized
centralized or
or decentral-
decentral-
dominating
dominatingcontingencies
contingencies
are are
forced
forced
to behave
to behave ized modes of IT governance.
governance.
as
as if
if these
thesedominating
dominatingcontingencies
contingencies
werewere
the the
only
only ones
onespresent.
present.
Two
Two examples
examplesillustrate
illustrate
thatthat
dominating
dominating
contin-
contin-
Research Methodology
gencies
genciesdo donot
notsimply
simplyrepresent
represent
an academic
an academic
exercise.
exercise.First,
First,consider
consider
a firm
a firm
withwith
related
related Data for testing the three hypotheses were drawn
operating
operatingunits
unitsthat
that
hashas
primarily
primarily
grown
grown from a research project that was initially
through
throughacquisition
acquisitionand
and
whose
whose
acquired
acquired
unitsunits designed to understand factors affecting the
have
have disparate
disparateITITinfrastructures.
infrastructures. Although mar- mar- modes of IT governance arrangements in busi-
Although
ket
ket relatedness
relatednesscouldcould present
present prospects
prospectsfor for ness firms (Rockness and Zmud 1989). The data
leveraging
leveragingeconomies
economies of of
scope
scope
through
through
a cen-a cen- were gathered through multiple case studies of
tralized
tralizedITITgovernance
governance mode,
mode, the the
disparate
disparate
IT IT the governance arrangements at eight firms.
infrastructures
infrastructures ofof
thetheacquired
acquiredbusiness unitsunits Eisenhardt (1989) suggests that researchers con-
business
could
could present
presentprohibitive
prohibitive costs
costs
to impede
to impede ducting multiple sites case studies should identi-
attempts
attemptsatatenterprise-wide
enterprise-wide integration
integration
of ITof IT fy their theoretical unit of interest and carefully
assets
assets and,
and,hence,
hence, thetheadoption
adoption of more
of more
cen- cen- select the case sites so as to produce sufficient
tralized
tralizedITITdecision-making
decision-making arrangements.
arrangements. Here,Here, variation in the variables of greatest research
diseconomies
diseconomiesofof scope
scopeduedueto growth
to growth through
through interest. Accordingly, first, a field survey was
acquisitions
acquisitionscould
could become
become a dominating
a dominating con- con- conducted to gather data on IT governance
tingency
tingencyinfluencing
influencing thethedecentralization
decentralizationof ITof IT arrangements in 35 firms. The data included the
decision
decisionmaking,
making, even
eventhough
though market
market related- IT decision roles of corporate IT, divisional IT,
related-
ness
ness might
mightotherwise
otherwise suggest
suggestthe the
desirability
desirability
of of and line management at 73 business units within
a centralized
centralizedgovernance.
governance. As As
a second
a second
exam- exam- the 35 firms. Details of this field survey and the
ple,
ple, consider
considera afirm
firmwithwith
unrelated
unrelatedoperating
operating manner in which the locus of IT decision making
units
units whose
whoseline
linemanagers
managers have have
limited
limited
IT IT was measured are described in Appendix A.
knowledge
knowledgeand andexperience
experience and,and,
therefore,
therefore,
low low
These survey data were used to select the case
absorptive
absorptivecapacity.
capacity. TheThe
inability
inability
and and
(likely)
(likely)
sites as follows. First, data on the locus of IT deci-
limited
limitedwillingness
willingness of of
these
these
lineline
managers
managersto to
sions for infrastructure, technology use, and pro-
assume
assumesignificant
significant IT IT
responsibilities
responsibilities might might
ject implementation decisions across business
inhibit
inhibitany
anyefforts
effortsto to
assign
assign
responsibility
responsibility
for a for a
units of each sampled firm were examined to
significant
significantportion
portionof of
IT IT
decisions
decisions
to the
to operat-
the operat-
detect if there were any significant variations
ing
ing units.
units.Thus,
Thus,even
eventhough
thoughlow low
market
market
relat-relat-
across the units. In almost all the 35 firms, the
edness
ednessmight
mightotherwise
otherwise suggest
suggest
the the
desirability
desirability
locus of authority for IT decisions was found to
of
of adopting
adoptinga adecentralized
decentralized IT governance
IT governance
be similar across the three sampled business
mode,
mode,thethedominating
dominating nature
nature
of low
of low
absorptive
absorptive
units of the enterprise. This provided the justifi-
capacity
capacitycould
coulddictate
dictatethethe
reality
reality
of aofcentral-
a central-
cation for inferring an overall firm-level pattern
ized
ized IT
ITgovernance
governance mode.
mode.
in the locus of authority for IT decisions. Next,
Given
Giventhetheprohibitively
prohibitively high
highopportunity
opportunitycosts costs the locus scores were examined across the 35
associated
associatedwith
withignoring
ignoringthethe
effects
effects
of dominat-
of dominat- firms to select firms that exhibit very different
ing
ing contingencies,
contingencies,it it
is expected
is expected
thatthat
firms will will patterns. In addition, COMPUSTAT data were
firms
design
designIT ITgovernance
governance modes
modesin response
in response
to the
to the gathered on sales revenues over the last three
singular
singulareffects
effects ofofthethe
dominating
dominatingcontingency.
contingency. years prior to the data gathering effort. These
Such
Such governance
governance arrangements
arrangements are,are,
therefore,
therefore, data were used as a proxy for the size of the firm.
likely
likelyto tobebeeither
eitherpurely
purelycentralized
centralized
or decen-
or decen- Together, the locus scores and firm size were
tralized.
tralized.Therefore,
Therefore, thethe
study's
study's
third
third
hypothesis
hypothesis
is is used as the criteria for case site selection to select
stated as follows: the eight case sites (Table 3):
Hypothesis 3: Organizations facing dominating *Two firms (C and H) were selected because
contingencies regarding corporate gover- they reflected a centralization of infrastructure
decisions,
decisions,but
butmoderate
moderatelocus
locus
scores
scores
on IT
onuse
IT use the submission of a practitioner-oriented report
and
and project
projectmanagement
management decisions.
decisions. to the study's sponsors. Table 4 portrays how the
* Firm A was selected because it reflected a cen- case sites vary according to the conceptualiza-
tralization of infrastructure and project man- tions of IT governance developed for the current
agement decisions, but a decentralization of IT study. Four of the firms exhibit a federal mode of
use decisions. governance (sites A, C, G, and H), three exhibit a
* Two firms (F and G) were selected because decentralized governance (sites B, D, and E), and
they reflected a centralization of all IT deci- one a centralized governance (site F).
sions.
* Three firms (B, D, and E) were selected During the case site selection, it was also con-
because they reflected a decentralization of all sidered important that all the case sites be char-
IT decisions. acterized as competent regarding their IT prac-
* Four of the firms (A, B, C, and D) were large in tices. The questionnaire phase of the original
size, two others (E and F) were medium in size, study (Rockness and Zmud 1989) also provided
whereas the other two (G and H) were small in measures of the firms' success in using IT; there-
size. fore, only those firms exhibiting better-than-aver-
age success in deploying IT within their business
The study's conceptualization scheme of central- strategies and value chain activities were consid-
ization, decentralization, and federal modes of ered as potential case sites. If the firms were not
governance arrangements was developed subse- competent at managing IT, one reason for that
quent to the completion of the case studies and could be the selection of inappropriate IT gover-
Site Decentralization of IT Decisions" Firm Size Description

A Low on infrastructure and Large Industrial company operating in
project management, high on $3.25 billion very diverse markets
IT use
B High on all IT decisions Large Industrial chemicals firm

$2.5 billion
C Low
Low on
oninfrastructure,
infrastructure, Large Durable goods firm selling to
moderate on IT use and $1.5 billion both industrial and consumer
project management markets
D High on all IT decisions Large Multinational manufacturer of

$1 billion industrial durable goods and
components
E High on all IT decisions Medium Food processor
$600 million
F Low on all IT decisions Medium Single-product energy company
$420 million
G Low on all IT decisions Small Computer equipment
$230 million manufacturer
H Low on infrastructure, Small Specialty manufacturer in the

moderate on IT use and $200 million mechanical and chemical
project management bonding industry
aA low locus score corresponds to the centralization of IT decisions, while a high locus score
corresponds to the decentralization of IT decisions.
IZ . -y n SI bE~ I ECIOIYL I aId mie4C.pIt m.[ uL lIZ 114M qICTT-6IM t(i
Locus for Locus for Project

Infrastructure Locus for IT Use Management Mode of IT
Case Site Decisions Decisions Decisions Governance
Corporation A Corporate IS Divisional IS Corporate IS Federal
Corporation B Divisional IS Divisional IS Line management Decentralized
Corporation C Corporate IS Divisional IS Divisional IS Federal
Corporation D Divisional IS Divisional IS Divisional IS Decentralized
Corporation E Divisional IS Divisional IS Divisional IS Decentralized
Corporation F Corporate IS Corporate IS Corporate IS Centralized
Corporation G Corporate IS Line management Line management Federal
Corporation H Corporate IS Line management Line management Federal
nance
nance arrangements.
arrangements.TheTheselection
selectionstrategy
strategy used
used review
review ofofthe
thetype
typedescribed
describedearlier;
earlier;
consequent-
consequent-
here
here thus
thus ensured
ensuredthe
thecredibility
credibility ofof
thethe
gover-
gover- ly,
ly, the
the interviewers
interviewerswere
werereasonably
reasonably blind
blind
to the
to the
nance
nance arrangements
arrangementsasasbeing
beingeffective
effectivein in
those
those theoretical
theoreticalmodel
modelthat
thatframes
framesthethe
current
currentstudy.
study.
firms
firms and,
and, therefore,
therefore,worthy
worthy ofofexamination
examination in in
this study.
Data
Data Analysis
AnalysisStrategy
Strategy
After
After the
theproject
projectwas
wascompleted,
completed,
a practitioner
a practitioner
Data Gathering Strategy report
report was
wasgenerated
generatedtoto
report
report
upon
upon
thethe
trends
trends
in in
IT
IT governance
governancearrangements
arrangementsdiscovered
discovered
through
through
Telephone interviews at each case site were con-
the study (Rockness and Zmud 1989).
ducted by the second author of this paper in col-
Subsequently, one of the original principal inves-
laboration with another principal investigator
tigators (this study's second author) teamed with
(Rockness and Zmud 1989). At each case site,
another researcher (this study's first author), who
telephone interviews were conducted with the
was not involved in the original data gathering
firm's CIO and the senior IS manager for the three
analysis, to re-interpret the case studies and to
business units involved with the field survey. At
produce the current study. Appendix C describes
most of the case sites, the firm's CFO was also
details of the process by which qualitative analy-
interviewed. Thus, a total of four to five inter-
ses were conducted.
views were conducted at each firm. On the aver-
age, each of these interviews lasted just under Our analysis revealed that two case sites were
one hour. Transcripts for each case site were gen- categorized as exhibiting reinforcing contingen-
erated and returned to those interviewed in order cies (sites B and F). Four case sites were found to
to validate the researchers' interpretation. exhibit conflicting contingencies (sites A, C, G,
Appendix B provides the structured interview and H). Finally, two firms exhibited this pattern of
protocol used in conducting these interviews. As contingencies (sites D and E).
can be seen from Appendix B, the focus of the
interviews was to understand how key IT deci-
sions were made at a firm, the reasoning that
underlay the pattern observed at a firm, and the Results
historical basis of this pattern. At the time these
transcripts were initially developed, the inter- Findings are presented separately for each of the
viewers had not engaged in an extensive literature three categories of multiple contingencies: rein-
_ _ m _ __~1 = i V-1 i SO I tsi 1o IMEE _ _ I I a

