1
Control Balancing in Information Systems Development
Offshoring Projects
Robert Wayne Gregory
Roman Beck
University of Göttingen
Goethe University Frankfurt
Platz der Göttinger Sieben 5
Grüneburgplatz 1
37073 Göttingen, Germany
60323 Frankfurt, Germany
gregory@wiwi.uni-goettingen.de
rbeck@wiwi.uni-frankfurt.de
Mark Keil
Georgia State University
Robinson College of Business
35 Broad Street
Atlanta, GA 30303, USA
mkeil@gsu.edu
WORKING PAPER VERSION, ACCEPTED FOR
FORTHCOMING PUBLICATION IN
MANAGEMENT INFORMATION SYSTEMS QUARTERLY
2
Control Balancing in Information Systems Development
Offshoring Projects1
Abstract: While much is known about selecting different types of control that can be exercised in
systems development (ISD) projects, the control dynamics associated with ISD offshoring
projects represents an important gap in our understanding. In this paper, we develop a
substantive grounded theory of control balancing that addresses this theoretical gap. Based on a
longitudinal case study of an ISD offshoring project in the financial services industry, we
introduce a three-dimensional control configuration category that emerged from our data,
suggesting that control types is only one dimension in which control configuration decisions need
to be made. The other two dimensions that we identified are control degree (tight vs. relaxed) and
control style (unilateral vs. bilateral). Furthermore, we illustrate that control execution during
the lifecycle of an ISD offshoring project is highly intertwined with the development of clientvendor shared understanding and that each influences the other. Based on these findings, we
develop an integrative process model that explains how offshoring project managers make
adjustments to the control configuration periodically to allow the ISD offshoring project and
relationship to progress, yielding the iterative use of different three-dimensional control
configurations that we conceptualize in the paper. Our process model of control balancing may
trigger new ways of looking at control phenomena in temporary inter-firm organizations such as
client-vendor ISD offshoring projects. Implications for research on organizational control and
1
An earlier version of this paper appeared in the Proceedings of the International Conference of Information Systems
(ICIS). The full citation is Beck, R.; Gregory R.W.; Prifling, M. (2008) Cultural Intelligence and Project
Management Interplay in IT Offshore Outsourcing Projects. ICIS; Paris, France. The ‘cultural intelligence’ aspect of
the ICIS paper was further developed and appeared in Information Technology & People (IT&P). The full citation is
Gregory, R.W.; Prifling, M.; Beck, R. (2009) The Role of Cultural Intelligence for the Emergence of Negotiated
Culture in IT Offshore Outsourcing Projects. IT&P, Vol. 22, No. 3, pp. 223-241.
3
ISD offshoring are discussed. In addition, guidelines for ISD offshoring practitioners are
presented.
Keywords: Control Balancing, Control Dynamics, Organizational Control, Information Systems
Development, Offshoring Projects, Outsourcing Relationships, Longitudinal Case Study,
Grounded Theory, Process Model, Project Management
4
Control Balancing in Information Systems Development
Offshoring Projects
INTRODUCTION
Information systems development (ISD) offshoring continues to grow in importance as a means
to exploit differences (e.g., labor costs, time zone, available expertise) across countries
(Ghemawat 2007; Sarker and Sarker 2009). Offshoring is a subset of the broader global
information systems (IS) outsourcing industry, which was estimated by Gartner in 2012 to be
worth $309 billion. The challenges associated with ISD outsourcing are well documented, and
statistics suggest that only 63 percent of IS outsourcing projects are successful (Lacity and
Willcocks 2012). Critical for achieving IS outsourcing success is effectively controlling projects
and relationships, which is extremely difficult in this context (Tiwana and Keil 2009) due to
outsourcing-specific challenges, such as developing client-vendor shared understanding (Rai et
al. 2009). These challenges are arguably even greater when the outsourcing involves an offshore
vendor. Thus, more research is needed on how to exercise control effectively in such projects.
Prior IS research on control has largely focused on why and under which circumstances specific
controls (e.g., monitoring an ISD task), control modes (e.g., behavior control), or control types
(e.g., formal control) are selected and used (e.g., Kirsch 1996), how controls are combined into a
portfolio (e.g., Kirsch 1997), how different control types interact at the portfolio level (e.g.,
Maruping et al. 2009; Tiwana 2010), and how the use of controls impacts performance (e.g.,
Henderson and Lee 1992; Tiwana 2008). Comparatively less research attention has been placed
on the dynamics of control, although there has been some work on how portfolios of control
change over time across different phases of an IS project (e.g., Kirsch 2004).
5
In general, prior research on control can be characterized as having a one-dimensional focus on
the construction of control portfolios and a unidirectional perspective on the relationship between
project- or relationship-related events and the triggering of changes in the control portfolio.
Indeed, in one of the few studies to examine control dynamics in outsourced software
development projects, Choudhury and Sabherwal (2003) focused on how certain events trigger
changes in the portfolio of controls and not on the dimensions of control use or the interrelationship between control behavior and the client-vendor relationship. Accordingly, there is a
theoretical gap regarding the dimensions of control that must be managed over time and the
reciprocal nature of the relationship between control use and the client-vendor relationship.
To address this theoretical gap, we studied control behavior in a large offshore outsourced ISD
project for several years. Our data directed us toward the notion of ‘control balancing,’ which we
define as making adjustments to the control configuration periodically in terms of control types,
control degree, and control style, to allow the ISD offshoring project and relationship to progress.
The image of balancing is used as a metaphor for finding the appropriate configurations along
these three control dimensions that will allow the ISD offshoring project and relationship to
progress without toppling over. We identified three control configurations from our data: (1)
authoritative control, which is based on a traditional outsourcing perspective with clear
boundaries between ‘client’ and ‘vendor,’ (2) coordinated control, in which client and vendor
work toward shared goals, and (3) trust-based control, grounded in shared understanding. Just as
a gymnast on a balance beam must recurrently shift her weight in order to keep from falling while
executing her performance, a project manager must make adjustments to the control
configuration periodically and shift between these three control configurations in order to
maintain the delicate balance that allows the project and relationship to move forward. Based on
the dynamics observed in our case study, we developed a process model of control balancing that
6
explains the dimensions of control use and the dynamic interaction between control
configurations and the ISD offshoring relationship.
We construct a theory of control balancing in ISD offshoring that explains what needs to be
balanced (i.e., what factors should be considered for explaining control balancing?), how control
balancing occurs (i.e., how are the factors related and what are the dynamics over time?), and
why control balancing matters (i.e., why is this an important phenomena and what are the
implications?). We begin with the theoretical background in which we present what is known
about control portfolios, their evolution, and control balancing in organizational and IS
management. In the methods section that follows, we describe our research design. Following
this, we present our case analysis and the theory that emerged therefrom. We conclude with a
discussion of findings in the light of existing literature, present practical and theoretical
implications, and provide directions for future research.
THEORETICAL BACKGROUND
Control in IS Outsourcing and Offshoring
In order for organizations, including temporary organizations such as projects, to achieve their
objectives, some type of strategy must be implemented that “effectively controls members’
activities in a manner functional for the organization” (Barker 1993, p. 409). Ouchi (1977; 1979;
1975) proposed one of the most widely adopted control frameworks, which focuses on formal
behavioral and outcome controls as well as informal clan controls. An extension of this
framework to the context of IS projects was proposed by Kirsch (1996; 1997), who built upon
this work as well as related studies (e.g., Henderson and Lee 1992). More recently, controlling IS
outsourcing and offshoring projects has become an important area of inquiry for IS researchers.
Table 1 summarizes this stream of research.
7
Study
Context
Focus
Findings
(Choudhury
and
Sabherwal
2003)
Outsourced ISD
projects
Evolution of portfolios of
formal and informal controls
and factors influencing the
choice and evolution of
controls
At the start of the project, control portfolios are
dominated by formal outcome controls, influenced by
the client’s perception of the controllee’s knowledge,
the consequent role expectations, and perceptions of
difficulty in monitoring vendor behavior. Later in the
project, formal behavior controls and informal controls
are added to the portfolio, influenced by vendor
performance.
(Gopal and
Gosain
2010)
Outsourced ISD
projects
The role of boundary spanning
between the vendor and client
for the performance effects of
formal and informal controls
Both formal and informal controls have positive
performance impacts. In addition, boundary spanning
moderates the impacts of outcome controls on
software quality and project efficiency and the impacts
of behavior controls on project efficiency, such that as
the level of boundary spanning increases, the
relationship between outcome/behavior controls and
the respective outcomes is strengthened.
(Heiskanen
et al. 2008)
Public-sector
outsourced ISD
projects
Co-evolution of control and
trust during relationship
development
The public sector client’s actions oscillate between
trust and control during the course of relationship
development in three areas: performance, price level,
and observed behavior.
(Rustagi et
al. 2008)
Outsourced IS
projects (i.e.,
applications,
infrastructure, or
business processes)
Amount of formal control
Clients with technical or relationship management
knowledge or who highly trust their vendors, use lower
amounts of formal control. However, the amount of
formal control generally increases with higher levels of
task uncertainty.
(Tiwana and
Keil 2007)
Outsourced ISD
projects
The role of specialized
knowledge in the domain of
outsourced activities, i.e.,
‘peripheral’ knowledge, for
control execution
Peripheral knowledge increases the positive
performance impacts of formal outcome controls, but
not formal process controls.
(Tiwana
2008)
Outsourced ISD
projects
The role of intentional
decoupling of interoperating
subsystems of a larger
system, i.e., technological
modularity, for control
execution
Technological modularity lowers the positive
performance impacts of formal process controls, but
not formal outcome controls.
(Tiwana and
Keil 2009)
Internal and
outsourced ISD
projects
Differences in the use and
effectiveness of control
mechanisms in internal versus
outsourced ISD projects
Control mechanisms are used to a greater extent in
outsourced projects relative to internal projects.
However, formal controls do not enhance outsourced
ISD project performance and the positive impact of
informal controls depends upon the level of
requirements volatility.
(Tiwana
2010)
Outsourced ISD
projects
Relationship and interactions
between formal and informal
control mechanisms in a
project’s control portfolio
The combination of informal controls with formal
process-based controls simultaneously enhances the
fulfillment of project goals and development flexibility.
However, combining informal controls with formal
outcome-based controls can impair these objectives.
Table 1: Overview of existing research on control in IS outsourcing and offshoring
In summary, prior research on control in IS outsourcing and offshoring contexts has focused on
formal and informal controls, including their choice, amount, interactions, performance impacts,
factors that moderate the relationship between control and performance, and the evolution of
control portfolios over time. In contrast, what is lacking is research on the different dimensions of
control use and the reciprocal effects of control behavior on the client vendor relationship in an
8
ISD offshoring project. Furthermore, our understanding of the dynamics and evolution of
organizational control in general is limited.
