ASYMMETRICAL JOINT ACTION EXPECTATIONS AND PRODUCT
INNOVATION PERFORMANCE IN THE SUPPLY CHAIN
By
Hugo A. DeCampos

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Business Administration - Operations and Sourcing Management – Doctor of
Philosophy
2014

ABSTRACT
ASYMMETRICAL JOINT ACTION EXPECTATIONS AND
PRODUCT INNOVATION PERFORMANCE IN THE SUPPLY
CHAIN
By
Hugo A. DeCampos
Using three essays, this dissertation establishes a case for asymmetrical joint action
expectations (AJAE) as a valid phenomenon in interfirm relations and as relevant to innovation
performance in the supply chain. Essay one lays the theoretical foundation of understanding
how differing marginal cost and benefit curves associated with joint action between two firms
can lead to differing optimal levels and therefore differing expectations regarding the desired
level of joint action across those firms. Essay two investigates this phenomenon in the context of
six case studies of interfirm new product development projects and finds that indeed, gaps
between actual and desired levels of joint action do exist and are relevant to innovation
performance. Further, the case studies reveal that not only is the size of the gap in AJAE
relevant, but so also is the clarity of those expectations. Essay three places AJAE within the
context of supply chain interoperability and tests the hypothesized model using an empirical
secondary data set of R&D powertrain projects in the automotive industry. Findings support
hypothesized relationships between AJAE, behavioral interoperability and innovation
performance.

Copyright by
HUGO A. DECAMPOS
2014

Dedicated to my beautiful wife, Kirsten Reynolds DeCampos

iv

ACKNOWLEDGEMENTS

This dissertation is the product of not only my own efforts, but also of those that have stood
beside me along the way and have given of themselves to help me along this journey.
I first want to acknowledge Dr. Steven A. Melnyk, my advisor and dissertation chair, for his
intellectual contributions to this dissertation and for the academic mentoring provided to me
throughout my years in the doctoral program. Dr. Melnyk’s ability to think out of the box and
engage in interesting and impactful research has stretched me and helped me grow as a
researcher. His guidance allowed me to pursue my own line of inquiry while his probing
questions helped steer me away from pitfalls and towards a more rigorous and impactful
outcome. I greatly appreciate his care and interest for not only me but for my family as well.
Dr. Melnyk is in the truest sense both a gentleman and a scholar and I look forward to his
continued association throughout my career.
I am also grateful for the generosity of the other committee members. Dr. John E. Ettlie, my
dissertation co-chair, believed in me from day one and despite being at a different university,
took an interest in me as if he were at Michigan State University. Dr. Ettlie has given much,
including helping to fund my attendance at a conference and taking time to travel and visit with
me in Michigan. I am especially grateful that Dr. Ettlie made available the secondary data set
used for essay three of this dissertation.
I am also grateful to committee member Dr. Stanley E. Fawcett, who I first met in China
while I was living there on an overseas assignment. It was Dr. Fawcett who nudged me towards
a career in academia and helped mentor me through the quantum career change from a General
Motors purchasing director to a Michigan State doctoral student. His guidance in my

v

dissertation, especially in the case-study portion, was invaluable. His and his family’s concern
and support for my family as we progressed through the Ph.D program is appreciated beyond
measure. The positive impact of his interaction with the DeCampos family will be felt for
generations to come.
Dr. Roger Calantone has been more than generous with his contribution to my development
as a researcher. I appreciate the methodological insights he has provided to me both in the
classroom and also in his office during one-on-one visits. I am also grateful for contributions Dr.
Gary Ragatz has provided on buyer-supplier relationships and on my case study methodology.
I appreciate the support that Dr. David Closs, chair of the department of supply chain
management at MSU, has provided to me both financially and academically. Dr. Closs has
helped organize the funding and infrastructure needed to have a successful Ph.D. program at
MSU – for this I am grateful. Regarding the Ph.D. program, I am also grateful for my fellow
student colleagues in the supply chain and marketing departments with whom I have spent
countless hours in classrooms, offices and restaurants sharing in this journey. A special thanks
goes out to my two office mates I had over the years whom I consider life-long friends: Piyas
Bandyopadhyay and Robert Wiedmer.
In addition to the doctoral program funding provided by the MSU Eli Broad College of
Business, and the Department of Supply Chain Management, I gratefully acknowledge the
following sources of funding that contributed to completion of my dissertation:
•

Robert P. Poland Doctoral Dissertation Fellowship

•

George and Marion Plossl Dissertation Fellowship (runner up) - APICS E&R Foundation