Category of Contingency
Contingency Corporation B Corporation F
Contingencies Factors Factor Level Influence Factor Level Influence
Corporate Corporate Decentralized Decentralized Centralized Centralized

governance governance mode
Firm size Small Centralized
Economies of Diversification Acquisition Decentralized

scope mode
Diversification Low market Decentralized High market Centralized

breadth relatedness relatedness
Exploitation Partnerships Decentralized Partnerships Decentralized

strategy for scope IT asset Centralized
economies consolidation
Absorptive Line IT knowledge High Decentralized

capacity
Mode of IT Governance Decentralized Centralized
Notes: The boldfaced cells indicate the salient contingencies at each firm
corresponding specific contingency factor was not identified by manag
influence. The other cells indicate less salient contingency factors.
forcing,
forcing, conflicting,
conflicting,
and dominating.
and The tabular With Corporation
dominating. B, the salient corporate gover-
The tabular
summaries
summaries distinguish
distinguish relative nance
between the between and scope
the pressures reinforce one another
relative
strength
strength of the
offorces
thefound
forces
to be found
influencing be in
tothe creating a strong
influencing the governance disposition
firms'
firms' IT governance
IT governance
mode (Tables
mode
4, 5, and 6). toward
(Tables 4, decentralization.
5, and 6). Corporation B operated
The
The forces
forces
producing
producing
strong effects
strong
are termed in a very
effects dynamic
are termed marketplace (specialty mar-
"salient"
"salient"contingencies,
contingencies,
while weakerwhile
forces are kets), where
weaker marketare
forces trends had to be identified
termed
termed "other"
"other"
contingencies.
contingencies. and reacted to quickly, and where acquisitions,
divestment, and joint venture activities had
become the norm rather than the exception. As a
Reinforcing
Reinforcing Contingencies
Contingencies result, the firm's overall corporate governance
Table
Table 5 illustrates
5 illustrates
the contingency forces at the modeforces
the contingency is characterized
at the by high decentralization.
two
two case
case
sites. sites.
As illustrated,
As illustrated,
all of the salient In of
all fact,the
the firm's
salientoperating philosophy envi-
contingency
contingency forces produce
forcessimilar
producedispositions sioneddispositions
similar each operating unit as being self-con-
toward
toward centralization
centralization F) or decen- tained.
(Corporation(Corporation F)Scope pressures emerged from low mar-
or decen-
tralization
tralization (Corporation
(Corporation
B). Therefore,
B).
these firms ket relatedness
Therefore, these among
firms the operating units, with a
are
are considered
considered
to exhibit
toreinforcing
exhibit contingen- result that
reinforcing these operating units had quite distinct
contingen-
cies.
cies. Further,
Further,
as might
asbemight
inferred from
be inferred
knowl- information
from knowl-system requirements. Two of the
edge
edge of the
of likely
the effect
likely of effect
these salient theseother
ofcontin- pressures
salient (an "acquisition" diversification
contin-
gency
gency forces,
forces,
Corporation
Corporation
B displayed a decen- mode and aadesire
B displayed decen-to build IT/line partnerships)
tralized mode of IT governance while also reinforced the governance disposition
Corporation F displayed a centralized mode. toward decentralization. Only a single, though
weak,
weak,contradictory
contradictory disposition
disposition
toward toward requests,
requests,inindefining
central- central- defining thethe
essential
essential
features
features
of of
ization
izationexisted:
existed:leveraging
leveraging
economies
economies
of scope of scopeapplications,
applications,andandin in
software
softwaresourcing.
sourcing.
In some
In some
through
through thetheconsolidation
consolidation
of IT assets.
of IT This
assets.
factor cases,
cases, line
This factor linemanagers
managers undertook
undertook applications
applications
produced
produced a desire
a desire
for for
the corporate
the corporate
IS groupIS togroup development
to
development without
without thethe
involvement
involvement
or aware-
or aware-
serve
servea astrong
strong rolerole
in setting
in setting
enterprise-wide
enterprise-wide ness
ness of
ofthe
thecorporate
corporate IS group.
IS group.
Thus,
Thus,
applications
applications
strategic
strategic IT IT
direction.
direction. are
are acquired
acquiredininthree
three ways:
ways:
by line
by line
managers
managers
using
using aafourth
fourth generation
generation language,
language,
by the
by cor-
the cor-
As
As aaconsequence
consequence of these
of these
reinforcing
reinforcing
contin- contin-
porate
porateISISgroup,
group, ororfrom
froman outside
an outside
vendor.
vendor.
The The
gencies,
gencies, Corporation
Corporation B hadBdecentralized
had decentralized
the the
corporate
corporateISISgroup
groupis always
is always
involved
involved
whenwhen
locus of its infrastructure and IT use decisions
applications
applicationsare
are
developed
developed
through
through
the latter
the latter
two two
and highly decentralized its project implementa-
means.
means.InIngeneral,
general,Corporation
Corporation
F's management
F's management
tion decisions. Most IT-related decision makingviews the IS function to be somewhat more
was, in fact, being handled by the divisional IS
decentralized than other staff departments.
staff. The corporate IS group served three major
roles: development of corporate-wide infrastruc-However, during the interviews, respondents
ture standards, with the divisions retaining themade it clear that they believed their firm's small
autonomy to conform to these standards; plan-size and the homogenous nature of their markets
ning, development, and operation of corporate-offered very real economies through the central-
wide applications; and approval of all major ITization of IT governance. Hence, even though
projects from the perspective of their fit with the line managers possessed high IT knowledge and
corporate infrastructure standards. Overall, thethere was a desire to build partnerships
qualitative evidence confirms the decentralizedbetween the line and IS managers, the salient
IT governance mode at Corporation B. contingencies stem from size and market relat-
edness. These salient contingencies reinforce
At Corporation F, the salient reinforcing contin-
each other to influence the emergence of a
gencies involved corporate governance and centralized mode of IT governance at
scope pressures. Corporation F consisted of a
Corporation F.
number of small subsidiaries located in the
Observations regarding the forces influencing
same geographic area and the firm's manage-
IT governance modes at Corporations B and F p
ment personnel were housed within a central
vide strong support for Hypothesis 1: reinforc
facility. Additionally, the firm was characterized
by high market relatedness: 95% of the sales contingencies are associated with a centralized
revenues were generated by a single producta decentralized mode of IT governance.
category. Not surprisingly, given this context,
Corporation F had adopted highly centralized Conflicting Contingencies
decision-making governance-subsidiary man- Table 6 summarizes the findings regarding
agement handled day-to-day operations, but four case sites that were judged to represent t
most major decisions were made at the corpo- conditions of conflicting contingencies
rate level. Corporations A, C, G, and H. Each one of th
firms exhibited a federal mode of IT governan
Two other contingency factors were observed at
Corporation F: the desire to enhance partnershipWith Corporation A, strong corporate gove
relations between IS and line managers and the nance and scope pressures produced opposin
fairly high IT absorptive capacity possessed by dispositions toward the centralization and
line managers throughout the firm. Although decentralization. Corporation A was structured
Corporation F was a late-adopter of IT relative around
to four operating groups, with each group
its competitors, it had vaulted ahead of these comprised of numerous business units within
various (often distinct) automotive and construc-
competitors in infusing IT within its value chain
activities. Although corporate IT held ultimate tion lines-of-business. The firm operated in mar-
decision authority for most IT decisions, line kets of low relatedness: a large, high market
share manufacturing business, a commodity
managers across the firm were very active in ini-
business where the firm seeks to be the low cost
tiating and prioritizing application development
C,)
0
Lb
Q)
Category of Contingency Corporation A Corporation C Corporation
Contingencies Factors Factor Level Influence Factor Level Influence Factor Level Influ
Corporate Governance Decentralized Decentralized |Decentralized Decentralized Decentralized Cent

"I
governance mode
(0
Iz
Firm size Small Cen
Geographic High Decentralized

dispersion
Economies Diversification Internal Centralized Acquisition Decentralized Internal Central

of scope mode
Diversification Low market Decentralized High market Cent

breadth relatedness relatedness relatedness
Exploitation IT asset Centralized

strategy for consolidation
scope
economies
Partnerships Decentralized Partnerships Decentra
Absorptive Line IT Low Centralized High Decentralized

capacity knowledge
Mode of IT governance Federal FeeralFederal Federal
Notes: The boldfaced cells indicate the salient contingen