Control Balancing in Organizations
In organizational management research, Barker’s (1993) analysis of self-managed teams was one
of the first studies to examine the evolution of control. His study illustrated the dynamic nature of
organizational control, which motivated further longitudinal research in this area. Choudhury and
Sabherwal (2003) investigated the dynamics of control in offshore outsourced ISD projects and
found that during the course of such projects, portfolios of formal and informal controls change
over time (see Table 1). Kirsch (2004) investigated the dynamics of control in IS projects where
common systems are deployed globally. She found that control choices in such projects change
across project phases due to changes in in the project context (e.g., poor project performance), the
stakeholder context (e.g., changing role expectations), and the global context (e.g., challenges
arising from cultural differences). More recently, Heiskanen et al. (2008) examined the evolution
of trust and control during the process of client-vendor relationship development in the context of
public sector ISD outsourcing. They found that the public sector client’s actions oscillated
between trust and control during the course of relationship development (see Table 1).
A relatively underexplored, yet highly relevant concept that has found its way into the
organizational control literature is the notion of ‘balance.’ The basic idea is that different
combinations of controls are needed depending upon the context and that a control portfolio that
is balanced to suit a particular context is going to be more effective than one that is not (Bradach
and Eccles 1989). Within organizations, the need for balancing results from an increasingly
turbulent and dynamic business environment (Eisenhardt et al. 2010; Smith and Lewis 2011).
Applying the notion of balance to organizational control, Cardinal et al. (2004) examined the
evolution of control and found evidence for a repetitive balance-imbalance-rebalance sequence in
9
which balance is defined as “a state where an organization exhibits a harmonious use of multiple
forms of control” (Cardinal et al. 2004, p. 412). Despite these important theoretical
advancements, research on control balancing in organizations is still in its infancy (Long and
Sitkin 2006; Sitkin et al. 2010). Thus, we have only very limited understanding about the nature
and characteristics of control balancing in temporary organizations (e.g., IS projects) and clientvendor relationship contexts.
RESEARCH METHODOLOGY
We conducted a longitudinal single-case study to analyze the management and control of an ISD
offshoring project. Given the scarcity of research on the dynamics of control in ISD offshoring
(King and Torkzadeh 2008), our goal was to build grounded theory with case study research
(Urquhart et al. 2010). The case we studied involved a German bank that was offshoring
development work to India. This case was particularly suitable for investigating the phenomenon
of interest because it afforded an opportunity to conduct a longitudinal analysis of the
management of the project including both client and vendor perspectives.
Primary and Secondary Data
Primary data collection consisted of 56 interviews conducted between April 2007 and August
2011, with both client and vendor representatives in Germany and India respectively. For the
interviews in Germany, we took extensive notes and transcribed them immediately after each
interview session (Walsham and Sahay 1999). The interviews conducted in India were taperecorded and transcribed. A typical interview lasted 1.5 hours, although we also had ‘long
interviews’ in the very beginning of the research process that lasted up to three hours and short
follow-up interviews of less than an hour in later stages of our research. Interviews were
conducted in an open-ended and unstructured manner initially and became semi-structured over
10
time in accordance with the development of our grounded theory. Similarly, data sampling was
closely aligned with our analysis and emergence of theory (Urquhart et al. 2010). Our primary
data collection was complemented by observations and informal face-to-face discussions during
our field research in Germany and India (see Table 2).
Type of primary data
Description of primary data
56 interviews
Role of interviewee
# of interviews with
client employees
# of interviews with
vendor employees
# of interviews in
total
Project manager/leader
21
11
32
Project team member
6
11
17
Business/functional
management
3
2
5
Top IT management
2
0
2
Total # of people interviewed
22
22
44
Total # of interviews (some
people were interviewed more
than once)
32
24
56
>300 pages of notes and
memos based on
fieldwork
Extensive notes were taken during field visits to capture observations, impressions, results of informal
discussions and meetings. Analytic memos were also crafted based on our reading of the data. These
captured joint reflections that were made within the research team as we combed through the interview
notes and transcripts and discussed the data. In total, more than 300 pages of notes and memos were
created during the analysis process from 2007 to 2011.
Table 2: Primary Data Consisting of Interviews and Fieldwork
In addition, for triangulation purposes, secondary data was collected and analyzed (see Table 3).
Type of secondary data
Description of secondary data
21 steering committee
meeting presentation
slides, each consisting of
20-50 slides
Contained information that was reported on a quarterly basis from the IT department to the business
and IT top management stakeholders. This information included among other things status reports
about the project and the relationship as well as critical issues that required top management
attention. This information helped in reconstructing the evolution of events and activities in the project.
8 sets of project
management slides,
including 5-30 slides each
Project management slides included information on escalation paths, client-vendor communication
plans, project reviews, kick-off presentation slides, business case presentation slides, and project
member matrices. This project documentation helped in understanding the context and the
characteristics of the project and the nature of the client-vendor relationship.
Table 3: Secondary Data that was Collected and Analyzed for Comparisons with Primary Data
Data Analysis and Research Process
Our methodology for investigating the case was ‘pluralist,’ meaning that we engaged in a variety
of interrelated research activities which are summarized in Table A1 in the Appendix (Mingers
2001, p. 245). We adopted a critical realist stance in the sense that ontologically we assumed the
existence of an independent reality and from an epistemological perspective we sought to “posit
11
descriptions of reality based on an analysis of the experiences observed and interpreted by the
participants, along with other types of data” (Wynn and Williams 2012). Our approach was also
consistent with Mingers (2004) who advocates a pragmatic approach in which researchers
embrace different research perspectives with the aim of generating “a useful model of reality”
(Van de Ven 2007, p. 59). We also followed the principle of ‘emergence’ from grounded theory:
grounded theory must ‘fit,’ the data under study, and it must meaningfully explain the behavior
under study (Glaser and Strauss 1967, p. 3). In addition, techniques were selected and used that
helped us increase the theory’s scope and the degree of conceptualization (Urquhart et al. 2010).
Our research process started with formulating the problem (Van de Ven 2007), designing the case
study (Yin 2003), and engaging in intertwined data collection and analysis for building grounded
theory, i.e., theorizing from our data by developing abstract categories and specifying the
relations between them (Bryant and Charmaz 2007). In coding for theory generation, an
important technique in grounded theory (Urquhart 2001; Urquhart 2007), we were guided
primarily by Glaser’s approach to the Grounded Theory Method (Glaser 1978; Glaser 1992). The
table in Appendix A illustrates in detail the coding and data analysis process. For pragmatic
reasons, i.e., “in order to participate in the current theoretical conversation” (Lempert 2007, p.
254) and for “relating the theory to other theories in the same or similar field” (Urquhart et al.
2010, p. 369), we gradually conducted an extensive literature review (see Table 4).
Type of
Literature
Review
Criterion for Literature Selection and
Examples
Scope of Literature Review and
Extant Knowledge Engagement
Goals of Literature Review
Problem
domain
literature
review
Literature related to the problem domain of
our study, i.e., managing ISD outsourcing/
offshoring projects and relationships (e.g.,
Dibbern et al. 2008; Koh et al. 2004; Levina
and Vaast 2008; Rai et al. 2009)
 Identification of >200 and careful
reading of approx. 40 works
 Presentations at >10 workshops
(e.g., Global Sourcing Workshops
and IS conferences (e.g., ICIS))
 Identification of domain
literature gaps
 Selection of case study
 Domain literature
integration
Theoretical
domain
literature
review
Literature related to the theoretical domain
of our study, i.e., organizational control
(e.g., Cardinal et al. 2004; Choudhury and
Sabherwal 2003; Kirsch 1997; Kirsch 2004;
Tiwana 2010; Tiwana and Keil 2007)
 Identification of >300 and careful
reading of approx. 60 works
 Participation in related research
projects operating in the same
theoretical domain
 Identification of theoretical
gaps
 Definition of research
question(s)
 Theoretical integration
Meta-
Non IS conceptual literature that increases
 Identification and screening of ~10
 Scaling up grounded
12
theoretical
literature
review
theoretical sensitivity and supports
theorization, i.e., the notion of balancing in
organizations (e.g., Bradach and Eccles
1989; Eisenhardt et al. 2010; Smith and
Lewis 2011; Sundaramurthy and Lewis
2003)
potentially interesting metatheoretical concepts and lenses
prior to the selection of ‘balancing’
as core lens to conceptualize data
theory
 Increasing theory’s scope
and facilitating
conceptualization
Table 4: Role and Use of Theory or Extant Literature in our Study
During our research and analysis process, we also gave special attention to the sequencing of
events and delineating control phases (stable states of a particular control configuration) from
triggering conditions (i.e., the conditions that led to a change in the control configuration)
(Langley 1999; Mohr 1982; Van de Ven 2007). Finally, we integrated our theoretical results into
the theory and problem domain literatures.
BRIEF CASE DESCRIPTION
Our case involved a 5-year-long project between a German client and an Indian vendor in the
financial services industry. The project involved the technical reengineering and integration of a
core banking system used for daily financial transactions worth billions of Euros. This system
consisted of two distinct transaction processing systems (TPS) with overlapping functionality,
i.e., a 25-year-old TPS focused on checking accounts with 1.3 million lines of code operating on
a batch-processing architecture, and a 7-year-old TPS focused on payment with 4.5 million lines
of code operating on a real-time processing architecture. Specific goals of the project included the
elimination of function and data redundancy, the reduction of system complexity, and avoiding
further dependency on key knowledge (e.g., about old programming languages) for which it
would be difficult to recruit new people to replace retiring specialists. The endeavor required
more than 170 project members and the overall project manager compared it to a heart surgery
due to its high level of risk and criticality to the business. To accomplish it, the bank outsourced
approximately 60% of the project work to a Capability Maturity Model Integration (CMMI)-5
certified offshore service provider from India which was contracted on a fixed-price basis.
CASE ANALYSIS RESULTS: CONTROL BALANCING IN ISD OFFSHORING
13
Gradually, the research process summarized above directed us toward the notion of ‘control
balancing.’ As a result, we decided to embark on a journey to develop a grounded theory of
control balancing in ISD offshoring, which we present in this paper. We start our theory
development by discussing key assumptions and boundary conditions.
Assumptions and Boundary Conditions
Our examination of control balancing in ISD offshoring was based upon certain assumptions that
both guided our theory development and set the boundary conditions for the application of our
theory (Bacharach 1989). One of our assumptions was that there cannot be a single best control
configuration that holds across all phases of a project and that control configurations must change
over time. Another assumption was that offshoring relationships evolve over time due to
differences in work practices, barriers to knowledge transfer, and cross cultural challenges.