•

NSF Grant No. 0725056 – helped fund the secondary data set used for essay three

vi

•

Dr. Steven A. Melnyk donation of personal frequent flier miles to help fund my travel for
data collection

I would be remiss to not mention those I hold dear to my heart in my personal life who
sacrificed on my behalf, took an interest in my doctoral studies and provided me encouragement
and advice along the way. My brothers Rob and Daniel have both been some of my greatest
supporters since childhood and have always believed in me and in my potential. Even though we
live far apart I still benefit from their continued friendship. I am also grateful to David Hall and
Ben Palmer who are like brothers to me and who have been my loudest cheerleaders outside of
my immediate family. I am grateful for my angel mother, Gloria, who I know prays and worries
over me as much as any mother ever has. She is all the mother a son could ever wish for. My
father, Hugo, continues to be my biggest hero in life. His example of the pursuit of knowledge
and education, love of God, kindness to all and patience in affliction fuels my passion for life
and for continuous improvement. I am honored to carry his name.
Finally, and most of all, I am grateful for my wife and children who have sacrificed much so
that I could pursue this dream. Hannah, my oldest child, helps keep me sane by laughing at all
my jokes and puns and by making beautiful music on her cello. Leila, my virtuoso violinist, is as
precious as the day she was born and never fails to give me the hugs every dad needs while
writing a dissertation. Little Hugo, my first born son, is my pride and joy and helps bring me
back to reality by asking me to drop what I am doing to play and spend time with him and to
always tuck him in bed. Natalie, born while we lived in Shanghai China, would probably have
written my dissertation in half the time it took me if she were in my shoes. Her focused and
tenacious personality inspires me. Joshua, born the month I started the Ph.D. program, is my
doctoral program timekeeper. Every time we celebrate Joshua’s birthday, I realize that time is

vii

ticking on my degree program. Joshua never fails to have a smile and a hug waiting for me when
I arrive home. Theodore is the greatest blessing to come to my life this past year. Theo’s
miraculous birth, survival and development have been a reminder of the fragile and joyous
nature of the lives we all live. His precious smile these past few months have energized me each
day and helped me get this dissertation over the goal line. Finally, I am eternally grateful for
Kirsten, the most amazing wife a man could have and the best mother our children could have.
Her love, friendship, forgiveness, patience, companionship and encouragement have been
unwavering since the day we met – she is both my best friend and my eternal companion. There
should be a medal of honor bestowed upon any person who can support their spouse through a
dissertation while raising six children! I couldn’t have completed this dissertation without her. It
is no understatement to say that I married “up”. This dissertation is dedicated to you, Kirsten.

viii

TABLE OF CONTENTS

LIST OF TABLES ...................................................................................................................... xii
LIST OF FIGURES ................................................................................................................... xiii
1 ESSAY ONE – ASYMMETRICAL JOINT ACTION EXPECTATIONS AND
INNOVATION PERFORMANCE: A CONCEPTUAL MODEL ........................................... 1
1.1 Introduction ...................................................................................................................... 1
1.2 Innovation and Relationships ......................................................................................... 3
1.2.1 Joint Action ................................................................................................................. 4
1.2.2 Relationships and Interoperability .............................................................................. 6
1.2.3 Asymmetrical Joint Action Expectations.................................................................... 7
1.3 Towards a Theoretical Framework of Asymmetrical Joint Action Expectations ..... 8
1.3.1 Model 1 – Simplifying Assumptions: Shared MB and MC curves .......................... 10
1.3.2 Model 2 – Relaxed Assumptions: Differing MC and MB Curves............................ 12
1.3.3 Model Limitations and Assumptions ........................................................................ 16
1.4 Research Propositions ................................................................................................... 17
1.5 Conclusions and next steps............................................................................................ 24
2 ESSAY TWO – ASYMMETRICAL JOINT ACTION EXPECTATIONS AND
INNOVATION PERFORMANCE: A CASE STUDY ANALYSIS ....................................... 26
2.1 Introduction .................................................................................................................... 26
2.2 Methodology ................................................................................................................... 28
2.2.1 Analytic induction ..................................................................................................... 28
2.2.2 A priori model ........................................................................................................... 29
2.2.3 Multi-Case Sampling ................................................................................................ 31
2.2.4 Interview protocol ..................................................................................................... 36
2.3 Data collection and analysis .......................................................................................... 38
2.4 Within-Case Analysis..................................................................................................... 40
2.4.1 Case A: Petroleum Equipment Corp ........................................................................ 40
2.4.1.1 Idea Conception ................................................................................................. 41
2.4.1.2 Idea and Application Development ................................................................... 41
2.4.1.3 Launch................................................................................................................ 43
2.4.1.4 Performance ....................................................................................................... 47
2.4.1.5 Key points .......................................................................................................... 47
2.4.2 Case B: Armored Vehicles Inc. ............................................................................... 47
2.4.2.1 Idea Conception / Idea Development ................................................................. 48
2.4.2.2 Application Development .................................................................................. 48
2.4.2.3 Launch................................................................................................................ 50
2.4.2.4 Performance ....................................................................................................... 55
2.4.2.5 Key points .......................................................................................................... 55
2.4.3 Case C: Blaze Inc. .................................................................................................... 55
2.4.3.1 Idea Conception ................................................................................................. 56