identified by managers as being an important influence.
producer,
producer, another
another
business
business
which emphasizes
which emphasizeswith
with these
these
salient
salient
forces, conflicting
forces, conflicting
disposi- disposi-
the
theniche
nichemarketing
marketing
of specialty
of specialty
items, anditems,
a tions
tions
and awere
were
also observed
also observed
regarding other
regarding
less other less
small
smallhigh-technology,
high-technology,high-growth
high-growth
business. business.
salient
salient contingency
contingencyforces. The
forces.
fairly large
Thegeo-fairly large geo-
Individual
Individual operating
operating
groupsgroups
enjoyed consider-
enjoyed consider-
graphic
graphic distances
distances
separating
separating
many of themany
firm's of the firm's
able autonomy in order to enable effective operating
operating divisions
divisions
induced pressures
inducedtowardpressures toward
strategies across these markets of low related- the
thedecentralization
decentralization of IT decision
of IT making,
decision making,
ness. This decentralized corporate governance whereas
whereas the the
relatively
relatively
high market-relatedness
high market-relatedness
mode along with low market relatedness had of
ofthese
thesedivisions
divisions
suggestedsuggested
that centralization
that centralization
strongly influenced Corporation A's mode of IT could be beneficial.
governance.
As a result of these countervailing dispositions,
However, Corporation A's senior management
was also very concerned with improving the IT governance arrangements at Corporation C
economies of scope through consolidation of IT reflected a federal governance structure: central-
assets. This factor produced a strong countervail- ized for infrastructure decisions, yet decentral-
ing disposition toward centralization. The firm ized for both IT use and project implementation
had a history of building market share through decisions. Significantly, decentralization stopped
aggressive cost reduction. As a result, consider- with the divisional IS staff; line management
able emphasis was placed on implementing played a fairly passive role regarding IT decision
common applications and sharing, where advan- making.
tageous to do so, common resources and knowl-
edge; and the corporate IT group was expected Most of Corporation G's product lines had
to play an influential role in identifying avenues evolved from a closely related technology base.
for enhancing the efficient use of IT assets across Therefore, the firm was functionally organized
the corporation. and had historically employed a centralized
governance mode. This corporate governance
These conflicting contingencies produced an IT
mode has produced strong pressures to corre-
governance mode where divisional IS had deci-
spondingly centralize the firm's IT decision
sion authority for IT use, but the corporate IT
making. However, since it competed in a seg-
group had decision authority for both infrastruc-
ment of the computer industry, Corporation G's
ture and project implementation decisions. It is
managers were very knowledgeable about and
noteworthy that even the less salient contin-
interested in IT. Line managers often initiated IT
gency forces at Corporation A (an internal
projects, set IT priorities, and managed key IT
growth strategy and the desire to build IS/line
projects. Therefore, the high IT knowledge of
partnerships) were conflicting, and hence
line managers produced a disposition toward
impelled the movement toward a federal gover- decentralization, which conflicted with the
nance structure.
effects of the corporate governance mode.
These strong conflicting contingencies were
Corporation C, which had grown through a series
accompanied by a number of less salient con-
of acquisitions and divestitures, had steadily
decentralized its overall governance mode to contingencies. The firm's history of
flicting
internal growth and high market-relatedness of
enable each operating division to better config-
ure its work and management systems to itsits operations induced moderate pressures
spe-
toward centralization. On the other hand, a
cific market environment. While senior manage-
ment recognized the benefits of aligning its IT dissatisfaction among line management
growing
governance mode with its overall governance
with corporate IS staff had led senior manage-
structure, line managers across the firm hadment
lim-to recognize the necessity of building bet-
ited experience (and little desire to gain such
ter partnership relations between corporate IS
experiences) with IT decision making. Thus,
and line managers (and, hence, decentralize IT
while line management knew where the businessdecision making). Despite building a strong IT
was going, the firm's corporate and divisional IS
infrastructure and implementing critical appli-
staffs were seen as being best positioned to cations,
react corporate IS had little credibility with
technologically to this business direction.lineAsmanagers.
Thus,
Thus, although
althoughinfrastructure
infrastructure
decisions
decisions
werewere
cen- cen- Table 4, the four federal
federal governance
governance firms
firms exhib-
exhib-
tralized
tralizedat atCorporation
Corporation G,G,senior
senior
management
management ited variations in how
how the
the specific
specific responsibilities
responsibilities
believed
believedthat
thatmoving
moving IT IT
useuse
andand
project
project
imple-
imple- were allocated among
among corporate
corporate IS,
IS, divisional
divisional IS,
IS,
mentation
mentationdecisions
decisions
totoline
line
management
management wouldwould and line managers. Corporate
Corporate IS
IS assumed
assumed respon-
respon-
both
both (1)
(1)enable
enableline
line
managers
managers to become
to becomemoremore sibility for IT infrastructure
infrastructure management
management at
at all
all the
the
deeply
deeply involved
involvedininthethedevelopment
development activities
activities four case sites. However,
However, the
the locus
locus of
of authority
authority for
for
and
and better
betterunderstand
understandthe
the
true
true
costs
costs
and and
benefits
benefits IT use and project management
management varied
varied across
across the
the
associated
associatedwith
withsuccessful
successful IT IT
use,use,
andand
(2) forge
(2) forge firms. In particular,
particular, line
line management
management assumed
assumed
improved
improvedrelationships
relationships between
betweenthethe
corporate
corporate
IS IS responsibilities for IT
IT use
use and
and project
project manage-
manage-
staff
staff and
andthe
thefirm's
firm's line
line
managers.
managers.Therefore,
Therefore, ment decisions at both
both Corporation
Corporation G
G and
and
Corporation
CorporationG'sG'sITITgovernance
governance structure
structure
reflects
reflects Corporation H. The high
high levels
levels of
of line
line IT
IT knowl-
knowl-
a federal
federalstructure.
structure. edge are responsible for
for the
the significant
significant role
role of
of line
line
management in these
these two
two firms
firms with
with federal
federal
Corporation
CorporationH's
H'stwo
twoprincipal
principal
areas
areas
of business
of business
modes of governance.
governance. InIn contrast,
contrast, Corporation
Corporation
were
were organized
organizedinto
into
corresponding
corresponding
business
business
C's federal mode of governance
governance does
does not
not involve
involve
groups;
groups;one onebusiness
business group
group hadhad
twotwo
subunits
subunits
and and
line management precisely
precisely because
because of
of their
their low
low
the
the other
otherfour
foursubunits.
subunits. Although
Although Corporation
Corporation
IT knowledge. Thus, line
line IT
IT knowledge
knowledge andand the
the
H's
H's relatively
relativelysmall
smallsize
size
hadhad
created
created
pressures
pressures
for for
consolidation of IT assets
assets appear
appear to
to be
be influencing
influencing
it
it to
to maintain
maintain(for
(for
reasons
reasons of of
cost)
cost)
centralized
centralized
IT IT
the differentiation of
of governance
governance forms
forms within
within the
the
governance,
governance,the thefirm
firm as as
a whole
a wholehadhad
decentral-
decentral-
federal mode of governance.
governance.
ized
ized its
itscorporate
corporategovernance
governance modemode
to enable
to enable
its its
operating
operatingmanagers
managers totomore
morequickly
quickly
respond
respond
to to Dominating Contingencies
Contingencies
competitors'
competitors'actions.
actions. InIn
addition,
addition,there
there
had had
beenbeen
Table 7 illustrates the
the contingency
contingency factors
factors felt
felt at
at
a growing
growingbelief
beliefatat
Corporation
Corporation H that
H that
the the
scopescope
the two case sites exhibiting
exhibiting dominating
dominating contin-
contin-
and
and priority
priorityofofITITapplications
applications werewere
bestbest
deter-
deter-
gencies: Corporation
Corporation D
D and
and Corporation
Corporation E.
E. Both
Both
mined
mined by byline
linemanagers.
managers. Correspondingly,
Correspondingly, the the
of these firms possess
possess purely
purely decentralized
decentralized IT
IT
firm's
firm's line
linemanagers
managers were
werebecoming
becoming veryvery
knowl-
knowl-
governance modes, thus
thus providing
providing strong
strong sup-
sup-
edgeable
edgeableabout
aboutITITandandwere
wereassuming
assuming increasing
increasing
port for Hypothesis
Hypothesis 3.
3.
responsibility
responsibilityfor
forstrategic
strategic
IT IT
applications.
applications.
Similarly,
Similarly,asassome
someofof the
the
operating
operatingunits
units
grewgrew
in in The dominating contingency
contingency that
that exists
exists for
for
size
size and
andwere
wereable
abletoto
support
supportfull-time
full-time
IS staff,
IS staff, Corporation D arises
arises from
from the
the firm's
firm's extremely
extremely
they
they had
hadbegun
beguntototake
takeover
over
an an
increasing
increasingnum- num- low market relatedness:
relatedness: Corporation
Corporation D D compet-
compet-
ber
ber of
of IT
ITdecision
decisionresponsibilities.
responsibilities. ed, both domestically
domestically and
and internationally,
internationally, in
in aa
number of distinct industrial
industrial markets,
markets, producing
producing
Thus,
Thus, as
asititgrew,
grew,Corporation
CorporationH had
H had
moved
moved
from
from finished goods as well
well as
as component
component parts.
parts.
a centralized
centralizedITITgovernance
governancemode
mode
to atofederal
a federal Further, it operated
operated manufacturing
manufacturing facilities
facilities in
in
structure,
structure,where
wherecorporate
corporate
IS handled
IS handled
most
most
infra-
infra- North America, South
South America,
America, and
and Europe.
Europe.
structure
structuredecisions
decisions and
and
line
line
management
managementhan-han- Because of the firm's
firm's low
low market
market relatedness,
relatedness,
dled
dled most
mostITITuseuseand
andproject
project
implementation
implementation divisional IS and line
line management
management were were per-
per-
decisions.
decisions.ISISstaff
staffhadhad
been
beenphysically
physically
moved
moved
to to ceived to be much better
better positioned
positioned to
to determine
determine
the
the larger
largeroperating
operatingunits
units
while
while
stillstill
reporting
reporting
to a to a the operating units' IT
IT requirements
requirements than
than were
were
corporate
corporateISISmanager.
manager. Such
Such
a movement,
a movement,in turn,
in turn, the corporate IS staff.
staff. Therefore,
Therefore, the
the firm
firm exhibit-
exhibit-
had
had spurred
spurredthetheline
line
management
management to develop
to develop
an an ed a very strong orientation
orientation toward
toward decentraliza-
decentraliza-
interest
interestin,
in,and
andparticipate
participate
in,in,
IT decision
IT decision
mak-mak- tion of IT decision making.
making. AsAs might
might be
be expect-
expect-
ing. As line managers' absorptive capacity ed, the firm's overall
overall corporate
corporate governance
governance mode
mode
increased, they assumed greater responsibility for was also quite decentralized:
decentralized: operating
operating divisions
divisions
IT use and project implementation decisions. were fully responsible
responsible for
for the
the vast
vast majority
majority of
of
The four case sites just assessed provide strong their strategic product-market
product-market andand value
value chain
chain
operation decisions.
support for hypothesis H2: conflicting contingen-
cies are associated with a federal IT governance Table 7 shows that the
the dominating
dominating contingency
contingency
mode. Interestingly, however, as evident from was consonant with other salient contingencies
_=UMIMn spomtatua ;s-
Category of Contingency
Contingency Corporation D Corporation E
Contingencies Factors Factor Level Influence Factor Level Influence
Corporate Governance Decentralized Decentralized Decentralized Decentralized

governance mode
Geographic High Decentralized

dispersion
Economies of Diversification Acquisition Decentralized Acquisition Decentralized

scope mode
Diversification Low market Decentralized High market Centralized

breadth relatedness relatedness
Exploitation Partnerships Decentralized

strategy for scope
economies
Exploitation IT asset Centralized

strategy for scope consolidation
economies
Absorptive Line IT Low Centralized

capacity knowledge
Mode of IT Governance Decentralized Decentralized
Notes: The cells with double bold lines indicate dominating c