Finally, we focused on inter-organizational control as opposed to intra-organizational control
under the assumption that in the offshore outsourcing context, it is the client-vendor dyad that
shapes the context for control.
The boundary conditions of our theory are set by the above assumptions as well as both the
organizational and project context of our study. Specifically, our theory is grounded in a 5-yearlong project between a German client and an Indian vendor in the financial services industry. The
project involved the technical reengineering of a core banking system and the task required more
than 170 project members. A Capability Maturity Model Integration (CMMI)-5 certified offshore
service provider from India was contracted on a fixed-priced basis to outsource approximately
60% of the project work. Particular organizational characteristics that need to be considered when
applying our theory to new contexts include the large size of the project, the long duration of the
project, the complexity and risk of the core banking system reengineering task, the German-
14
Indian offshoring relationship context, the selective outsourcing arrangement in which client and
vendor distributed the workload almost evenly, the fixed-priced contract, and the special
commitment of the vendor to enter into the German market.
Core Category and Related Concepts
To develop a shared vocabulary for the subsequent process analysis of our case, we start the
development of our grounded theory by introducing and defining the core category ‘control
balancing’ that we developed from our data. This category explains what needs to be balanced in
terms of control in ISD offshoring projects and depicts along which dimensions control
configurations are changed over time in ISD offshoring projects. Accordingly, we define control
balancing as making adjustments to the control configuration periodically in terms of control
types, control degree, and control style, to allow the ISD offshoring project and relationship to
progress. The three dimensions of control balancing decisions that emerged from our data are
explained and defined in Table 5, together with codes and indicators that illustrate their
emergence from our data as well as properties and associated ranges of each dimension.
Dimension
Definition / Meaning
Properties / Ranges
(control configuration level)
Control
types
Refers to the selection of control
mechanisms for the control portfolio,
which can be of a procedural type, i.e.,
oriented toward improving efficiency and
effectiveness, social type, i.e., oriented
toward developing shared understanding,
or hybrid, i.e., serving both goals
simultaneously
- goal orientation
(relatively procedural –
relatively social)
- Procedural controls (e.g., status
reviews, definition of roles and
responsibilities, tracking project
milestones)
- Social controls (e.g., information
exchange, socialization activities,
intercultural workshops)
- Hybrid controls (e.g., site visits,
reflection-in-action sessions,
common understanding
workshops)
Control
degree
Refers to the degree of control use, which
can vary from tight, i.e., high amount of
control in terms of a large overall number
of controls used simultaneously and/or a
high frequency of control execution,
usually paired with a high intensity of
control use, to relaxed, i.e., low amount of
control in terms of a small number of
controls used simultaneously and a low
frequency of control execution, usually
paired with a low intensity of control use
- amount: number
(large – small)
-
- amount: frequency
(high – low)
- intensity
(high – low)
Codes / Indicators (Selection)
Multiple control use
Amount of control
Control desire
Control completeness
Ability to let go
Excessive control
Frequency of control use
Relaxed control attitude
Exerting pressure with control
Intensity of control
Focus on control
15
Control
style
Refers to the style of control use, which
can vary from unilateral, i.e., focus on one
party (typically the client) controlling the
other party (typically the vendor) in a onesided manner without emphasizing
mutual agreement and with a clear
separation of controller/controllee roles,
to bilateral, i.e., focus on the client and
the vendor using control mechanisms
jointly with an emphasis on mutual
agreement of control selection and use,
resulting in blurring boundaries between
controller/controllee roles
- role ambiguity
(high – low)
- use of controls
(jointly – one-sided)
- mutual agreement
(strong – weak)
-
Agreement with control
Controller vs. controllee roles
Mutual control use
Control direction
Control harmony
Control coordination
Perceived control necessity
Joint control use
One-sided control use
Control perception
Table 5: Dimensions of the Core Category Control Balancing
By examining the changing three-dimensional control configurations and recurrent control
balancing decisions over time in our case, we realized that control balancing decisions in ISD
offshoring drive and are driven by shared understanding in the client-vendor relationship. In
Table 6, the four areas of shared understanding that emerged from our data when examining
control balancing behavior over time are defined and explained, together with codes and
indicators from our data.
Areas of Shared
Understanding (SU)
Definition / Meaning
Codes / Indicators (Selection)
Project processes
(SU1)
Shared understanding about
project processes refers to
mutual agreement about project
goals, objectives, and tasks, as
well as operational approaches,
processes, and procedures for
achieving them
-
Business-functional
knowledge
(SU2)
Shared understanding about
business-functional knowledge
refers to joint knowledge about
the client’s business application
domain, the client’s systems, IT
infrastructure, and the functional
requirements
-
Mutual expectations
(SU3)
Shared understanding about
mutual expectations refers to
negotiated views and consensus
about mutual expectations, roles
and responsibilities, as well as
their fulfillment
- Roles and responsibilities residing with the client and the
vendor
- Distribution of workload between client and vendor in each
phase and task area
- Expectations fulfillment or nonfulfillment
- Unrealistic expectations
- Clarifying and communicating expectations
- Adaptation of expectations
- Clear separation of client and vendor responsibilities
- Shared client and vendor responsibilities
Reengineering goals and objectives
Project milestones and timelines
Project phases, sub-projects, and work streams
ISD approaches
Project methodology
Initial requirements gathering
Functional specifications
Technical specifications
Coding
Unit test / component integration test / quality control test /
integration test / user acceptance test
- Configuration / rollout
Financial services industry domain knowledge
Understanding of client’s business and processes
Functionality of client’s systems
Functional requirements
Business consequences of systems failure
Understanding risk of core banking reengineering
Client-specific and banking-specific terminology
Implications of systems shutdown
16
- Development of shared expectations
Social practices
(SU4)
Shared understanding about
social practices refers to shared
views and compromises about
social communication and work
practices, resulting from mutual
behavioral adaptation
-
Openness / frankness / transparency of communication
Discussion and consensus finding
Communication of problems and issues
Handling of software documentations
Motivation to learn new social practices
Style of asking questions
Punctuality / time attitude
Perfectionism
Role of family
Respect
Aloofness
Accommodation and avoiding disagreement
Planning / risk averseness
Hierarchy and status boundaries
Negotiated culture
Mutual learning and adaptation
Table 6: Areas of Shared Understanding Relevant to Control Balancing in ISD Offshoring
The introduction of these foundational definitions and conceptualizations provides the necessary
vocabulary to explain the process of control balancing in ISD offshoring projects, based on our
longitudinal case analysis. In the case presentation and analysis that follows, we characterize the
evolution of control balancing as proceeding through four phases involving three distinct control
configurations. While the particular phases and trajectory that we observed may be case specific,
we believe that the three control configurations can be generalized to other contexts.
Control Balancing Phase 1: Coordinated Control
Client and vendor in our case initially implemented a control configuration that we call
coordinated control (summarized in Table 7).
Phase
Triggering Conditions
Control Balancing Decisions
Consequences
Phase 1:
Coordinated
control
Gaps in shared understanding
 Need to reach an
agreement on joint project
processes (SU1)
 Need to transfer businessfunctional knowledge (SU2)
 Need to clarify mutual
expectations (SU3)
 Need to deal with
differences in social
practices (SU4)
 Combination of procedural
controls (operational process
documents, formal communication
plan, mirror structure), social
controls (socialization activities
outside of work, intercultural
workshops), and hybrid controls
(common understanding
workshops), resulting in a hybrid
control portfolio
 Tight control degree: High amount
of control
 Bilateral control style: Focus on
 Development of good
working relationships in
various client-vendor work
streams and sub-projects
based upon mutual
agreement about joint
project processes (SU1),
negotiated mutual
expectations (SU3), and a
perceived good level of
shared social practices
(SU4), as well as joint
business-functional
17
joint control based on mutual
agreement of control selection
and use
knowledge (SU2)
Table 7: First Phase of Control Balancing in Our Case
When the project started in late November 2004, client and vendor faced a situation of sharing
only a minor relationship history with significant gaps in shared understanding. Joint processes
for working together in the project had to be established (SU1) and mutual expectations had to be
clarified (SU3) as this project was the first major engagement between the client and the vendor.
A fixed-price contract had been negotiated during the 9 months prior to project kick-off in which
high-level expectations had been already clarified, but they had to be translated into operational
roles and responsibilities on the project timeline. In addition, the vendor program manager
explained the need to deal with differences in social practices (SU4):
“What I recognized as one of the major challenges, which we had to handle upfront…this was
one of the units in [name of client], which had not been exposed to outsourcing at this scale
before…and all of a sudden so many Indians arrive [laughs] and you [the Germans] don’t know
how to deal with them.”
While the vendor did have prior experiences in the U.S. and European financial services industry,
working together with German clients was relatively new for them. More importantly, the task of
reengineering a 25-year-old legacy and core banking system of this magnitude and size was also
new to the vendor and business-functional knowledge had to be transferred (SU2). These
antecedent conditions triggered coordinated control:
“We [client and vendor] realized early on that we both had a strong interest in this project and
that the complexity of the project, requiring deep technical and business-functional knowledge,
required intensive cooperation…we knew we could only do this together…the first thing we did
was to decide jointly on a number of mechanisms to get the relationship going.”
[client program manager]
Even before the first vendor team members arrived in Germany, the client and the vendor
planned and coordinated the first coordinated control configuration:
18
“We asked ourselves the question: How can we bring the team up to speed early on, clarify and
explicate mutual expectations, and create an effective shared operational plan? We coordinated
our first actions closely with [name of vendor]-management. One of the things we discussed was
that we needed joint workshops in the beginning to establish common understanding, a joint
working model, and deal with cultural differences.”
[client chief architect]
Accordingly, the above explained gaps in shared understanding triggered client and vendor to
implement both procedural controls (e.g., formal communication plan involving regular project
status meetings, sub-project reviews, steering committee meetings and operational process
documents documenting and tracking goals, roles, and responsibilities), social controls (e.g.,
socialization activities outside of work and ‘intercultural’ workshops to foster shared
understanding about mutual social practices), and hybrid controls (e.g., ‘common understanding’
workshops involving discussions about project methodology, templates, coordination tools, as
well as configuration and release management), resulting in a hybrid control portfolio.
Furthermore, all control mechanisms were planned, coordinated, and implemented jointly, which
was the consequence of the bilateral control style decision to foster mutual cooperation early on
in the relationship. Finally, the control degree was relatively tight. This is illustrated by the
relatively high amount of control in this first phase, particularly in terms of the overall number of
controls used simultaneously.