ix

2.4.3.2 Idea Development .............................................................................................. 57
2.4.3.3 Application Development .................................................................................. 58
2.4.3.4 Launch................................................................................................................ 59
2.4.3.5 Performance ....................................................................................................... 63
2.4.3.6 Key points .......................................................................................................... 63
2.4.4 Case D: Construction Truck Corp. and Gear Box Corp. ......................................... 64
2.4.4.1 Idea Conception ................................................................................................. 64
2.4.4.2 Idea Development .............................................................................................. 65
2.4.4.3 Application Development .................................................................................. 65
2.4.4.4 Launch................................................................................................................ 66
2.4.4.5 Performance ....................................................................................................... 71
2.4.4.6 Key points .......................................................................................................... 71
2.4.5 Case E: Construction Truck Corp. and Composite Manufacturing ......................... 72
2.4.5.1 Idea Conception ................................................................................................. 72
2.4.5.2 Idea Development .............................................................................................. 73
2.4.5.3 Application Development .................................................................................. 73
2.4.5.4 Launch................................................................................................................ 74
2.4.5.5 Performance ....................................................................................................... 79
2.4.5.6 Key points .......................................................................................................... 79
2.4.6 Case F: Electronics Connect and Design Service Corp ........................................... 79
2.4.6.1 Idea Conception ................................................................................................. 80
2.4.6.2 Idea Development .............................................................................................. 80
2.4.6.3 Application Development .................................................................................. 81
2.4.6.4 Launch................................................................................................................ 82
2.4.6.5 Performance ....................................................................................................... 87
2.4.6.6 Key points .......................................................................................................... 87
2.5 Cross-Case Analysis ....................................................................................................... 87
2.5.1 AJAE Gap Size ......................................................................................................... 87
2.5.2 AJAE Gap Clarity ..................................................................................................... 94
2.5.3 Performance .............................................................................................................. 96
2.6 Propositions and Refined Model................................................................................... 99
2.7 Limitations and Conclusions....................................................................................... 102
3 ESSAY THREE - ASYMMETRICAL JOINT ACTION EXPECTATIONS, SUPPLY
CHAIN INTEROPERABILITY AND INNOVATION PERFORMANCE: AN
EMPIRICAL ANALYSIS OF INTERFIRM NPD POWERTRAIN PROJECTS IN THE
AUTOMOTIVE INDUSTRY .................................................................................................. 106
3.1 Introduction .................................................................................................................. 106
3.2 Interoperability ............................................................................................................ 110
3.2.1 Interoperability in military and IT fields of research .............................................. 110
3.2.2 Interoperability in supply chain management ......................................................... 112
3.3 Hypothesized model ..................................................................................................... 116
3.4 Research methodology and data analysis .................................................................. 123
3.4.1 CB-SEM (with bootstrapping) ................................................................................ 124
3.4.2 Secondary data set................................................................................................... 125
3.4.3 Operationalizing the constructs............................................................................... 126
3.5 Results ........................................................................................................................... 130
x

3.5.1 Data properties ........................................................................................................ 130
3.5.2 Measurement model ................................................................................................ 133
3.5.2.1 Reliability......................................................................................................... 137
3.5.2.2 Convergent validity.......................................................................................... 137
3.5.2.3 Discriminant validity ....................................................................................... 137
3.5.2.4 Statistical power analysis – bootstrapping ....................................................... 138
3.5.3 Path model .............................................................................................................. 139
3.6 Discussion...................................................................................................................... 143
3.6.1 Results ..................................................................................................................... 143
3.6.2 Managerial implications.......................................................................................... 147
3.6.3 Limitations .............................................................................................................. 147
3.6.4 Extensions ............................................................................................................... 148
3.7 Conclusion .................................................................................................................... 150
4

Conclusion ........................................................................................................................... 152