indicate the salient contingencies at each firm. Blank cells in
factor was not considered to be an important influence. Oth
factors.
reflecting
reflectingcorporatecorporate
governance (decentralized
governance located (decentralized
located "deep"
"deep"within
withinoperating
operating divisions,
divisions,so so
as as
governance)
governance) and economies
and scope (acquisi- toof
of economies enable
enable each
eachbusiness
scope business unit
(acquisi-unittotoshape
shapeanan
IT IT
tion
tiongrowth
growth
mode) pressures.
mode) In fact,
pressures.
the less infrastructure
In fact,
infrastructure and the less
anda aportfolio
portfolioofof
IT IT
applications
applications
salient
salientcontingencies
contingencies
(desire to build line/IS
(desire to build
for its market situation. line/IS
partnerships
partnerships and high geographic
and high dispersion)
geographic dispersion)
also
alsoare consonant
are consonant
with the dominating
with theThe
contin- dominating contingency
dominating contin- observed with
gency.
gency. However,
However,
interviews madeinterviews
it quite clear Corporation
made Eit quite
involved clearof growth
its strategy
that
that the mode
the ofmodeIT governance
of IT was governance
primarily through acquisitions.
was primarilyA major food processor,
influenced by the low market relatedness. Corporation E has primarily grown through the
Interviewees constantly stressed that since their acquisition of regional food companies, with the
firm was diversified into different businesses acquired firms becoming regional divisions.
and across several countries, each operating Reflecting their origins, the acquired divisions
division had to focus on the unique nature of tended to have vastly differing management
its particular market and singularly shape its philosophies, operating styles, and governance
business strategies and operations. As a conse- systems. They also had very distinct IT strategies,
quence, IT decision responsibilities were being architectures, applications, and equipment. The
firm
firm might
mightbest
bestbebedescribed
describedas as
a relatively
a relatively
loose
loose Summary
confederation of autonomous divisions.
Corporation E had adopted a decentralized Table
gov-8 summarizes the results of the analyses of
ernance mode in order to accommodate the the case sites. Strong support was found for all
pressures of its growth through acquisitions. three of the hypotheses derived from the applica-
tion of the multiple contingencies theory toward
Table 7 shows that two salient contingencies
the modes of IT governance in firms. An inter-
(low line IT management knowledge and high
pretation of these findings as well as a discussion
market relatedness) acted in a direction that is of their implications to both research and prac-
opposite to the effects of this dominating contin- tice is provided below.
gency. The firm's high market relatedness was
recognized as being ideally suited to a central-
ized approach to IT decision making. Further,
Discussion
perhaps because of its being a "loose confeder-
ation" of autonomous operating units, Corpor- Scholars seeking explanations of the variance in
ation E had very lean managerial staffs at both firms' modes of IT governance arrangements
corporate and divisional levels. Because of the have often restricted their search to the influence
workload pressures that accompany lean man- of a single variable. Even when researchers have
agerial structures, line managers were very focused upon the influences of several factors in
reluctant to take on significant IT authority and a single study, they have examined these factors
would have preferred to move as many IT as though they operate in isolation. This study
responsibilities as possible from their divisional was motivated by the realization that, in reality,
IS staffs to a corporate staff. Thus, Corporation these contingency factors operate together in
E exhibited a dominating contingency that was shaping firms' IT governance arrangements. The
dissonant with the effects of other salient con- results of this examination of the forces in play at
tingencies. During the interviews, this dissonanteight case sites indicate that the theory of multi-
situation was underscored in statements that the ple contingencies provides a useful frame for
firm's senior management regularly examined understanding how these contingency forces
the feasibility of moving toward a more cen- operate together in influencing the location of IT
tralized IT governance. Senior management had decision rights.
recognized that significant opportunity costs
were being borne by not moving toward a cen- Prior to discussing the implications of these find-
tralized approach to IT decision making and that ings, first the study's limitations are summarized
the current IT governance represented a subopti- so that the findings can be placed in a proper
mal design. However, as each acquired division perspective. Then, implications of these findings
had retained its unique IT infrastructure and appli- are provided for both research and practice.
cations, the exorbitant costs associated with con-
solidating and standardizing IT architectures and The Study's Limitations
applications inhibited senior management from
moving toward a more suitable IT governance The salience of individual contingency factors is
arrangement. subjectively assessed by retrospectively drawing
upon managers' explanations for why their firms
Evidence obtained from these two case sites sug- chose the existing IT governance arrangements.
gests that, consistent with the predictions of Two potential concerns may surface with such a
Hypothesis 3, observed IT governance under a strategy. First, managers could be applying faulty
dominating contingency would reflect either a assessments of the salience of their firm's contin-
centralized or a decentralized mode. In the two gencies. However, the fact that four to five man-
instances observed within the case sites, the agers were interviewed at each firm and inter-
dominating contingency resulted in a decentral- pretations were based on this collective rather
ized IT governance mode. than on a single individual, we believe greater
wmg&.- zr me migam - 1- -
Hypotheses Supported? Findings
H1 Reinforcing contingencies Yes Corporation B exhibited a decentralized mode of

will induce a centralized IT governance primarily because of its decentralized
or decentralized mode of corporate governance mode and operations in
IT governance markets of low relatedness.
Corporation F exhibited a centralized mod
governance primarily because of its small size,
centralized corporate governance, and operations in
highly related markets.
H2 Conflicting contingencies Yes Corporation A operates in markets of low relatedness

will induce a federal mode and had a decentralized mode of corporate
of IT governance governance. However, with the recent desire to
consolidate IT assets, it experienced conflicting
contingencies. As a result, corporate IS managed IT
infrastructure and projects, whereas divisional IS
managed IT use.
Corporation C has grown through acquisitions and
has a decentralized mode of corporate governance.
However, line managers have low IT knowledge. As a
result of these conflicting contingencies, corporate IS
manages the IT infrastructure, while divisional IS
manages IT use and projects.
Corporation G has a centralized mode of corporate
governance. However, line managers have high IT
knowledge. As a result of these conflicting
contingencies, corporate IS manages the IT
infrastructure, while line management has authority for
IT use and project management.
Corporation H is a small firm. However, its line
managers have high IT knowledge and operate in a
decentralized mode of corporate governance. As a
result of these conflicting contingencies, corporate IS
manages the IT infrastructure, while line management
as authority for IT use and project management.
H3 Dominating contingencies Yes Corporation D operates in markets of low relatedness.

will induce a centralized As a result, divisional IS manages the IT infrastructure
or decentralized mode of and IT use, while line management has authority for
IT governance. project management.
Corporation E has grown primarily thro
acquisitions. As a result of the prohibit
integrating acquired businesses, divisio
authority for all IT activities.
MIS Quarterly Vol. 23 No. 2/Jun
validity can be ascribed to the measures of practice.