The main effect of coordinated control was the “development of good working relationships
between client and vendor personnel” [excerpt from the steering committee meeting slides after 3
months of cooperation], which was based mainly upon mutual agreement about joint project
processes (SU1) and negotiated mutual expectations (SU3). In addition, both client and vendor
thought they had reached a fairly good level of shared understanding about social practices (SU4)
and business-functional knowledge (SU2). However, what both client and vendor senior
19
management did not realize at that point in time were the persistent business-functional
knowledge gaps (SU2):
“They [client members] gave us unimaginable amounts of raw software code and related
documentation … they expected us to analyze the old software to derive functional requirements
and translate them into technical requirements for the new software which from our point of view
was simply not possible.” [vendor project manager]
While knowledge gaps (i.e., about the client’s business application domain, the client’s systems,
IT infrastructure, and the functional requirements) were the main reasons for the challenges that
were to follow, persistent differences in social practices (SU4), especially those relating to
communication, exacerbated the situation. For example, the vendor preferred agreement over
conflict while client project members emphasized open and frank communication, even if
conflictive discussions would result. As a result of these and other differences, there were
misunderstandings and subsequent heated discussions to clarify them which hampered further
shared understanding development. When the first software deliveries and test results were
produced, it became clear that things were not developing as expected and the client sensed that
their initial expectations had been somewhat unrealistic.
Control Balancing Phase 2: Authoritative Control
The challenges explained above relating to shared understanding in the client-vendor
relationship, which occurred in spite of careful control balancing decisions and a coordinated
control configuration, culminated in unmet expectations (SU3). The client architecture lead
commented on his experiences from the time when the project approached its first deadlines in
mid-2005:
“Something was always delivered to a certain deadline, even if they [vendor developers] had to
work throughout the whole night. But what was delivered usually did not meet our expectations.
We had to invest a lot of additional resources to rework what was delivered.”
20
Mutual expectations also remained unfulfilled from the vendor’s point of view who had expected
more efforts on behalf of the client to help build up the necessary business-functional knowledge.
For example, parts of software code had German language comments and many banking-specific
or even client-specific concepts that were difficult to understand for an external vendor.
Unfulfilled expectations brought significant shortcomings in shared understanding to the surface
and triggered control balancing decisions, resulting in an authoritative control phase (Table 8).
Phase
Triggering Conditions
Control Balancing Decisions
Consequences
Phase 2:
Authoritative
control
Unfulfilled expectations (SU3)
 Project goals and timeline
threatened
 Need to emphasize more
procedural efficiency and
effectiveness
 Need to increase the
pressure on the vendor
 Control portfolio dominated by
procedural controls (status
reviews, inspection of
documentations and reports,
testing deliverables, tracking
goals and milestones)
 Tight control degree: high amount
of control paired with a high
intensity of control use
 Unilateral control style: focus on
the client controlling the vendor
without emphasizing agreement
Increase in project effectiveness
at the expense of shared
understanding deterioration
(SU3)
 Conflicts and communication
breakdown
 Shared understanding in the
offshoring relationship
deteriorates, leading to further
unfulfilled expectations
=> vicious cycle
Table 8: Second Phase of Control Balancing in Our Case
The reasoning behind this shift to authoritative control was explained to us by a senior client
project manager:
“The use of this sort of authoritative control approach was triggered by quality or time issues.
Sometimes we were still on time, but the quality was miserable which is when we got
disappointed with the vendor who delivered just ‘something’ to the deadline. Then our status
light turned red which alarmed us and so we needed to change something. Our reaction was: we
need to control the vendor more tightly. We consciously sent out a signal that there were
unacceptable issues with the deliverables.”
Accordingly, gaps in shared understanding could not be closed on a consistent basis, leading to
vendor deliverables that did not fulfill the client’s expectations and that posed a threat to the ontime achievement of the project’s goals and objectives with the expected quality. This further
alarmed the client and led to the perception that an even tighter degree of control was needed and
that the pressure on the vendor needed to be increased. This was done by emphasizing procedural
efficiency and reminding the vendor that they were responsible for delivering results. As a
21
consequence, the control degree was tightened by increasing further the amount of control (in
particular the frequency with which controls were used) as well as by focusing even more
attention and resources to rigorous and intensive control as illustrated by the following quotes:
“You need to control everything yourself, even a test document. Nothing, but really nothing can
be accepted without control.” [client architecture lead]
“Very detailed control was the important thing. At the end of the day you needed to inspect every
single thing…that was more important than anything else.” [client project manager]
The client program manager even explained that at times certain control mechanisms such as
status reviews were used on a daily basis to control more tightly. In addition, consciously sending
out a signal to the vendor by controlling deliverables more intensively illustrates the decision for
a more unilateral control style, which is also a characteristic of authoritative control. Finally, with
the decision to control the vendor more tightly and to increase the efficiency and effectiveness in
the project, the overall control portfolio became relatively procedural. As a consequence of these
control balancing decisions, the vendor clearly realized that something had to be changed and
improved, which triggered the search for higher degrees of project effectiveness. The vendor
program manager explained:
“It was after around one year of working together that in one of the sub-projects we faced an
issue. The level of detail expected by the client project managers turned out to be completely
different than what we had been delivering ... We established extra ‘focus teams’ in [offshore
location] to address performance issues.”
However, authoritative control also caused a further deterioration in shared understanding. In
particular, as illustrated by the following quote, there was a breakdown in communication:
“This [referring to the change in client control behavior] initiated multiple rounds of
discussions. The question raised in our team was ‘why are they doing this?’ Yes, details are
important, but results are also equally important. The situation escalated…somewhere that
element of discussion was lost, people were not talking to each other anymore. Yes, if I talk to
you, I can hopefully explain to you why I think this is sufficient and you can explain to me your
22
views on why you think it is not sufficient. That discussion thread had started to break down.”
[vendor program manager]
This breakdown of communication hampered the accumulation of shared business-functional
knowledge (SU2) and ultimately led to further unfulfilled expectations (SU3) and a vicious cycle.
After the problems were escalated to the top management level, the client was on the verge of
“pulling out the contract” [client program manager] and the project was headed toward failure.
Control Balancing Phase 3: Coordinated Control
Both the client and the vendor were driven by a strong mutual commitment toward project
success. As the client program manager commented: “Failure was no option.” In a series of joint
reflections between key client and vendor project members, gaps in shared understanding in all
four above mentioned areas (SU1-4) were identified. Realizing and reflecting upon these gaps led
to control balancing decisions and a new phase of coordinated control (see Table 9).
Phase
Triggering Conditions
Control Balancing Decisions
Consequences
Phase 3:
Coordinated
control
Gaps in shared understanding
 Need to renegotiate joint
project processes (SU1)
 Need to close businessfunctional knowledge gaps
(SU2)
 Need to renegotiate mutual
expectations (SU3)
 Need to foster negotiation of
shared social practices
(SU4)
 Combination of procedural
controls (redefinition of roles and
responsibilities, joint parallel
testing approach), social controls
(workshops, feedback-driven
communication, lessons learned
sessions) and hybrid controls
(offshore site visits, reflection-inaction sessions, replay sessions),
resulting in a hybrid control
portfolio
 Tight control degree: High amount
of control paired with a high
intensity of control use
 Bilateral control style: Focus on
joint control based on mutual
agreement of control selection
and use
Shared understanding
development (SU1-4)
 Collaborative working
relationships
 Shared understanding
development and learning in
the offshoring relationship
fosters the use of coordinated
control
=> virtuous cycle based on
shared experiences and
learning
Occasional setbacks to
authoritative control
 Client and vendor cycled
several times back and forth
between authoritative and
coordinated control
Table 9: Third Phase of Control Balancing in Our Case
While the control degree remained tight in this phase of coordinated control, a conscious decision
was made to adapt the control style. The following quote by a senior client project manager
illustrates the control balancing decision toward a bilateral control style:
23
“At some point in time we arrived at the conclusion that we could not continue to simply control
the performance of the vendor… rather, we realized, in fact both we and the vendor realized that
we could only master the challenges in the project by working together very closely.”
A senior vendor project manager, who was involved in most of these critical joint reflections,
explained:
“It was in one of these quarterly meetings, the project reviews that we usually conducted
iteratively in [client location in Germany] or [vendor location in India]. It was a gradual process
of joint reflections, but… the key tipping point was probably this meeting where we discussed the
gaps and decided [on] a change in methodology and mutual expectations”
From this point forward the distribution of workload changed and new role expectations were
negotiated (SU3). In the end, client and vendor agreed that they both had to invest additional
resources and that the client would have to take over more responsibility concerning key issues
such as cooperating with the vendor to build up business-functional knowledge in the relationship
(SU2). The vendor assigned a very senior banking expert to the project who worked closely with
the client in Germany for one year. One of his achievements was producing a 22-page long
‘vocabulary list’ with key banking- and client-specific concepts and their explanations as a
learning document for the project teams.
New social and hybrid controls were added to the control portfolio and the degree of control
remained tight, reflecting the shift from authoritative control back to coordinated control. An
example is the decision (that was taken jointly by client and vendor) to initiate offshore site visits
in which client project managers were sent in small groups for one to two weeks to the offshore
vendor location. One of the client project managers commented on this experience:
“The site visits to India were helpful …to talk to them [vendor offshore personnel]…all of us
[project leaders] went to [vendor location in India] at some point in time…now we [client and
vendor staff] feel much more comfortable to communicate with each other. Afterwards, the
cooperation was totally different. Personal contact is extremely important. People now have the
heart to call each other and talk openly about the issues.”
24
The site visits helped to close the gaps in shared understanding with regards to social practices
(SU4) as well as business-functional knowledge (SU2). They also facilitated joint problem
solving by client and vendor project members, a key implication of coordinated control. The
client program manager commented:
“Always when they [client project managers] were there, the project status was on a green light!
When you get to know the people there locally, then this helps to build up a good project
climate.”
As a further example of coordinated control in this phase of cooperation, a client project manager
commented on the use of joint ‘reflection-in-action’ in client-vendor face-to-face meetings as
well as extensive coaching of less experienced team members.
“I had regular meetings with my counterpart from the Indian organization. We discussed
different perceptions, misunderstandings, cultural and communication issues, and reflected
jointly on the experiences (positive and negative) we gained from working together in the project.
Afterward we carried out extensive coaching with our project members on both sides… That way,
we were able to solve many conflicts and misunderstandings in the project and motivate our
project workers.” [client project manager]
With coordinated control in place, client and vendor groups started to understand each other
better and started to learn more intensively about the social practices that were deeply embedded
in client and vendor work teams (SU4). For example, vendor personnel started to communicate
more openly concerning any critical issues or problems they faced. Client personnel also started
to learn and adapt. For example, after learning about the Indians’ feelings in situations of
disagreement and conflict and their desire for accommodation, they started to develop new
communication techniques. A client project manager explained:
“What I do in such situations is [I] ask my communication counterpart actively whether he really
meant ‘yes’ or ‘no’ when I realize that he is not being totally serious with his statement.