APPENDICES ........................................................................................................................... 157
APPENDIX A: CASE STUDY INTERVIEW PROTOCOL ............................................ 158
APPENDIX B: SURVEY MEASURES FROM US POWERTRAIN R&D STUDY ..... 163
APPENDIX C: DATA NORMALITY AND TRANSFORMATION ANALYSIS ........ 165
REFERENCES .......................................................................................................................... 176

xi

LIST OF TABLES

Table 2.1: Case Demographics ..................................................................................................... 35
Table 2.2: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case A .................. 45
Table 2.3: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case B .................. 53
Table 2.4: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case C .................. 61
Table 2.5: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case D .................. 69
Table 2.6: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case E................... 77
Table 2.7: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case F ................... 85
Table 2.8: AJAE Gap Size ............................................................................................................ 91
Table 2.9: AJAE Gap Size, Gap Clarity and Performance ........................................................... 92
Table 2.10: Performance Score Validation ................................................................................... 93
Table 3.1: Supply Chain Interoperability.................................................................................... 114
Table 3.2: Research Streams Potentially Related to Interoperability ......................................... 116
Table 3.3: Variance/Covariance Matrix ...................................................................................... 132
Table 3.4: Key Model Statistics .................................................................................................. 134
Table 3.5: Squared Correlation Matrix - AVE on Diagonal ....................................................... 135
Table 3.6: Hypotheses Overview ................................................................................................ 144

xii

LIST OF FIGURES

Figure 1.1: Net Benefit Curve ......................................................................................................... 9
Figure 1.2: Alpha and Beta with shared MC and MB curves ....................................................... 12
Figure 1.3: Alpha and Beta with differing MC and MB curves ................................................... 14
Figure 1.4: Posited Model ............................................................................................................. 23
Figure 2.1: Posited Model from Essay 1 ....................................................................................... 30
Figure 2.2: Case Sampling ............................................................................................................ 33
Figure 2.3: Blank Assessment Tool .............................................................................................. 37
Figure 2.4: Performance Assessment Survey ............................................................................... 38
Figure 2.5: Actual vs. Desired Joint Action - Case A................................................................... 44
Figure 2.6: Performance - Case A................................................................................................. 46
Figure 2.7: Actual vs. Desired Joint Action - Case B ................................................................... 52
Figure 2.8: Performance - Case B ................................................................................................. 54
Figure 2.9: Actual vs. Desired Joint Action - Case C ................................................................... 60
Figure 2.10: Performance - Case C ............................................................................................... 62
Figure 2.11: Actual vs. Desired Joint Action - Case D................................................................. 68
Figure 2.12: Performance - Case D............................................................................................... 70
Figure 2.13: Actual vs. Desired Joint Action - Case E ................................................................. 76
Figure 2.14: Performance - Case E ............................................................................................... 78
Figure 2.15: Actual vs. Desired Joint Action - Case F ................................................................. 84
Figure 2.16: Performance - Case F ............................................................................................... 86
Figure 2.17: Revised Model ........................................................................................................ 101

xiii

Figure 2.18: Graphical Model of Asymmetrical Joint Action Expectations .............................. 102
Figure 3.1: Hypothesized Model................................................................................................. 122
Figure 3.2: Measurement Model ................................................................................................. 136
Figure 3.3: Path Model Step 1 – Mediation Test ........................................................................ 141
Figure 3.4: Path Model Step 2 –Mediation Test ......................................................................... 142
Figure A.1: Interview protocol………………………………………………………………....158
Figure B.1: R&D study survey protocol………………………………………………………..163
Figure C.1: Normality analysis and data transformation analysis…………………………...…165

xiv

1 ESSAY ONE – ASYMMETRICAL JOINT ACTION
EXPECTATIONS AND INNOVATION PERFORMANCE: A
CONCEPTUAL MODEL
1.1 Introduction
Increasingly, the locus of innovation is shifting upstream from OEMs in the supply chain
towards first tier suppliers (Rothenberg & Ettlie, 2011). For example, as noted by Mosquet,
Russo, Wagner, Zablit, and Arora (2014), since 2008 there has been a 37 percent increase in the
number of patents filed (an indication of innovation) by first tier suppliers in the auto industry, as
compared to a 28 percent increase in patent filings by automotive OEMs. OEM’s seeking to
capitalize on the innovation potential and investments taking place in the supply chain must
therefore view innovation activity differently than they have in the past; namely as a supplychain rather than as a single-firm activity. Critical to this consideration is management of the
relationship between the participating firms in the value chain, specifically between the customer
and the supplier. For the innovation to take place, the relationship must first be established, and
then managed strategically and purposefully.
This change in the locus of innovation has affected not only how firms in these relationships
work with each other, but it also puts into question the relevance of past management practices
employed when the locus of innovation was more heavily weighted towards the OEM. For
OEMs, supplier management strategies that worked in the past are increasingly less effective in
this new environment. For suppliers, rules of customer engagement are being rewritten as
suppliers are increasingly choosing with whom to partner with, especially for those suppliers
seen as innovation leaders (Chew & Whitbread, 2002). In short, the rules of how interfirm
relationships emerge and are managed may not be the same in an environment where the OEM