practice.Recall
Recallthat
that
we had
we had
assessed
assessed
the appro-
the appro-
salience. Second, the researchers' own biases priateness
priatenessand
and effectiveness
effectivenessof these
of these
firms'firms'
IT IT
may be confounding these measures of salience. management
management practice
practice
in 1988.
in 1988.
WhileWhile
we do we
not do not
However, as explained earlier, the researcher have
havesimilar
similar
data
dataabout
about
their
their
contemporary
contemporary
per- per-
who assessed the salience of the contingency formance, anecdotal evidence indicates that
factors did not participate in developing assess- these firms do continue to be cited in the trade
ments of the modes of IT governance. Further, the press for the quality of their IT management prac-
transcripts used for the analysis had already been tices. Still, the absence of rigorous evidence
read and validated by the interviewees. regarding the contemporary performance of
these firms might raise concerns about the gener-
Another potential concern regarding our meth-
alizability of our findings. However, since this
ods relates to the relatively small number of case
study focuses on the relationship between pro-
sites. However, as our case site selection process
files of multiple contingencies and modes of IT
insured that considerable variation was observed
governance, not on specific practices or tactics
in the modes of IT governance across the case
evidenced by these firms, we do regard our
site firms, we are confident that our implications
results as generalizable to contemporary firms.
are relevant to most large organizations.
Finally, the focus of the current study was to
A third concern might be raised regarding the
explain the formation of IT governance modes at
age of this data set (collected in the late 1980s).
an enterprise level of analysis. However, the
Is this study relevant given today's IT manage-
ment context? We would ourselves be con- eight case sites in our study exhibited relatively
undifferentiated governance arrangements across
cerned along these lines if the intent of this
their business units. Brown (1997) argues that
study was to document contemporary trends
when business units exhibit similar governance
regarding the modes of IT governance.
arrangements across the enterprise, it is appropri-
However, as our goal instead is to assess fun-
ate to examine the influential contingency con-
damental explanations underlying the design of
ditions at the level of the enterprise. However, if
firms' IT governance arrangements, we believe
significant variations exist in the mode of gover-
the data set to be reasonable and that findings
similar to those obtained here would be nance across business units, Brown presents evi-
dence that factors at the business unit level might
observed with data collected from contempo-
provide stronger explanations of the governance
rary organizations. We strongly encourage such
research efforts. A more relevant concern relat-
arrangements within individual units. Therefore,
we are satisfied that our focus at the enterprise
ed to the age of the data relates to the chang-
level is appropriate for this study.
ing landscape of the IT marketplace over the
past few years. Today, external partners (e.g.,
vendors, systems integrators, and consultants)
Implications for Research
serve a very active role in many, if not most,
firms' IT decision making arrangements. This study offers a robust conceptual model with
Therefore, our conceptualization of IT gover- theoretical linkages for understanding the loca-
nance should be extended in future research to tion of a firm's IT decision responsibilities.
Although previous studies have identified contin-
incorporate these external partners as an addi-
gency factors that are thought to influence firms'
tional actor in measuring locus of authority.
modes of IT governance arrangements, explana-
However, we are confident that the underlying
tions for their effects have been relatively sparse.
explanations derived from the theory of multi-
ple contingencies would remain valid given One of the significant contributions of our work
such an extension. is the identification of corporate governance,
economies of scope, and absorptive capacity
forces as providing underlying explanations for
Further, another possible concern related to the
the effects of specific contingency factors. These
age of the data might question whether or not the
forces have strong roots in the contemporary
companies that constituted our 1988 study have
remained exemplars of good IT managementresearch on organization theory, strategy, and
innovation.
innovation.We
Wefeel
feel
that
that
these
these
forces
forces
can serve
can serve
as as Simson
Simson 1990;
1990;Wilder
Wilder 1990).
1990). Our Ourstudy
studyprovides
provides
powerful
powerfultheoretical
theoreticallenses
lenses
bothboth in interpreting theoretical
in interpreting theoreticalinsights
insightsinto into thisthisphenomenon
phenomenon by by
the
the results
resultsofofprior
priorinvestigations
investigations and and in shaping linking
in shaping linking this
thisgovernance
governance mode
mode withwithconflicting
conflicting
rigorous
rigorousresearch
researchmodels
models forfor
future
futureinquiry.
inquiry. contingencies.
contingencies.Specifically,
Specifically, mostmostbusiness
businessfirmsfirms
are
are experiencing
experiencingthe the simultaneous
simultaneous andandcompet-
compet-
The
The second
secondmajor
majorcontribution
contribution lies lies
in our
in our
appli-
appli- ing
ing pulls
pullsand
andpressures
pressures ofoftheir
theircompetitive
competitive strat-
strat-
cation
cation of ofthetheuse
useofofthetheTheory
Theory of Multiple
of Multiple egy,egy, desire
desireto toexploit
exploit enterprise-wide
enterprise-wide synergies,
synergies,
Contingencies
Contingenciestoto provide
provide a more
a morerobustrobust
explana-
explana- and
and the
the growing
growingITITknowledge
knowledge of of
their
their
lineline
man-man-
tion
tion of
ofthe
thedeterminants
determinants of of
firms'
firms'
IT governance
IT governance agers.
agers. The
Thefederal
federalmode mode ofof governance
governance allows
allows
arrangements.
arrangements.For Forexample,
example, thethe
more more
sophisti-
sophisti- unbundling
unbundlingthe theauthority
authority forfordistinct
distinctsegments
segmentsof of
cated
cated lens
lensoffered
offeredbybythethemultiple
multiplecontingencies
contingencies IT IT activities
activitiesandandvesting
vesting thethelocus
locus
of of
authority
authorityfor for
perspective
perspectiveenables
enables
us us
to to
understand
understand whatwhat
other-other- those
those segments
segmentsininresponse
response to to
thetheeffects
effects
of dis-
of dis-
wise
wise would
wouldhavehavebeen
been a perplexing
a perplexing observation,
observation, tinctive,
tinctive,and
andconflicting,
conflicting, contingencies.
contingencies.
viz.,
viz., why
whyCorporation
Corporation E exhibited
E exhibited a decentral-
a decentral-
In
In terms
termsof offuture
futureresearch
research directions,
directions, there
there
is a is a
ized
ized IT
ITgovernance
governance mode
mode despite
despite
operating
operatingin in
highly
highlyrelated
relatedmarkets.
markets. WeWe strongly
strongly encourage
encourage great
great need
need totobetter
better understand
understand the the
dynamics
dynamics of of
others
othersto toconsider
considerhowhow this
this
theoretical
theoreticalperspec-
perspec- organizational
organizational adaptation
adaptation of of
IT IT
governance
governance
tive
tive may
mayinform
inform other
otherIT IT
research
research
domains.
domains.
modes
modes in inresponse
responsetoto changing
changing organizational
organizational
and
and industry
industrycontingencies.
contingencies.WeWe
believe
believe
thatthat
Further,
Further,thethethree
threetypes
types
of of
contingencies
contingencies
identi-
identi- evolving
evolvingbusiness
businessoror
technological
technological
forces
forces
regu-
regu-
fied
fied in
inthis
thiswork
work have
have
interesting
interestingimplications.
implications. larly
larly alter
alterthe
thenature
natureand
and
relative
relative
strength
strength
of the
of the
Both
Both the
thereinforcing
reinforcing and
and
dominating
dominating contingen-
contingen- contingency
contingencyfactors
factorsand
and
demand
demand
accommoda-
accommoda-
cies
cies produce
producesimilar
similaroutcomes
outcomes in the
in the
sensesense
that that tion
tion of
of their
theireffects
effectsonon
the
the
firms'
firms'
existing
existing
IT gov-
IT gov-
they
they lead
leadtotoeither
eithercentralized
centralizedor decentralized
or decentralized ernance
ernance modes.
modes.Thus,
Thus,repeated
repeated
observations
observations
by by
modes
modes of ofITITgovernance.
governance. Yet,
Yet,
thethe
dynamics
dynamicsof of practicing
practicingIT
ITmanagers
managersthat
that
their
their
firms
firms
regularly
regularly
how
how these
thesecontingencies
contingenciesproduce
produce
alikealike
out-out- restructure
restructuretheir
theirITITgovernance
governance
arrangements
arrangements
would be a direct reflection of the need to
comes
comes arearedifferent.
different. With
With
reinforcing
reinforcing
contingen-
contingen-
cies,
cies, all
allof
ofthe
thesalient
salient
contingencies
contingencies
produce
produce
like like adapt to ever-evolving contingency forces.
dispositions
dispositionsononthe
the
mode
mode
of of
IT governance.
IT governance.
In In Since such periodic alterations of IT gover-
contrast,
contrast,with
withdominating
dominatingcontingencies,
contingencies,
the the nance arrangements can be both disruptive and
resulting
resultinggovernance
governance
arrangement
arrangementis the
is result
the result
of of expensive, are there other means of adaptation?
a single,
single,dominant
dominantcontingency.
contingency.
If the
If the
effects
effects
of of For example, Brown and Sambamurthy (1998)
other
other salient
salientcontingencies
contingenciesareare
in conflict
in conflict
withwith and Zmud (1988) have examined the role of
coordination mechanisms as elements of IT
the
the effects
effectsofofthe
the
dominating
dominating
contingency,
contingency,
as is as is
the case with Corporation E, then it would governance. How are coordination mechanisms
appear that these firms might incur some struc- utilized along with the locus of IT decision
tural maladpatation costs because they are being rights to respond to the pressures of evolving
forced to ignore other salient contingencies. We multiple contingencies? We think that this
would expect that the governance arrangements might be a fascinating and fertile ground for
future research.
induced by reinforcing contingencies would
inherently be more stable than those produced Another direction for future research would be to
by dominating contingencies. Obviously,
examine the influence of multiple contingencies
researchers might wish to examine some of the
on the design of customized governance arrange-
performance-related consequences of reinforcing ments within individual business units. Brown
versus dominating contingencies.
(1997) identifies a variety of business unit level
Conflicting contingencies appear to be a particu- factors as influencing distinct modes of gover-
larly interesting scenario for IS researchers since nance in individual business units. Brown and
they often lead to a federal mode of IT gover- Magill (1998) provide conceptual arguments
nance. During the 1990s, observations of prac- about impacts of enterprise and business unit
tice suggest that more firms are adopting the fed- level factors on modes of governance arrange-
eral governance form (LaPlante 1991; von ments within business units. Applying the multi-
pie
pie contingencies
contingenciesperspective,
perspective,
Brown
Brown
and and
Magill
Magill Similarly,
Similarly, senior
seniorexecutives
executivesmust
mustendeavor
endeavorto to
(1998)
(1998) offer
offerpropositions
propositionsabout
about
thethe
mode
mode
of gov-
of gov- accurately
accurately gauge
gaugethe
thesalience
salienceand
andthe
the
implica-
implica-
ernance for IT use within business units under tions
tions of
of individual
individualcontingency
contingencyforces.
forces.For
For
different scenarios of corporate and business-unitinstance,
instance, in
in Corporation
CorporationE,E,were
wereitit
notnot
for
forthethe
contingencies. Similar to our study, their proposi-
dominating
dominating effects
effectsof
ofgrowth
growththrough
through acquisition,
acquisition,
tions are directed at examining governancethe the firm
firm would
wouldhavehavemoved
movedinto intoa conflicting
a conflicting
arrangements in firms where the patterns arecontingencies
contingencies modemodeandandwould
wouldhavehave found
foundthethe
similar across business units. Collectively, ourfederal
federal governance
governancemode modetotobebemore
more appropri-
appropri-
study and the Brown (1997) and the Brownate. ate. What
What is
is the
thereal
realcost
costofofa amaladaptive
maladaptive gover-
gover-
and Magill (1998) studies could help nance
nance mode
mode brought
broughton onby
bya adominating
dominating contin-
contin-
researchers empirically examine the influence gency?
gency? What
What would
wouldbebethe
thereal
realbenefit
benefitofof
resolv-
resolv-
of multiple contingencies on differentiated gov-ing
ing the
the dominating
dominatingcontingency
contingencyand andmoving
moving to to
ernance arrangements across business units in the
the more
more appropriate
appropriategovernance
governancemode? mode? Senior
Senior
firms. executives
executives responsible
responsiblefor
fordesign
designofof
their
their
firm's
firm's
IT IT
governance
arrangementsmust
mustbebe
vigilant
vigilant
in in
Finally, our conceptualization of absorptive understanding
understandingthe
thesalience
salienceand
andimplications
implications
capacity forces focused on the IT-knowledge of (regarding
(regarding both
boththe
themode
modeofofITITgovernance
governanceand
and
line managers. We also recognize the impor- the
the costs
costs and
and benefits
benefitsattributed
attributedtoto
particular
particular
gov-
gov-
tance of the converse side of absorptive capaci- ernance
ernance modes)
modes)of
ofthe
themultiple
multiplecontingency
contingency
ty-the business knowledge of IT managers forces
forces impacting
impactingtheir
theirfirm
firmatatany
any
given
given
point
point
in in
(Boynton et al. 1992). We would expect low time.
time.
business knowledge of IT managers to influence
pressures toward decentralization. Future
research must examine how limited business Conclusion
understanding on the part of IT managers (either
In practice, it is frequently observed that organi-
at an enterprise or divisional level) influences the
zational IT governance moves through recurrent
nature of firms' IT governance structures.
centralization/decentralization cycles-both for
specific organizations and for all organizations in
Implications for Practice general. The theory of multiple contingencies,
along with our identification of three core
This study has provided robust descriptions of
domains of contingency forces, offers the
alternative modes of IT governance and the con-
promise of richer conceptual and practical
tingency conditions associated with their adop-
understandings of the complex dynamics under-
tion. Further, the study has offered detailed
lying the evolutionary nature of organizational IT
explanations as to why particular IT governance
governance arrangements.
modes may be appropriate when they may at
first glance appear to "contradict" acceptedOur long-standing research program on IT gover-
practice. All of a firm's senior executives (notnance has only served to reinforce our view of
just senior IT executives) must recognize the the critical importance of establishing appropri-
advantages of different IT governance modes ate IT decision rights to direct and coordinate an
and must understand the nature of the forces organization's effective use of and exploitation of
likely to become salient in directing the evolu- IT. IT governance arrangements create the con-
tion of their firm's IT governance arrangements. text, the rules, and the expectations that frame
In particular, as existing contingency forces and filter the IT-related decision making occur-
change in nature and salience, and as some of ring throughout all organizations. We anticipate
these forces become dominant in their effects, that our conceptual ideas and our interpretations
senior executives must realize the necessity of of the forces influencing IT governance arrange-
making adjustments in their IT governance ments at eight case sites will support, and ideal-
arrangements in order to maintain a dynamic ly stimulate, the work of other researchers inter-
alignment with their ever-changing organiza- ested in these and related organizational design
tional and environmental contexts. issues regarding the effective management of IT.
We
Wealsoalsoanticipate
anticipatethatthat
our discussions
our discussions
of both of both Blau,
Blau, P.
P. M.
M.and
andSchoenherr,
Schoenherr, R. R.
A. A.
TheThe
Structure
Structure
the
thelikely
likely effects
effects
of specific
of specific
contingency forces forcesof
contingency of Organizations,
Organizations,BasicBasic Books,
Books,
NewNewYork,
York,
on
on ITITgovernance
arrangements and theand the dynam-1971.
dynam-
ic
ic nature
nature ofofthethe
salient
salient
forcesforces
at playatat play
any point Boynton, A. C., Jacobs, G. C. and Zmud, R. W.
at any point
in
in time
timewillwillprove
proveusefuluseful
to those those executives "Information Technology Management: Just
to executives
faced
facedwithwith thethe
responsibility
responsibility of designing
of designing
(and (andWhose Responsibility Is It? Sloan Management
revising)
revising) their
their
firm's
firm's
modemode
of IT governance.
of IT governance. Review (4), 1992, pp. 32-38.
Boynton, A. C., Zmud, R. W., and Jacobs, G.
Finally,
Finally,regardless
regardless of whether
of whetherthe focustheisfocus is"The Influence of IT Management Practice on
research
research ororpractice,
practice,
IT governance
IT governance
issues invari-
issues invari- IT Use in Large Organizations," MIS Quarterly
ably
ablytake
take thethe
form
formof very
of very
messymessy
and ephemeral
and ephemeral(18:3), September 1994, pp. 299-318.
phenomena.
phenomena. WeWe are are
convinced
convinced
that efforts
that efforts
to to
Broadbent, M., and Weill, P. "Management by
understand
understand andand
direct
direct
thesethese
phenomena
phenomena
can not can notMaxim: How Business and IT Managers Can
follow a "business as usual" stance. We hope Create IT Infrastructures," Sloan Management
that our application of holistic theoretical and Review (38:3), Spring 1997, pp. 77-92.
methodological postures will encourage other Brown, C. V. "Examining the Emergence of
researchers-and practitioners-to likewise
Hybrid IS Governance Solutions: Evidence
innovate in their explorations of effective IT gov-
from a Single Case Site," Information Systems
ernance arrangements. Research (8:1), March 1997, pp. 69-94.
Brown, C. V., and Magill, S. L. "Alignment of the
IS Function With the Enterprise: Toward a
Acknowledgements
Model of Antecedents," MIS Quarterly (18:4),
Funding for this research was obtained from December
the 1994, pp. 371-403.
Financial Executives Research Foundation. Brown, C. V., and Magill, S. L. "Recon-
However, the views expressed here are oursceptualizing
and the Context-design Issue for the
not those of the research sponsors. Prior versions
Information Systems Function, Organization
of the manuscript have benefited from presenta-Science (9:2), March-April 1998, pp. 176-194.
tions at the Florida State University, University
Brown,of C. V., and Sambamurthy, V. "Linking
California, Los Angeles, and the University of
Intra-Organizational Stakeholders: CIO
Pittsburgh. We would like to thank the seniorPerspectives
edi- on the Use of Coordination
tor, the associate editor, and the reviewersMechanisms,"
for Florida State University,
their comments on earlier versions of the manu- Tallahassee, FL, 1998.
script. Clark, T. D. "Corporate Systems Management:
An Overview and Research Perspective,"
Communications of the ACM (35:2), February
References
1992, pp. 61-75.
Ahituv, N., Neumann, S., and Zviran, M. "Factors Clark, C. E., Cavanaugh, N.C., Brown, C. V., and
Affecting the Policy for Distributing Sambamurthy, V. "Building Change-Readiness
Computing Resources," MIS Quarterly (13:4), Capabilities in the IS Organization: Insights
December 1989, pp. 389-401. From the Bell Atlantic Experience," MIS
Applegate, L. M, McFarlan, F. W., and Quarterly (21:4), December 1997, pp.
McKenney, J. L. Corporate Information Systems425-455.
Management: Text and Cases, 4th ed., RichardCohen, W. M., and Levinthal, D. A. L.
D. Irwin, Chicago, 1996. "Absorptive Capacity: A New Perspective on
Bastien, D. T. "Common Patterns of Behavior and Learning and Innovation," Administrative
Communication in Corporate Mergers and Science Quarterly (35), 1990, pp. 128-152.
Acquisitions," Human Resource Management Cooprider, J. G. and Victor, K. M. "The
(26:1), 1987, pp. 17-33. Contribution of Shared Knowledge to IS
Blau, P. M. "A Formal Theory of DifferentiationGroup Performance," in Proceedings of the
in Organizations," American Sociological Fourteenth International Conference on Infor-
Review (35), 1970, pp. 201-218. mation Systems, J. I. DeGross, R. P. Bostrom,
and
andD.D.
Robey
Robey
(eds.),(eds.),
Orlando,
Orlando, FL, 1993,Henderson,
FL, 1993, pp. Henderson,
pp. J. J.
C., C.,
"Plugging
"Plugging
Into Strategic
Into Strategic
285-300. Partnerships:
Partnerships: TheThe
Critical
Critical
IS Connection,"
IS Connection,
Cross, J., Earl, M., and Sampler, J. Sloan
SloanManagement
ManagementReview,
Review,
Spring,
Spring,
1990, pp.
1990, pp
"Transformation of the IT Function at British 7-18.
Petroleum," MIS Quarterly (21:4), December Henderson, J., and Venkatraman, N. "Strategic
1997, pp. 401-424. Alignment: A Framework for Strategic
Curtis, B., Krasner, H., and Iscoe, N. "A Field Information Technology Management," in
Study of the Software Design Process for Large Transforming Organizations, T. Kochan and M.
Systems," Communications of the ACM Useem (eds.), Oxford Press, New York, NY,
(31:11), 1988, pp. 1268-1287. 1992, pp. 97-117.
Dixon, P. J., and John, D. A. "Technology Issues Hoskisson, R., and Hitt, M. "Strategic Control
Facing Corporate Management in the 1990s," and Relative R&D Investment in Large
MIS Quarterly (13:3), September 1989, pp. Multiproduct Firms," Strategic Management
247-255. Journal (9:6), 1988, pp. 605-621.
Ein-Dor, P. and Segev, A. "Organizational Jemison, D. B., and Sitkin, S. B. "Corporate
Context and MIS Structure," MIS Quarterly (6), Acquisitions: A Process Perspective," Aca-
1982, pp. 55-68. demy of Management Review (11:1), 1986,
Eisenhardt, K. "Building Theories from Case pp. 145-163.
Jones, G. R., and Hill, C. W. L. "Transaction Cost
Study Research," Academy of Management
Review (14), 1989, pp. 532-550. Analysis of Strategy-Structure Choice,"
Strategic Management Journal (9:2), 1988, pp.
Fisher, J. "Technological Interdependence, Labor
159-172.
Production-functions and Control Systems,"
Kazanjian, R. K., and Drazin, R. "Implementin
Accounting Organizations and Society (19:6),
Internal Diversification: Contingency Factor
1994, pp. 493-505.
for Organization Design Choices, Academy o
Galbraith, J. R. Designing Complex Organiza-
Management Review (12:2), 1987, pp.
tions, Addison-Wesley, Reading, MA, 1967. 342-354.
Ghoshal, S. "Global Strategy: An Organizing
Keen, P. G. W., Shaping the Future, Harvard
Framework," Strategic Management Journal
Business School Press, Boston, 1991.
(8:5), 1987, pp. 425-440.
King, J. L "Centralized Versus Decentralized
Ginsberg, A. "Connecting Diversification to
Computing: Organizational Consideration
Performance: A Sociocognitive Approach,"
and Management Options" Computing
Academy of Management Review (15:3), Surveys (15:4), December 1983, pp. 320-349.
1990, pp. 514-535. LaBelle, A., and Nyce, H. E. "Whither the IT
Gresov, C. "Exploring Fit and Misfit with Multiple Organization?" Sloan Management Review
Contingencies," Administrative Sciences (28), 1987, pp. 75-85.
Quarterly (34:3), September 1989, pp. LaPlante, A. "Here Come the Hybrids,"
431-453.
Computerworld, June 1991, pp. 57-61.
Grinyer, P. H., Yasai-ardekani, M., and Al-Bazaz,
Lawrence, P. R., and Lorsch, J. W. Organization
S. "Strategy, Structure, the Environment,and
andEnvironment, Richard D. Irwin:
Financial Performance in 48 UK Companies," Homewood, IL, 1969.
Academy of Management Journal (23),Leifer, 1980, R. "Matching Computer-based Information
pp. 193-220. Systems with Organizational Structures," MIS
Hamel, G., and Prahalad, C.K. Competing for the
Quarterly (12:1), March 1988, pp. 63-74.
Future: Breakthrough Strategies for Lubatkin,Seizing M. "Mergers and the Performance o
Control of Your Industry and Creatingthe theAcquiring Firm," Academy of Management
Markets of Tomorrow, Harvard Business Review (8), 1983, pp. 218-225.
School Press, Boston, 1994. Maglitta, J., and Mehler, M. "The New
Hannan, M. T., and Freeman, J. "Structural Centralization," Computerworld, April 27,
Inertia and Organizational Change," American 1992, pp. 85-88.
Sociological Review (49), April 1984, pp. Napier, N. K. "Mergers and Acquisitions, Human
149-164. Resource Issues, and Outcomes: A Review and
Suggested
Suggested Typology,"
Typology,"
Journal ofJournal
Management and B.
of Management B. Gallupe
Gallupe (eds.),
(eds.), Lake
Lake Louise,
Louise,Albert
Albert
Studies
Studies (26:3),
(26:3),
1989, 1989,
pp. 271-289.
pp. 271-289. 1993, pp.pp. 194-204.
194-204.
Nelson,
Nelson, R. R.,
R. and
R.,Winter,
and Winter,
S. G., An Evolutionary Sambamurthy, V., and Zmud, R. W. IT
S. G., An Evolutionary
Theory
Theory of Economic
of Economic
Change, Change,
Belknap Press,
Belknap Press,
Management Competency Assessment: A Tool
Cambridge,
Cambridge, MA, MA,
1982. 1982. for Creating Business Value Through IT,
Olson,
Olson, M.,M.,
and Chervany,
and Chervany,
N. "The Relationship Financial Executives Research Foundation,
N. "The Relationship
Between
Between Organizational
Organizational
Characteristics
Characteristics
and and
Morristown, NJ, 1994.
the Structure of the Information Service Simmonds, P. "The Combined Diversification
Function," MIS Quarterly(4), 1980, pp. 57-68. Breadth and Mode Dimensions and the
Panzar, J. C. and Willig, R. D. "Economies ofPerformance of Large Diversified Firms
Scope," American Economic Review (71), Strategic Management Journal, (11), 1990, p
1981, pp. 268-272. 399-410.
Pitts, R. A. "Diversification Strategies and Slevin, D. P., and Pinto, J. K. "Balancing Strategy
Organizational Policies of Large Diversified and Tactics in Project Implementation," Sloan
Firms," Journal of Economics and Business Management Review, (29:1), Fall 1987, pp.
(28:3), 1976, pp. 181-188. 33-41.
Rathnam, S., Mahajan, V., and Whinston, A. B. Tavakolian, H. "Linking the Information
"Facilitating Coordination in Customer Technology Structure with Organizational
Support Teams: A Framework and Its Competitive Strategy: A Survey," MIS Quarterly
Implications for the Design of Information (13:3), September 1989, pp. 309-317.
Technology," Management Science (41:12), Teece, D. J. "Economies of Scope and the Scope
1995, pp. 1900-1921. of the Enterprise," Journal of Economic
Reed, R. and Luffman, G.A. "Diversification: The Behavior and Organization (1), 1980, pp.
growing confusion," Strategic Management 223-247.
Journal (7), 1986, pp. 29-35. Thompson, J. D. Organizations in Action,