Basically, you have to ask: ‘What other options are there?’ and somehow try to stimulate the
communication through open instead of closed ‘yes’ or ‘no’ questions. In such situations you
have to communicate ‘the Indian way’.”
25
Client and vendor also identified the need to renegotiate joint project processes and made an
adaptation to the testing approach (SU1):
“We implemented a joint parallel testing approach. Basically, we introduced one more phase and
let old and new programs run in parallel both in Germany and India. Recurrently, we would
compare results and output and if it was different, then we analyzed why and asked ourselves
whether we missed something and if yes where…in the requirements stage or some other phase?
The new testing approach proved to be so effective that we adopted this as standard practice for
subsequent sub-projects and work streams.” [senior vendor project manager]
As a result of the shift toward coordinated control, client and vendor began to cooperate more
intensively with each other and started to make significant improvements in the areas of shared
understanding about business-functional knowledge (SU2) and mutual expectations (SU3), which
gradually also led to better deliverables. Based on the joint experiences in the project, a virtuous
cycle of continued shared learning and understanding and coordinated control developed.
However, the shift from authoritative control to coordinated control was anything but smooth and
linear:
“We cycled back and forth several times… But the key thing was that we were always able to sit
together with [name of vendor] at the table, identify the gaps, and decide jointly the necessary
changes. That’s what I liked about [name of vendor], this constructive cooperation.” [client
program manager]
Over time, the virtuous cycle mentioned above led to the selection and use of all kinds of new
mechanisms that were instantiations of coordinated control. Project managers started to
experiment with regular telephone calls to India, the placement of client and vendor onsite teams
in joint open plan offices, and the use of shared blackboards for status tracking in onsite teams.
Client and vendor members would also meet regularly to discuss open issues and solve problems
jointly which was an implication of coordinated control. Gradually, the deliverables in the project
improved significantly. A critical roll-out of the most important sub-project was successfully
26
completed in June 2007 and the deliverables of the vendor had improved to a level that the client
was satisfied with the results.
Control Balancing Phase 4: Trust-Based Control
Fulfilled mutual expectations (SU3) triggered control balancing decisions, resulting in trustbased control (Table 10).
Phase
Triggering Conditions
Control Balancing Decisions
Consequences
Phase 4:
Trust-based
control
Fulfilled expectations (SU3):
 Need to decrease the
amount of resources
dedicated to control
 Need to foster trust and
increase the maturity of the
relationship
 Control portfolio dominated by
social controls (spontaneous
phone calls, brainstorming
sessions, informal exchange of
ideas, direct and pragmatic
communication)
 Relaxed control degree: low
amount of control paired with a
low intensity of control use
 Bilateral control style: Focus on
joint control based on mutual
agreement of control selection
and use
 Shared understanding became
established (SU1-4) and
paved the way for a trustbased client-vendor
relationship
Occasional setbacks to
coordinated control
 The relaxation of control
degree gave room for new
issues to emerge that
challenged client-vendor
shared understanding, caused
issues in the project, and
triggered more coordinated
control from time to time
Table 10: Fourth Phase of Control Balancing in Our Case
The following quotes by the client program manager and two project managers illustrate why the
control balancing decision toward a relaxed control degree was eventually taken, a key
characteristic of trust-based control:
“Our project managers went through a steep learning curve, both with regards to social and
procedural aspects. This learning in the offshoring relationship gradually led to shared
understanding and trust… you don’t have to control every single thing…after five successful
deliverables the sixth will also be fine…project managers then relaxed the reins.” [client
program manager]
“The moment in which three spot tests in a row meet your requirements, that’s when you start
trusting each other [referring to client and vendor project members]. Trust starts replacing the
need for controlling tightly.” [client project manager]
“Trust made a dense nexus of control obsolete.” [client project manager]
When our participants used the word ‘trust,’ they referred to the fulfillment of expectations (SU3)
and the essential consequential belief that they could rely on the vendor’s capability to deliver
27
high quality results within an agreed upon timeframe without having to be tracked and controlled
intensively. In this phase of trust-based control, the selection and use of control mechanisms
changed. Many control mechanisms were of a social type, focusing on improving further shared
understanding in the relationship:
“In the beginning we had many discussions [referring to conflicts over the quality of deliverables
and related issues], but today conversations are more toward optimizing performance; our
relationship matured, we now talk about more advanced topics, not day-to-day disagreements.”
[senior vendor project manager]
Client and vendor participants in our study commented on the use of frequent spontaneous
communication, the exchange of ideas, and joint brainstorming sessions. Communication also
became more direct (not following always the previously defined formal pathways) and
pragmatic. Furthermore, trust-based control led to a reverse situation in comparison with
authoritative control: now it was frequently the vendor who was in the ‘lead’ and providing
direction for where to go and how. A senior client project manager explained:
“For me a key experience was the moment when a problem could not be solved internally in our
organization anymore and [name of vendor] took over full responsibility. Then I realized that
something fundamentally had changed. I didn’t have to be the one anymore who knows
everything.”
A vendor project manager gave us a few illustrative examples which made clear that the vendor
was taking over the lead gradually and the client continuously reduced the size of its managerial
work force assigned to the project. Also, the client program manager commented on this:
“It was at the end of 2008 when I began to realize something. The interest calculations for many
customers had defects. Everybody was on vacation except myself and then we had a crisis
meeting. That was the first crisis meeting of such kind where [name of vendor] was at the table.
They analyzed the situation and came up with a solution to the problem already the next day.
They basically solved the entire bandwidth of problems all by themselves. It became clear to me:
They had understood.”
28
Gradually, trust-based control fostered the establishment of shared understanding (SU1-4) and
the development of a truly trust-based relationship, as illustrated by the following quotes:
“The element of success comes from trust…whether you trust me, that what I will do is something
which is good for you. That trust developed in our relationship and bound us together…that
really helped in making it happen.” [vendor program manager]
“There are no boundaries between the two teams. We always work together for the project. So,
whenever we interact … we always talk about the project; we never talk about the individuals.
We constantly ask ourselves how to improve things together… I talked to them as my colleagues,
working for the same project, working for the same goal. That was the very best thing.” [vendor
architecture lead].
However, similar to the transition from authoritative control to coordinated control, the transition
from collaboration control to trust-based control was not entirely smooth and linear:
“It was a balancing act. Controlling tightly and coordinating control actions intensively with the
vendor cost us a lot of resources, I call it ‘energy modus operandi.’ A lot of overhead and
time…of course you cannot manage in such a modus operandi completely, you need to try to get
to the next stage [referring to the shift to trust-based control which required less managerial
resources]. But then you recurrently get into a new situation where you need to grab the reins
more tightly again and so you jump back into energy modus operandi for some time.” [senior
client project manager]
A Process Model of Control Balancing in ISD Offshoring
The process model in Figure 1 synthesizes our case analysis results and abstracts the most
important findings.
29
Control Balancing and Evolution in ISD Offshoring
Authoritative
Control
Multiple iterations
possible
Unfulfilled
expectations
Shared
Understanding
Deterioration
Coordinated
Control
Multiple iterations
possible
Gaps in
shared
understanding
Trust-based
Control
Fulfilled
expectations
Shared
Understanding
Development
Shared
Understanding
Establishment
Evolution of ISD Offshoring Client-Vendor Shared Understanding
Legend:
Changes in shared understanding triggering control balancing
Control balancing affecting shared understanding
Evolution and dynamics of control balancing
Figure 1: Process Model of Control Balancing in ISD Offshoring
Accordingly, we find from our case analysis that gaps in client-vendor shared understanding
trigger the use of coordinated control with the consequence of shared understanding development
(see Tables 7 and 9). Unfulfilled expectations trigger the use of authoritative control with the
consequence of shared understanding deterioration (see Table 8). Finally, fulfilled expectations
trigger the use of trust-based control, leading eventually to the establishment of shared
understanding (see Table 10). As explained in our case analysis, multiple iterations between
coordinated control and authoritative control as well as between trust-based control and
coordinated control are possible. However, we did not find case-based evidence for a setback all
the way from trust-based control back to authoritative control.
Control Configurations and Balancing States in ISD Offshoring
Control balancing resulted in our case in three different control configurations, which are defined
and explained in Table 11. The assignment of control mechanisms to one particular control
30
configuration state (last column) is a simplified view. Mechanisms such as ‘status reviews’ have
been naturally used throughout the project. However, the mapping of control mechanisms to the
three different control configurations illustrates which control mechanisms were particularly
important in which phase.
Control
Configuration
Definition
Meaning and Implication
Typical Examples of Control
Mechanisms from our Case
Authoritative
control
- Procedural
control
portfolio
- Tight control
degree
- Unilateral
control style
An approach to
control based on
a traditional
client-vendor
perspective
Client-vendor perspective:
- Means that exchange parties see
themselves as ‘client’ and ‘vendor’ with
clearly separated roles and responsibilities
- Implies that the client specifies
requirements, the vendor delivers upon
them, and control selection and use is
driven by the client
- Results in situations in which the client
dominates the relationship from a
managerial perspective
-
-
Status reviews
Detailed examination of
deliverables
Tracking of project goals
Definition of client/vendor roles
and responsibilities
Operational process documents
Coordinated
control
- Hybrid control
portfolio
- Tight control
degree
- Bilateral
control style
An approach to
control based on
a coordination
perspective
Coordination perspective:
- Means that exchange parties see
themselves as partners that need to closely
coordinate activities
- Implies that client and vendor work toward
coordinated, shared goals, search jointly for
problem solutions, and select and use
control mechanisms accordingly
- Results in situations in which neither the
client nor the vendor dominates
-
Joint parallel testing approach
Site visits
Workshops
Reflection-in-action sessions
Coaching of team members
Joint communication plan
Lessons learned sessions
Feedback mechanisms
Trust-based
control
- Social control
portfolio
- Relaxed
control
degree
- Bilateral
control style
An approach to
control based on
a trust-based
perspective
Trust-based perspective:
- Means that exchange parties see
themselves as part of the same team based
upon mutual trust and shared
understanding
- Implies that the vendor delivers without the
client getting deeply involved in the process
and less resources are dedicated to
controlling the vendor
- Results in situations in which problems get
solved instantly, new ideas for
improvements are generated, and the
vendor takes over more responsibility in the
relationship
-
Direct and pragmatic
coordination
Brainstorming sessions
Spontaneous communication
(e.g., facilitated through open
plan offices)
Informal exchange of ideas
-
-
-
Table 11: Control Configurations and Balancing States in ISD Offshoring
Figure 2 depicts the three dimensions of control balancing decisions, illustrating that controllers
in ISD offshoring need to make recurrent decisions with regards to the appropriate control
configuration. The arrows in the figure depict the stylized non-deterministic evolution of control
balancing observed in our case study.
31
control
type
c)
social
a) Authoritative control
(procedural; tight; unilateral)
b) Coordinated control
(hybrid; tight; bilateral)
b)
hybrid
control
degree
c) Trust-based control
(social; relaxed; bilateral)
relaxed
procedural
a)
tight
unilateral
bilateral
control
style
Figure 2: The Three Dimensions of Control Balancing
DISCUSSION AND THEORETICAL INTEGRATION
In studying the dynamics of control associated with a 5-year long ISD offshoring project, our
data prompted us to examine the related phenomenon of ‘control balancing,’ which became the
basis for our theory development. Our grounded theory of control balancing in ISD offshoring
explains what needs to be balanced (see Figure 2, Table 5, and Table 11), how control balancing
occurs (see Figure 1 and Tables 7 to 10), and why control balancing matters (illustrated by the
dynamic interplay of control balancing decisions with shared understanding in Figure 1 and
Tables 7 to 10). In the following, we discuss our theory and integrate it into the literature.
Control Balancing
We conceptualized control balancing as making targeted adjustments to the control configuration
periodically along three different dimensions, i.e., control types, control degree, and control style.
This departs significantly from prior research, which has focused solely on one dimension, i.e.,
32
control types (Cardinal et al. 2004), acknowledging the need for control balancing (Gopal and
Gosain 2010, p. 978), without examining in detail what needs to be balanced, how control
balancing occurs, and why control balancing matters. Thus, the grounded theory of control
balancing in ISD offshoring developed in this paper represents the first systematic attempt to
explain organizational control balancing. With ‘balancing,’ we specifically refer to recurrently
and intentionally tuning and directing control actions toward key objectives (in our case, ISD
offshoring project and relationship progress), by finding the appropriate configurations along the
three identified control dimensions. Our findings, summarized in the integrative process model in
Figure 1, suggest that the process of control balancing is much more multifaceted, dynamic, and
iterative than assumed in existing literature, which basically has focused on linear shifts in the
creation of control portfolios over time (Cardinal et al. 2004; Choudhury and Sabherwal 2003;
Kirsch 2004).
Control types is one dimension of control balancing that emerged from our data. We found that
control mechanisms can be of a procedural type, i.e., oriented toward the goal of improving
efficiency and effectiveness (e.g., status reviews), social type, i.e., oriented toward the goal of
developing shared understanding (e.g., socialization activities), or hybrid type, i.e., oriented
toward both goals simultaneously (e.g., common understanding workshops). This goal-oriented
typology of control mechanisms fits our notion of control balancing, which emphasizes making
adjustments to facilitate both project and relationship progress. However, it also represents a
departure from past organizational and IS control theory that has focused on explaining control
mechanisms in terms of their degree of formality (Kirsch 1996; Ouchi 1979).
Control degree, is the second dimension of control balancing that emerged from our data, and
depicts how firmly control is exercised, ranging from tight to relaxed. We found that the amount
of control (measured in terms of both the frequency of control use and the number of controls
33
used simultaneously), and the intensity of control are useful properties to depict the degree of
control at any given point in time. Thus, our conceptualization of control degree significantly
extends and refines prior theorizing, which speaks only to the notion of ‘amount of control’
(Henderson and Lee 1992; Rustagi et al. 2008). To provide an illustrative example of the added
explanatory value, conducting frequent project status reviews is not necessarily sufficient for the
control degree to be tight. One must also consider the intensity with which the status reviews are
performed, i.e., the level of scrutiny to which project is exposed in terms of questioning details
and challenging results.
Control style is the third dimension of control balancing that we identified, and depicts how
control is enacted in the client-vendor relationship, i.e., from relatively unilateral to bilateral. In
particular, our findings provide evidence that in an inter-organizational relationship, such as an
ISD client-vendor offshoring relationship, ambiguity can arise as to who actually controls whom.
Our study illustrates that such ambiguity may result from applying different approaches to control
selection, implementation, and use (i.e., unilateral vs. bilateral) and that this aspect of control
balancing is not captured under existing control theory. In the unilateral control style, ambiguity
concerning the role of controller and controllee in the inter-organizational relationship is
relatively low, as control is exercised mainly in a one-way direction (typically the client
controlling the vendor). Further, in the unilateral control style, controllers do not emphasize
mutuality with respect to control selection and use. On the other hand, in a bilateral control style
ambiguity concerning the role of controller and controllee is relatively high and boundaries can
begin to blur. In the bilateral control style, exchange parties engage in mutual coordination of
control actions and exercise control jointly. As illustrated by our case study, there can be mutual
agreement with regards to control, though this finding is not one that would be predicted based on
prior literature in which the client is typically viewed as the controller and the vendor as
34
controllee (e.g., Tiwana and Keil 2007). Thus, in surfacing the bilateral control style, we have a
counterintuitive finding that extends prior theorizing and allows us to conceptualize situations in
which client and vendor may be viewed simultaneously as controllers.
The Dynamics of Controlling IS Outsourcing and Offshoring Relationships
By analyzing recurrent control balancing decisions in our longitudinal case study of an ISD
offshoring project, we were able to identify and conceptualize three control configurations that
are relevant to this context (see Table 11): authoritative control (procedural; tight; unilateral),
coordinated control (hybrid; tight; bilateral), and trust-based control (social; relaxed; bilateral).
In our particular case study, the exchange parties started with coordinated control, which rests
upon viewing each other as partners who need to closely coordinate work activities. This results
in a hybrid control portfolio with simultaneous emphasis on procedural efficiency and
effectiveness as well as the development of shared understanding, a relatively tight control
degree, and a bilateral control style. While good working relationships were initially developed
through this control configuration, contextual challenges related to shared understanding (e.g.,
business-functional knowledge gaps, and differences in communication and work practices)
created problems that are documented in prior IS offshoring research (e.g., Dibbern et al. 2008;
Levina and Vaast 2008). These challenges ultimately led to an unsatisfying vendor performance
and unfulfilled mutual expectations, which has also been observed in past IS research (e.g.,
Sabherwal 1999). Choudhury and Sabherwal (2003) found that under such circumstances
changes are made to the control portfolio. We extend their findings by explaining that this can
trigger the use of authoritative control. This control configuration rests upon the traditional
client-vendor relationship perspective in which control selection and use is driven by the client.
Our findings suggest that the use of authoritative control may improve project efficiency, but at
35
the cost of a potential deterioration in client-vendor shared understanding. These findings are
related to the known challenge of vicious cycles between control and relationship quality in interorganizational contexts (Das and Teng 1998). In our case study, the identification of gaps in
shared understanding by key client and vendor project leaders and respective actions to address
them eventually triggered the repeated use of coordinated control (with occasional setbacks to
authoritative control), with an even tighter control degree than before and enabled by
accumulated shared experiences and learning. While the role of shared learning and
understanding development between client and vendor is recognized in the IS outsourcing
literature (e.g., Vlaar et al. 2008), its role in facilitating control balancing had not been previously
explored. In our case study, vendor performance improved gradually, mutual expectations were
fulfilled, and the use of trust-based control was triggered (with occasional setbacks to
coordinated control). This led to a situation in which client and vendor employees started to view
themselves as part of the same team, thus fostering further establishment of shared understanding.
Several important insights emerge from our detailed longitudinal analysis of control balancing.
First, contextual factors such as vendor (non-)performance and (un)fulfilled expectations not only
trigger changes to the control portfolio as suggested in prior research (Choudhury and Sabherwal
2003) but also may lead to significant changes in control degree and style. Thus, one important
contribution of our work is to refocus the discussion from constructing portfolios of control to the
idea of control balancing and to recognize that in this broader, more dynamic view there are
multiple dimensions that must be considered (i.e., control types, control degree, and control style)
and that these dimensions can be combined into specific control configurations. Second, we find
strong evidence for the reciprocal effects of control balancing decisions and shared understanding
development in an IS outsourcing relationship. In other words, control behavior is not only
triggered by project- and relationship-related events as suggested in past research, but control
36
behavior itself has a significant impact on the evolution of the project and relationship. These
reciprocal effects offer new insight into the dynamics of control balancing and add to our
understanding of the impacts of organizational control behavior. Specifically, we provide
evidence for multiple iterations and vicious/virtuous cycles in the process of control balancing
based on a dynamic interaction between control balancing decisions and the evolution of clientvendor shared understanding.
Guidelines for Managing ISD Offshoring Projects and Relationships
Based on the theoretical findings discussed, summarized, and integrated above, we derive the
following guidelines for managing ISD offshoring projects and relationships (Table 12).
Guideline
Challenge
Recommendation
#1) Make conscious
control balancing
decisions
Project and relationship challenges (e.g.,
performance is bad, expectations are not fulfilled)
oftentimes lead to ad-hoc control change
responses (e.g., quickly increase the degree of
control in the belief that more control will lead to
better outcomes) which may not yield the
expected benefits (e.g., shared understanding
deterioration, leading to a vicious cycle).
We recommend that client and vendor project
managers recurrently assess control balance
and ask the question: ‘How good is my control
configuration in terms of the three dimensions,
i.e., control types, degree, and style, given the
current situational requirements for achieving
both project-related and relationship-related
goals and objectives?’ Ideally, such ‘reflectionin-action’ leads to conscious control balancing
decisions that proactively tackle project and
relationship challenges.
#2) Leverage the
advantages of a
bilateral control style
Usually, the client in an outsourcing relationship is
viewed as the ‘controller’ and the vendor is
viewed as the ‘controllee’. This leads to a
unilateral control style which in combination with a
tight control degree and relatively procedural
control portfolio (i.e., authoritative control) can be
detrimental for developing shared understanding
in the relationship, an essential success factor in
offshore outsourcing.
We recommend that client and vendor project
managers sit together and think about control
jointly, thereby compensating for the natural
drift of (hierarchical, bureaucratic) organizations
toward efficiency, traditional project success
criteria, and authoritative control. The
implication is that project managers should try
to move as early as possible toward the right in
our process model (see Figure 1) in which
control configurations are characterized, among
other things, by a bilateral control style.
#3) Balance project
and relationship
requirements
Strategies for project management (e.g., status
reporting and tracking) sometimes may be at
odds with strategies for relationship management
(e.g., development of shared understanding),
producing tensions and undesirable reactions
(e.g., the vendor asking the question why s/he is
being treated as a subordinate instead of a
partner).
We recommend that client and vendor project
managers establish and maintain the delicate
balance between achieving short-term goals of
project success and achieving long-term goals
of relationship development which are equally
needed for ISD offshoring success.