Rockart, J. F. "The Line Takes the Leadership-IS McGraw-Hill, New York, 1967.
Management in a Wired Society," Sloan Venkatraman, N. "IT-enabled Business
Management Review, Summer 1988, pp. Transformation: From Automation to Business
57-64. Scope Redefinition," Sloan Management
Review (35:2), Winter 1994, pp. 73-87.
Rockart, J. F., Earl, M. J., and Ross, J. W. "Eight
Imperatives for the New IT Organization," von Simson, E. "The 'Centrally' Decentralized IS
Sloan Management Review, Fall 1996, pp. Organization," Harvard Business Review, July-
43-56. August 1990, pp. 158-162.
Rockness, H., and Zmud, R.W. Information Walz, D., Elam, J. J., and Curtis, B. "Inside a
Technology Management: Evolving Manager- Software Design Team: Knowledge Acquisi-
ial Roles, Financial Executives Research tion, Sharing, and Integration," Communica-
Foundation, Morristown, NJ, 1989. tions of the ACM (36:10), 1993, pp. 63-76.
Ross, J. W., Beath, C. M., and Goodhue, D. L. Weill, P., and Broadbent, M. Leveraging the New
"Develop Long-term Competitiveness Through Infrastructure, Harvard Business School Press,
IT Assets," Sloan Management Review, Fall Boston, MA, 1998.
1996, pp. 31-42. Wilder, C. "Bend Me, Shape Me: What's the Best
Sabherwal, R., and Kirs, P. "The Alignment Shape for IS in the 1990s?" Computerworld,
Between Organizational Csfs and Information December 24, 1990, p. 14.
Technology Capability in Academic Institu- Williamson, O. E. Markets and Hierarchies:
tions," Decision Science (25:2), 1994, pp. Analysis and Antitrust Implications, The Free
301-330. Press, New York, 1975.
Sambamurthy, V., and Zmud, R.W. "A Winter, S. G., "Knowledge and Competence as
Conceptual and Measurement Approach for Strategic Assets," in The Competitive
Examining IT Management Architectures," in Challenge: Strategies for Industrial Innovation
Proceedings of the Annual Conference of the and Renewal, D. J. Teece (ed.), Ballinger
Administrative Sciences Association of Publishing Company, Cambridge, MA, 1987,
Canada, Information Systems Division, P. 159-184.
Todd
Zerfane,
Zerfane,R.
R.M.,
M.,and
andGul,
Gul,F. F.
A. A.
"The
"The
Effects
Effects
of Task
of Task dispersed
dispersed IS IS
workforce.
workforce.In addition,
In addition,
he has also
he has als
Characteristics and Sub-unit Structure on examined
examined issues
issues
related
related
to thetoCIO
the
andCIO
top and
man-top man
Dimensions of Information Processing," agement
agement teams'
teams'roles
roles
in nurturing
in nurturing
organizational
organizationa
Information Process Management (29:6), success
successwith
withinformation
information technology
technology
use. Mostuse.
of Most o
1993, pp. 703-719. his work has been conducted in Fortune 500
Zmud, R. W. "Building Relationships Throughout firms and has been funded by the Financial
the Corporate Entity," in Transforming the IS Executives Research Foundation and the
Organization: The Mission, The Framework, Advanced Practices Council (APC) of the Society
the Transition, J. Elam, M. Ginzberg, P. Keen, for Information Management International (SIM)
and R. W. Zmud (eds.), ICIT Press, His previous work has been published in journals
Washington, DC, 1988. such as MIS Quarterly, Decision Sciences, and
Information Systems Research. Currently, he is a
associate editor on the board of MIS Quarterl
About the Authors
and a departmental editor for The Database fo
V. Sambamurthy (Ph.D., University of Advances in Information Systems. He has been
Minnesota, 1989) is associate professor at the active in managing and participating in the con
Robert H. Smith School of Business, University of ference programs for Decision Sciences, ICIS,
Maryland, College Park. His current research and the OCIS division of the Academy of
interests are focused on capabilities and organi- Management.
zation designs for the extraction of information
technology (IT) value. Some of the issues he has Robert W. Zmud is professor and Michael F
recently examined include: (1) the process by Price Chair in MIS in the Michael F. Price College
which contemporary firms envision, mobilize of Business at the University of Oklahoma. Hi
attention toward, and justify organizational fund- current research interests focus on the impact o
ing for visionary IT applications, (2) the IT man- information technology in facilitating a variety o
agement capabilities and competencies for suc- organizational behaviors and on organizationa
cess with business use of IT, and (3) the design of efforts involved with planning, managing, and
IS organizations to build value-added relation- diffusing information technology. Both his Ph.D
ships with senior management and business (University of Arizona) and his S.M. (MIT)
management and lateral relationships among the degrees are in management.
APPENDIX A
Details of the Field Survey