#4) Develop shared
understanding in the
Control balancing decisions in ISD offshoring are
driven by shared understanding in the
We recommend that client and vendor project
managers view critical offshoring issues such
37
relationship
relationship and the sooner such understanding is
created, the better. However, critical offshoring
issues (e.g., business-functional knowledge) are
frequently viewed in outsourcing from a one-sided
perspective (e.g., knowledge transfer / absorptive
capacity), which may limit shared understanding
development.
as project processes, business-functional
knowledge, mutual expectations, and social
practices from a shared understanding
perspective. For example, client and vendor
need to accumulate joint knowledge that is
embedded in the relationship; both client and
vendor need to learn and adapt to one another
to develop negotiated social practices.
Table 12: Practical Guidelines for Control Balancing in ISD Offshoring Projects
Limitations and Future Research
Our study is naturally not without limitations. While a single case study design affords the
opportunity to generalize to a theory (Lee and Baskerville 2003, p. 232), which we have strived
to achieve here, further research is needed to replicate our findings across different contexts
(Eisenhardt 1989; Eisenhardt and Graebner 2007; Yin 2003). Additional research is also needed
to determine what other control configurations, besides the three that we identified, may be
feasible and under what circumstances. Another direction for future research would be to explore
whether individual control modes and mechanisms vary on the dimensions of control degree and
style within the context of a particular control configuration. We suspect that for any given
control configuration, the individual control modes and mechanisms will tend to have a high level
of correlation with respect to control degree and style, but additional research is needed in order
to confirm this. Another avenue for future research would be to explore whether control
balancing is contingent upon the choice of systems development methodology, which in our case
was a waterfall ISD approach and might have affected the dynamics of control balancing due to
the associated structuring of the project and manner in which personnel were deployed.
Furthermore, we did not examine the role of technologies in promoting or inhibiting the
balancing of control, e.g., by facilitating the enforcement of controls. Thus, future research
should examine more explicitly the interactions between the use of information and
communication technologies and control balancing. Another limitation of our study is that we
38
focused on examining inter-organizational control transmission, while in reality, the use of interorganizational (e.g., client-vendor) controls could potentially interact with the use of intraorganizational controls (e.g., vendors controlling their offshore teams). This, too, deserves further
research attention. Finally, our theory is also constrained to some extent by temporal restrictions.
In particular, researchers applying and extending our findings to new contexts need to consider
that our examination of control balancing is based on the early phases of an IS offshoring
relationship. In our particular case, the client and the vendor cooperated for the first time in a
joint multi-year project and prior to this project had only cooperated by filling individual resource
gaps on a one-off basis. Thus, they shared only a minor relationship history. The phenomenon of
control balancing might look differently during later and more mature stages of such a
relationship.
Finally, we particularly call for more research on the process of control balancing. We hope that
our study will trigger more empirical research on control balancing in related contexts, including
domestic IS outsourcing projects and internal IS projects, in which different stakeholders may
also need to explore various control configurations and adapt them over time as a project unfolds.
Such exploration may provide insights into how controllers can acquire the skills needed to
enable effective control balancing across a variety of project contexts, thus improving the
chances of achieving successful project outcomes.
CONCLUSION
Our study elucidates the dynamic, reciprocal, and multifaceted nature of control balancing in ISD
offshoring projects. This perspective challenges the assumptions and addresses some of the
shortcomings of prior research, namely the one-dimensional focus on types of control in the
construction of control portfolios and the unidirectional focus on triggers affecting changes in
39
control portfolios. In summary, our grounded theory of control balancing in ISD offshoring
projects holds the potential to trigger new ways of looking at control phenomena in temporary
inter-firm organizations such as client-vendor ISD offshoring projects.
ACKNOWLEDGEMENTS
This work was developed as part of a research project of the E-Finance Lab at Goethe University
Frankfurt. Any opinions, findings, conclusions, or recommendations expressed in this paper are
those of the authors and do not necessarily reflect the views of the E-Finance Lab or its
supporting partners. The authors are indebted to the participating universities and gratefully
acknowledge the financial support of their industry partners. In addition, we would like to thank
Michael Prifling who has supported the data collection. Finally, we acknowledge the truly
developmental review process that this paper has gone through under the guidance of the SE and
AE and we appreciate the contribution of the entire review team in helping us to improve upon
the ideas presented in the paper.
REFERENCES
Bacharach, S.B. 1989. "Organizational Theories: Some Criteria for Evaluation," The Academy of Management
Review (14:4), pp. 496-515.
Barker, J.R. 1993. "Tightening the Iron Cage: Concertive Control in Self-Managing Teams," Administrative Science
Quarterly (38:3), pp. 408-437.
Bradach, J.L., and Eccles, R.G. 1989. "Price, Authority, and Trust: From Ideal Types to Plural Forms," Annual
Review of Sociology (15:1), pp. 97-118.
Bryant, A., and Charmaz, K. 2007. The Sage Handbook of Grounded Theory, Thousand Oaks: SAGE Publications.
Cardinal, L.B., Sitkin, S.B., and Long, C.P. 2004. "Balancing and Rebalancing in the Creation and Evolution of
Organizational Control," Organization Science (15:4), pp. 411-431.
Charmaz, K. 2006. Constructing Grounded Theory: A Practical Guide through Qualitative Analysis
Choudhury, V., and Sabherwal, R. 2003. "Portfolios of Control in Outsourced Software Development Projects,"
Information Systems Research (14:3), pp. 291-314.
Das, T.K., and Teng, B.-S. 1998. "Between Trust and Control: Developing Confidence in Partner Cooperation in
Alliances," Academy of Management Review (23:3), pp. 491-512.
Dibbern, J., Winkler, J., and Heinzl, A. 2008. "Explaining Variations in Client Extra Costs between Software
Projects Offshored to India," MIS Quarterly (32:2), pp. 333-366.
Eisenhardt, K.M. 1989. "Building Theories from Case Study Research," Academy of Management Review (14:4), pp.
532-550.
Eisenhardt, K.M., Furr, N.R., and Bingham, C.B. 2010. "Microfoundations of Performance: Balancing Efficiency
and Flexibility in Dynamic Environments," Organization Science (21:6), pp. 1263-1273.
Eisenhardt, K.M., and Graebner, M.E. 2007. "Theory Building from Cases: Opportunities and Challenges," Academy
of Management Journal (50:1), pp. 25-32.
40
Ghemawat, P. 2007. "Managing Differences," Harvard Business Review (85:3), pp. 58-68.
Glaser, B.G. 1978. Theoretical Sensitivity, Mill Valley: The Sociology Press.
Glaser, B.G. 1992. Emergence Vs. Forcing: Basics of Grounded Theory Analysis, Mill Valley, CA: Sociology Press.
Glaser, B.G., and Strauss, A. 1967. The Discovery of Grounded Theory: Strategies for Qualitative Research,
Chicago: Aldine Publishing Company.
Gopal, A., and Gosain, S. 2010. "The Role of Organizational Controls and Boundary Spanning in Software
Development Outsourcing: Implications for Project Performance," Information Systems Research (21:4),
pp. 960-982.
Heiskanen, A., Newman, M., and Eklin, M. 2008. "Control, Trust, Power, and the Dynamics of Information System
Outsourcing Relationships: A Process Study of Contractual Software Development," Journal of Strategic
Information Systems (17:4), pp. 268-286.
Henderson, J.C., and Lee, S. 1992. "Managing I/S Design Teams: A Control Theories Perspective," Management
Science (38:6), pp. 757-777.
Kelle, U. 2007. "The Development of Categories: Different Approaches in Grounded Theory," in The Sage
Handbook of Grounded Theory, A. Bryant and K. Charmaz (eds.). London: SAGE Publications, pp. 191213.
King, W.R., and Torkzadeh, G. 2008. "Information Systems Offshoring: Research Status and Issues," Management
Information Systems Quarterly (32:2), pp. 205-225.
Kirsch, L.J. 1996. "The Management of Complex Tasks in Organizations: Controlling the Systems Development
Process," Organization Science (7:1), pp. 1-21.
Kirsch, L.J. 1997. "Portfolios of Control Modes and Is Project Management," Information Systems Research (8:3),
pp. 215-239.
Kirsch, L.J. 2004. "Deploying Common Systems Globally: The Dynamics of Control," Information Systems
Research (15:4), pp. 374-395.
Koh, C., Ang, S., and Straub, D. 2004. "It Outsourcing Success: A Psychological Contract Perspective," Information
Systems Research (15:4), pp. 356-373.
Lacity, M., and Willcocks, L. 2012. "Outsourcing Business and I.T. Services: The Evidence of Success, Robust
Practices and Contractual Challenges," Legal Information Management (12:01), pp. 2-8.
Langley, A. 1999. "Strategies for Theorizing from Process Data," Academy of Management Review (24:4), pp. 691710.
Lee, A.S., and Baskerville, R.L. 2003. "Generalizing Generalizability in Information Systems Research,"
Information Systems Research (14:3), pp. 221-243.
Lempert, L.B. 2007. "Asking Questions of the Data: Memo Writing in the Grounded Theory Tradition," in The Sage
Handbook of Grounded Theory, A. Bryant and K. Charmaz (eds.). London: SAGE Publications, pp. 245264.
Levina, N., and Vaast, E. 2008. "Innovating or Doing as Told? Status Differences and Overlapping Boundaries in
Offshore Collaboration," MIS Quarterly (32:2), pp. 307-332.
Long, C.P., and Sitkin, S.B. 2006. "Trust in the Balance: How Managers Integrate Trust-Building and Task Control,"
in Handbook of Trust Research, R. Bachmann and A. Zaheer (eds.). Cheltenham: Edward Elgar, pp. 87–
106.
Mähring, M., and Keil, M. 2008. "Information Technology Project Escalation: A Process Model," Decision Sciences
(39:2), pp. 239-272.
Maruping, L.M., Venkatesh, V., and Agarwal, R. 2009. "A Control Theory Perspective on Agile Methodology Use
and Changing User Requirements," Information Systems Research (20:3), pp. 377-399.
Mingers, J. 2001. "Combining Is Research Methods: Towards a Pluralist Methodology," Information Systems
Research (12:3), pp. 240-259.
Mingers, J. 2004. "Real-Izing Information Systems: Critical Realism as an Underpinning Philosophy for Information
Systems," Information and Organization (14:2), pp. 87-103.
Mohr, L.B. 1982. Explaining Organizational Behavior, San Francisco: Jossey-Bass.
Ouchi, W.G. 1977. "The Relationship between Organizational Structure and Organizational Control," Administrative
Science Quarterly (22:1), pp. 95-113.
Ouchi, W.G. 1979. "A Conceptual Framework for the Design of Organizational Control Mechanisms," Management
Science (25:9), pp. 833-848.
Ouchi, W.G., and Maguire, M.A. 1975. "Organizational Control: Two Functions," Administrative Science Quarterly
(20:4), pp. 559-569.
Pan, S.L., and Tan, B. 2011. "Demystifying Case Research: A Structured-Pragmatic-Situational (Sps) Approach to
Conducting Case Studies," Information and Organization (21:3), pp. 161-176.
41
Rai, A., Maruping, L.M., and Venkatesh, V. 2009. "Offshore Information Systems Project Success: The Role of
Social Embeddedness and Cultural Characteristics," MIS Quarterly (33:3), pp. 617-641.
Rustagi, S., King, W.R., and Kirsch, L.J. 2008. "Predictors of Formal Control Usage in It Outsourcing Partnerships,"
Information Systems Research (19:2), pp. 126-143.
Sabherwal, R. 1999. "The Role of Trust in Outsourced Is Development Projects," Communications of the ACM
(42:2), pp. 80-86.
Sarker, S., and Sarker, S. 2009. "Exploring Agility in Distributed Information Systems Development Teams: An
Interpretive Study in an Offshoring Context," Information Systems Research (20:3), pp. 440-461.
Sitkin, S.B., Cardinal, L.B., and Bijlsma-Frankema, K. 2010. Organizational Control, New York: Cambridge
University Press.
Smith, W.K., and Lewis, M.W. 2011. "Toward a Theory of Paradox: A Dynamic Equilibrium Model of Organizing,"
Academy of Management Review (36:2), pp. 381-403.
Sundaramurthy, C., and Lewis, M. 2003. "Control and Collaboration: Paradoxes of Governance," The Academy of
Management Review (28:3), pp. 397-415.
Tiwana, A. 2008. "Does Technological Modularity Substitute for Control? A Study of Alliance Performance in
Software Outsourcing," Strategic Management Journal (29:7), pp. 769-780.
Tiwana, A. 2010. "Systems Development Ambidexterity: Explaining the Complementary and Substitutive Roles of
Formal and Informal Controls," Journal of Management Information Systems (27:2), pp. 87-126.
Tiwana, A., and Keil, M. 2007. "Does Peripheral Knowledge Complement Control? An Empirical Test in
Technology Outsourcing Alliances," Strategic Management Journal (28:6), pp. 623-634.
Tiwana, A., and Keil, M. 2009. "Control in Internal and Outsourced Software Projects," Journal of Management
Information Systems (26:3), pp. 9-44.
Urquhart, C. 2001. "An Encounter with Grounded Theory: Tackling the Practical and Philosophical Issues," in
Qualitative Research in Is: Issues and Trends, E. Trauth (ed.). Idea Group Publishing, pp. 104-140.
Urquhart, C. 2007. "The Evolving Nature of Grounded Theory Method: The Case of the Information Systems
Discipline," in The Sage Handbook of Grounded Theory, A. Bryant and K. Charmaz (eds.). London: SAGE
Publications, pp. 339-359.
Urquhart, C., Lehmann, H., and Myers, M. 2010. "Putting the ‘Theory’ Back into Grounded Theory: Guidelines for
Grounded Theory Studies in Information Systems," Information Systems Journal (20:4), pp. 357–381.
Van de Ven, A.H. 2007. Engaged Scholarship: A Guide for Organizational and Social Research, New York: Oxford
University Press Inc.
Vlaar, P.W.L., van Fenema, P.C., and Tiwari, V. 2008. "Cocreating Understanding and Value in Distributed Work:
How Members of Onsite and Offshore Vendor Teams Give, Make, Demand, and Break Sense," MIS
Quarterly (32:2), pp. 227-255.
Walsham, G., and Sahay, S. 1999. "Gis for District-Level Administration in India: Problems and Opportunities," MIS
Quarterly (23:1), pp. 39-65.
Wynn, D.J., and Williams, C.K. 2012. "Principles for Conducting Critical Realist Case Study Research in
Information Systems," MIS Quarterly (36:3), pp. 787-810.
Yin, R. 2003. Case Study Research - Design and Methods, Thousand Oaks, California, USA: Sage Publications.
42
Appendix A: Research Process
Table A1 depicts our research process, including the main steps as well as associated tasks and
outcomes. While the table depicts a linear process, we adhered to grounded theory practice and
were highly iterative in our approach. For example, the process analysis was conducted in
parallel with substantive and theoretical coding. Furthermore, data collection and analysis tasks
were highly intertwined with each other and each influenced the other over time. Therefore, we
do not clearly differentiate between data collection and data analysis tasks.
Research
Step
Tasks
Outcomes
Problem
formulation
 Establishing the phenomenon in terms of its
practical relevance as a prerequisite to produce
grounded theory that has ‘grab’ (Glaser and
Strauss 1967)
 State what the problem is from a practice and
theory perspective and why it is important (Van de
Ven 2007)
 Screening prior research to identify gaps in the
literature (Urquhart 2007)
Single case
study design
 Establishing engaged relationship with
practitioners and negotiating access to data (Pan
and Tan 2011; Van de Ven 2007)
 Selecting a case study site and motivating the
rationale for conducting a single case study, e.g.,
the main criterion for a revelatory case is “when
an investigator has an opportunity to observe and
analyze a phenomenon previously inaccessible to
scientific investigation” (Yin 2003, p. 42)
Open coding
& data
collection
 Gathering rich primary and secondary data,
including intensive interviewing (Charmaz 2006)
 Coding the data and understanding what it is
about by going through interview transcripts line
by line, assigning conceptual labels to data
segments, and identifying core categories (Glaser
1978)
 Adhering to the principle of emergence of
grounded theory: categories should emerge from
the data in the sense that they must ‘fit’ (they must
be readily, not forcibly, applicable to and indicated
by the data under study) and ‘work’ (they must be
meaningfully relevant to and be able to explain the
behavior under study) (Glaser and Strauss 1967)
 Triangulating and comparing different slices of
data to find similarities and differences (Charmaz
2006)
 Delimiting further coding to only those concepts
and variables that relate to the emerged
categories (Glaser 1978)
 Making constant comparisons between instances
of data labeled as a particular category and other
instances of data in the same category to
substantiate categories (Urquhart et al. 2010)
 Further data collection guided by the principle of
theoretical sampling, i.e., deciding on analytic
 Identified the management of ISD offshoring
projects as a practically relevant problem that
many managers are struggling with
 Problem identified as developing a relationship
with an offshore vendor and managing an ISD
offshoring project which is important due to the
rapid growth of ISD offshoring and the associated
well-known challenges (King and Torkzadeh 2008)
 Identified gaps in the literature on controlling ISD
offshoring projects
 Reached an agreement with a banking company
to conduct a longitudinal multi-year case study of
an ISD offshoring project and to obtain data from
both the client and the vendor
 Selected a ‘revelatory case’: longitudinal analysis
of an ISD offshoring project including both client
and vendor perspectives, which has been
inaccessible to scientific investigation before
(Choudhury and Sabherwal 2003, p. 313; Dibbern
et al. 2008, p. 359)
 Conducted multiple interviews, observed
meetings, and obtained project documentation
 Generated more than 300 initial codes and more
than 300 pages of notes and analytical memos
 Identified categories related to the simultaneous
use of multiple controls throughout the ISD
offshoring project without forcing existing concepts
from the literature onto the data. Insured that the
categories were relevant the understanding of
control in offshore ISD projects
 Compared multiple perspectives, including client
and vendor’s, and compared multiple sources of
data
Selective
coding & data
collection
 Delimited further coding to a set of tentative core
categories which evolved into control types,
control degree, and control style
 Followed the constant comparisons technique of
grounded theory research, focusing on the
development of categories and concepts by
constantly comparing data to data (e.g., primary
interview data to secondary data such as project
documentation, data from Germany to data from
43
grounds where to sample from next (Glaser and
Strauss 1967, p. 45)

Process
analysis &
data collection
Theoretical
coding & data
collection
Scaling up
Theoretical
integration
 Mapping of key events (e.g., decisions, actions,
outcomes) against a timeline with a focus on
depicting the exact sequencing of events with
respect to control behavior and decisions in the
case study (Mähring and Keil 2008)
 Delineation of states / phases from triggering
mechanisms (Langley 1999; Van de Ven 2007)

 Analysis and specification of theoretical
relationships between core concepts and
categories (Bryant and Charmaz 2007, p. 25).
This theoretical coding (Glaser 1978), also
referred to as iterative conceptualization (Urquhart
et al. 2010), is aimed at increasing the level of
abstraction, relating categories to each other, and
clarifying which categories may be properties of
others
 Engaging with other theories for theory building:
To raise the level of conceptualization and scale
up the emerging theory, existing theories or
concepts should be used for comparisons
(Urquhart 2007). Thereby, meta theories and
theoretical categories with limited empirical
content and general scope are particularly suitable
as heuristic or sensitizing devices (Kelle 2007)
 Grouping higher level categories into broader
themes with the goal of increasing the
generalizability of the theory and being able to
relate the theory to the broader literature
(Urquhart et al. 2010)
 Relating the theory to other theories in the same
or similar field by comparing the substantive
theory generated with other, previously developed
theories (Glaser 1978; Urquhart et al. 2010)

Table A1: Research steps, tasks, and outcomes


India, data from early data collection to later data
collection) as well as data to theoretical concepts
outside the domain of study
Collected and analyzed additional data as needed
to develop our emerging theory. As an example,
when our theorizing became focused on control
types, degrees, and styles, we decided to conduct
additional interviews to flesh out their properties in
more detail
Constructed a detailed case narrative and process
model describing sequences of events and
evolution of the project and the relationship over
time
Identified three different control configurations
which form the phases of our process model
Identified triggers and consequences of control
balancing decisions which evolved into four types
of shared understanding in the client-vendor
relationship
Analyzed relationships between the tentative
categories control type, degree, and style.
Conceptualized control types, degree, and style as
three different dimensions of control, each with
associated properties and ranges
 Engaged with literature on balancing in
organizational contexts (e.g., Bradach and Eccles
1989)
 Conceptualized control balancing as a higher level
category. Defined it as making adjustments to the
control configuration periodically in terms of control
types, control degree, and control style, to allow
the ISD offshoring project and relationship to
progress
 Conceptualized three different combinations of
control types, degree, and style (i.e., control
configurations): authoritative control, coordinated
control, trust-based control
 Compared our core category of control balancing
with the notion of ‘a portfolio of control modes’
(Kirsch 1997), the literature on control dynamics
(Choudhury and Sabherwal 2003; Kirsch 2004),
and the literature on control evolution (Cardinal et
al. 2004)
 Compared our findings on controlling ISD
offshoring projects with prior literature on
managing and controlling ISD offshoring projects
and relationships (e.g., Choudhury and Sabherwal
2003; Dibbern et al. 2008; Levina and Vaast 2008;
Rai et al. 2009; Sabherwal 1999)