The field survey sample was drawn from 140 firms that were members of the Financia
Institute (FEI), the professional society for chief financial officers (CFO), whose research
funded the original study (Rockness and Zmud 1989). While manufacturers represented a
firms participating in the study, a few firms representing the insurance, financial services,
munications industries were also included. At each sampled firm, the CFO was asked to id
business units on the basis of the business units' overall performance-a high-performing un
age performing unit, and a low-performing unit. Five questionnaires were then asked to be
and returned to the researchers: one to be completed by the CFO, one to be completed
information officer (CIO), and one to be completed by the senior executive of each of th
ness units. Usable responses were obtained from 35 firms (response rate of 25%), yielding a
business units. This response rate is consistent with that observed in other field surveys aim
standing phenomena at both the firm and business unit levels (Gaedeke and Tootelian 197
to verify
verify whether
whetherthe theresponse
responserate
rate
produced
produced
a biased
a biased
set of
setresponses,
of responses,
secondary
secondary
data on
data
netonrevenue,
net revenue,
net
net income,
income,and andassets
assetswere
were
gathered
gatheredfrom
fromthethe
COMPUSTAT
COMPUSTAT database
database
for both
for both
responding
responding
and nonre-
and nonre-
sponding
sponding firms.
firms.Analysis
Analysisrevealed
revealed
nonosignificant
significant
differences
differences
between
between
the two
the subsamples
two subsamples
(p > .10),
(p >par-
.10), par-
tially
tially alleviating
alleviatingconcerns
concernsabout
about
thethe
problem
problem
of nonresponse
of nonresponsebias.bias.
Published
Published material
materialininthe
thetrade
trade
press
press
andand
prior
prior
research
research
werewere
utilized
utilized
for developing
for developing
items2items2
to assess
to locus
assess locus
and
and sharing
sharingon
onseven
sevenclasses
classes
ofofIT IT
decisions:
decisions:
1. long-term
long-termhardware
hardwaredecisions
decisions
about
about
mainframes,
mainframes,
minicomputers,
minicomputers,
and networks,
and networks,
including
including
capacities,
capacities,
specific
specific vendors,
vendors,brands,
brands,and
andtechnologies;
technologies;
2. long-term
long-termmicrocomputer
microcomputerdecisions,
decisions,
including
including
capacities,
capacities,
specific
specific
vendors,
vendors,
brands,
brands,
and technologies;
and technologies;
3. long-term
long-termapplications
applicationsdecisions
decisions
about
about
IT-enabled
IT-enabled
value
value
chain
chain
activities
activities
(logistics,
(logistics,
manufacturing,
manufacturing,
etc.);
4. annual hardware decisions about enhancements and upgrades;
5. annual applications decisions about enhancements or additions to existing applications supporting

various value chain activities;
6. project management decisions; and
7. daily operations decisions, including service delivery priorities and overriding standard procedures
when special IS service needs arose.
Senior business unit (non-IS) executives served as respondents in indicating the extent to which corporate
IS, divisional IS, and line management assumed and shared responsibility for each decision.3 The
response cues ranged from "unilaterally makes decisions" through "shares equally in decisions" to "no
role in the decision," arranged on a five-point Likert scale. Factor analyses were used to confirm that the items
tapping each one of these dimensions of IT decisions loaded together onto their corresponding factors.
Next, rule sets were devised to operate on the data collected through the field survey and compute mea-
sures of IT decision-making locus.4 The locus of IT decision making was operationalized as follows:
* If the decision scores indicated that corporate IS played a primary or unilateral role in a particular deci-
sion, and that divisional IS and line management had a minor or no role in that decision, then corpo-
rate IS was identified as the locus of that decision (low decentralization, score = 1). Similar logic was
applied to identify cases where divisional IS (moderate decentralization, score = 2) or business man-
agement (high decentralization, score = 3) were viewed to be the locus of IT decision making. Such a
rule set reflects a logic that business management as the locus of IT decision making represents the
highest level of decentralization, a logic consistent with the literature (Rockart 1988).
* If two stakeholders (invariably either corporate IS and divisional IS or divisional IS and line manage-
ment) were reported to serve essentially equivalent roles, a mid-point score (1.5 or 2.5, respectively)
was assigned.
2A copy of the questionnaire can be obtained upon request from the first author.
3Responses on the locus of IT decisions were also obtained from the ClOs. A comparison of the responses obtained
from ClOs and senior business unit executives revealed consistency in structure and pattern. However, as the CIO
responses indicated a potentially self-serving bias toward centralization, the decision was made to rely on the senior
business unit executive's responses.
41n order to assess the reliabilty of the coding scheme for locus, two raters (blind to this study's objectives) indepen-
dently developed codes for each of the seven types of ITM decisions. Inter-rater reliability was assessed using the
Cohen's Kappa and the values ranged from .70 to .83; all of these values are significant at p < .05.
- zl6][NU :i.uih~Ei iI~(. E~ 11k~1IE] E1U IL4f :Z~i .i.
Architecture
Dimension Items Cronbach's Alpha
LOCUS
Infrastructure * long-range hardware decisions .77

* annual hardware decisions
IT Use * long-range decisions about micros .73

* long-range applications decisions
* annual applications decisions
* decisions about daily IS operations
Project * project management decisions
Implementation
Once
Once locus
locus
scoresscores
on the seven
on ITthedecisions
seven were
ITdeveloped
decisionsfor each
were of the
developed
73 business units,
for factor
each of the 73 bus
analysis
analysis was conducted
was conducted
to produce three
to produce
factors, as three
shown infactors,
Table Al: locus
as ofshown
infrastructure
in Table
decisions,
Al: locus of infras
IT
ITuse
use
decisions,
decisions,
and project
and implementation
project implementation
decisions. On the basisdecisions.
of this factorOn analysis,
theaggregate
basis of this factor a
scores
scores for for
the locus
theoflocus
these three
of these
decisionsthree
were developed
decisions
for each
werecase site.
developed for each case site.
APPENDIX
APPENDIX B B
Interview Protocol for Case Site Interviews
The following structured protocol was developed and used to guide the interviews:
1. Describe the distribution of IT management responsibilities between corporate IS and divisional IS

groups in applying IT in your firm. What factors explain this pattern of distribution?
2. How does such a distribution compare with the overall corporate centralization/decentralization
policies?
3. Describe the distribution of IT management responsibilities between divisional IS personnel and line
managers in applying IT. What factors explain this pattern of distribution?
4. Have these patterns of responsibilities changed over the last three to five years? If so, how and why?
5. Do these patterns of responsibilities differ across business units? If so, how do they differ and why to
they differ.
6. What are the critical factors that need to be in place for business units to apply IT effectively? Why?
7. What factors most limit a business unit's capability to apply IT effectively? Why?
8. Are these factors (both positive and negative) influenced by the distribution of IT management respon-
sibilities?
The interviews were conducted with the firm's CIO and senior IS executives from the three business units
identified by the CFO during the field survey. At certain sites, the firm's CFO was also interviewed. Extensive
notes were taken during the interview, with the interview guide designed to accommodate note taking. At
the end of interviews at each case site, the researchers generated descriptive summaries, which were then
returned to the interviewees for their comments. Their comments were used to amend the summaries.
APPENDIX C
Details of the Data Analysis Approach

Given the qualitative nature of the data analysis across the eight case sites, we deemed it important to
adopt an analysis strategy that would protect the results against potential biases due to the researcher's
knowledge of the hypotheses. Therefore, one author undertook the analysis of the contingency forces,
whereas the other author undertook analysis of governance modes at the eight case sites.
The study's first author read the transcripts of the individual case studies in order to develop profiles of
the different contingency forces in play at each case site. Next, the author reread the transcripts to gauge
the salience of the different contingency forces. The salience of individual contingency forces was
assessed by both the tone and repetition evidenced by respondents in their identification of specific con-
tingency forces and their discussion of the effects nd th sinese forces. For example, all respondents at
Corporation E talked about the history of their firm's growth through acquisition. They emphasized that
this diversification mode was a key characteristic of their business and that ignoring the influence of this
factor in designing their IT governance arrangements would be both impossible and prohibitively costly.
Thus, this factor was judged to be highly salient. As another example, respondents at Corporation C
repeatedly mentioned that their line managers lacked the IT knowledge to take on IT decision responsi-
bilities, even though other forces suggested that a more decentralized governance mode would be desir-
able. Thus, low line IT knowledge was judged to be a highly salient contingency factor in Corporation C.
As a final example, respondents at Corporation A indicated the existence of a desire to build partnerships
between the line and IS managers. However, during the discussions, respondents also pointed out that
this desire was not as strong an influence as the desire to leverage economies of scope through the con-
solidation of IT assets. In fact, respondents indicated that significant concerns existed to obtain greater
efficiencies through the enterprise-wide use of IT assets. Based on these conversations, we inferred that
the desire to consolidate IT assets was a salient contingency force at Corporation A, whereas the desire
to build partnerships was a less salient contingency force. Readers will recall that the transcripts serving
as the basis for such judgements regarding salience had been read and validated by the respondents.
After identifying and assessing the salience of the contingency forces at play at each firm, the study's first
author then diagnosed whether each force produced a disposition toward centralization or decentraliza-
tion. Again, this analysis was based on the transcripts of the interviews. The analysis described so far was
depicted through contingency-state diagrams for each case site. These diagrams reflected the following
elements: (1) contingency forces producing a disposition toward centralization and their relative salience,
(2) contingency forces producing a disposition toward decentralization and their relative salience, and (3)
the strength of these overall dispositions toward centralization and decentralization. Figure C1 illustrates
the contingency-state diagrams developed for one site, Corporation A; a full set of these diagrams can be
obtained on request from the authors.
As a next step, the first author examined the different contingencies and classified the case sites into rein-
forcing, conflicting, and dominating contingencies. Reinforcing contingencies were defined as organiza-
tions where all the salient contingencies produced similar dispositions toward centralization or decen-
tralization. Two case sites were categorized as exhibiting reinforcing contingencies (sites B and F).
Conflicting contingencies were defined as organizations where the salient contingencies produced con-
tradictory dispositions toward centralization or decentralization. Four case sites were found to exhibit
conflicting contingencies (sites A, C, G, and H). Finally, dominating contingencies were defined as orga-
nizations where a single contingency appeared to influence the mode of IT governance. Two firms exhib-
ited this pattern of contingencies (sites D and E). Thus, this process resulted in an exhaustive classifica-
tion of the eight case sites into reinforcing, conflicting, or dominating contingencies.
Independently, the second author, who had participated in the original case data gathering, produced
assessments of the modes of IT governance at each case site. This author examined both numerical scores
from the questionnaire survey and transcripts of the cases to classify the case sites as to the modes of IT
governance depicted in Table 4.
In summary, the two authors independently produced the profile of multiple contingencies and the mode
of IT governance for each case site. These, then, became the data examined in assessing the study's
hypotheses. Such an analysis approach provides assurance that the authors did not expose the study's
results to their subjective research biases.
Enhance line/IS partnerships

Internal
Internal growth
growth
modemode
IT governance
Infrastructure: Corporate IS
IT use: Divisional IS
+ Disposition toward
centralization
Project management:
r, Corporate IS
Low market
Decentralized
relatedness
corporate Improved efficiency throug
governance IT asset consolidation

Arrangements For Information Technology Governance

Uploaded by

Copyright:

Available Formats

Arrangements For Information Technology Governance

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Arrangements For Information Technology Governance

Uploaded by

Copyright:

Available Formats

Arrangements for Information Technology Governance: A Theory of Multiple

Management Information Systems Research Center, University of Minnesota is

ARRANGEMENTS FOR INFORMATION

Today, information technologies (IT) present

MIS Quarterly Vol. 23 No. 2, pp. 261-290/June 1999 261

262 MIS Quarterly Vol. 23 No. 2/June 1999

Spheres of IT Activities Patterns in the Locus of Authority

Project management Corporate IS

Spheres of IT Activities Patterns in the Locus of Authority

IT infrastructure Divisional IS Divisional IS Divisional IS Divisional

IT infrastructure Corporate IS Corporate IS Corpor

IT use Corporate IS Divisional IS or line management Divisional IS

264 MIS Quarterly Vol. 23 No. 2/June 1999

Corporate Overall Centralized Corporate IS Applegate et al. 1996; Brown and M

Firm size Small Corporate IS Ein-Dor and Segev 1982

Acquisition growth Division

Diversification Related markets Corporate IS Boynton et al. 1992; Brown and

Unrelated markets Division

Exploitation Enterprise-wide Corporate IS Henderson 1990; Keen 1991; L

Absorptive Line IT Low Corporate IS Boynton et al. 1992, 1994; Brown

266 MIS Quarterly Vol. 23 No. 2/June 1999

MIS Quarterly Vol. 23 No. 2/June 1999 267

268 MIS Quarterly Vol. 23 No. 2/June 1999

MIS Quarterly Vol. 23 No. 2/June 1999 269

Site Decentralization of IT Decisions" Firm Size Description

B High on all IT decisions Large Industrial chemicals firm

project management markets

D High on all IT decisions Large Multinational manufacturer of

H Low on infrastructure, Small Specialty manufacturer in the

270 MIS Quarterly Vol. 23 No. 2/June 1999

IZ . -y n SI bE~ I ECIOIYL I aId mie4C.pIt m.[ uL lIZ 114M qICTT-6IM t(i

Locus for Locus for Project

Corporation A Corporate IS Divisional IS Corporate IS Federal

Corporation B Divisional IS Divisional IS Line management Decentralized

Corporation C Corporate IS Divisional IS Divisional IS Federal

Corporation D Divisional IS Divisional IS Divisional IS Decentralized

Corporation E Divisional IS Divisional IS Divisional IS Decentralized

Corporation F Corporate IS Corporate IS Corporate IS Centralized

Corporation G Corporate IS Line management Line management Federal

Corporation H Corporate IS Line management Line management Federal

MIS Quarterly Vol. 23 No. 2/June 1999 271

_ _ m _ __~1 = i V-1 i SO I tsi 1o IMEE _ _ I I a

Corporate Corporate Decentralized Decentralized Centralized Centralized

Economies of Diversification Acquisition Decentralized

Diversification Low market Decentralized High market Centralized

Exploitation Partnerships Decentralized Partnerships Decentralized

Absorptive Line IT knowledge High Decentralized

Mode of IT Governance Decentralized Centralized

272 MIS Quarterly Vol. 23 No. 2/June 1999

MIS Quarterly Vol. 23 No. 2/June 1999 273

Corporate Governance Decentralized Decentralized |Decentralized Decentralized Decentralized Cent

Geographic High Decentralized

Economies Diversification Internal Centralized Acquisition Decentralized Internal Central

Diversification Low market Decentralized High market Cent

Exploitation IT asset Centralized

Partnerships Decentralized Partnerships Decentra

Absorptive Line IT Low Centralized High Decentralized

Mode of IT governance Federal FeeralFederal Federal