1

dominates innovation versus the environment where this innovation activity is shared, or even
dominated, by the OEM’s tier-one suppliers.
Past research has sought to predict where and when interfirm relationships are organized by
focusing primarily on economic considerations (Lampel & Giachetti, 2013; Luzzini, Caniato,
Ronchi, & Spina, 2012; McIvor, 2009). For example, transaction cost economics (TCE) draws
on economically based frameworks to predict when a firm should make a product/service inhouse, buy it in the marketplace, or ally with another firm to produce the product/service in
partnership (Williamson, 1979, 2008). Yet, there is strong anecdotal evidence that economic
considerations are not sufficient to explain whether or not such relationships emerge. In some
cases, relationships that should have been established based on strictly economic grounds were
not launched. As an example, consider the failed GM/Renault-Nissan alliance talks of 2006 as
described by Langley, White, and Boudette (2006) and LaReau (2006). At the heart of this
potential strategic alliance was recognition of the strong potential economic benefits that could
be generated. One estimate was that by sharing certain platform developments and purchasing
costs, there were over $10 billion USD in potential synergies to be gleaned. Yet, ultimately, this
relationship failed to even be established because of differences in expectations of how the
benefits would be distributed. Specifically, GM perceived itself as receiving fewer benefits than
Renault-Nissan; consequently, it asked for a significant equalization payment upfront – a request
that was rejected by Renault-Nissan.
While this is an obvious example of an interfirm relationship that should have been
established but was not, it is not unique. There are other examples of failed relationships. It has
been reported that over 50% of all alliances fail (Nidumolu, Ellison, Whalen, & Billman, 2014;
Parise & Casher, 2003). Research is replete with attempts to explain why so many B2B alliances

2

fail (Park & Ungson, 2001; Stuart, 1997). Annual buyer/supplier relationship quality surveys tell
a similar story – that of numerous failures, even with some of the largest and most mature firms
in industry (Zhang, Henke, & Griffith, 2009; Zhang, Viswanathan, & Henke Jr, 2011). It is
evident that economic issues, while important, are not enough to explain success and failure in
innovation-orientated partnerships. Other behavioral issues must be considered – issues such as
trust (Moldoveanu & Baum, 2011; Zaheer & Venkatraman, 1995), top management commitment
(Adobor & McMullen, 2007), and corporate culture (McAfee, Glassman, & Honeycutt Jr, 2002).
In this study, we introduce an additional issue – an issue with its roots in behavioral
economics – that of differences in expectations between the parties involved in the potential
relationship. Implicit in past research has been the assumption that the parties involved in such
relationships approach the relationship with identical expectations or that the expectations (or
differences) are not critical, compared to the economic considerations. Yet, as can be seen from
the previously cited example involving GM and Renault-Nissan, expectations can and do differ
and these differences do affect the resulting ability of the relationship to first launch and then
succeed. We regard these differences in expectations ultimately as part of a broader set of issues
dealing with interoperability.

1.2 Innovation and Relationships
In the context of the shifting locus of innovation, where innovation success resides more and
more across firm boundaries, focus on behavioral variables that define the relationships may be a
more appropriate paradigm for explaining innovation success as opposed to a purely economic
paradigm. While this study does focus on the role of joint action expectations, a behavioral
element of the relationship, it does so by building upon traditional economic models of marginal
cost and marginal benefit analysis commonly used in economic literature (Olson, 1965, pg.24).

3

As such, we integrate both behavioral and economic considerations into a single framework that
we posit is relevant in understanding differential performance for interfirm innovation initiatives
(Hirsch, Friedman, & Koza, 1990).
1.2.1

Joint Action

In interfirm relationships, both firms engage in joint action in order to solve problems related
to the value creation process, be it in operations, strategy formulation or new product
development. Joint action is defined by Heide and John (1990) as “the degree of interpenetration
of organizational boundaries” and by Gulati and Sytch (2007) as “the degree of dyadic
cooperation and coordination across a wide array of organizational activities, such as design, cost
control, and quality improvement” (Gulati & Sytch, 2007, pg.40; Heide & John, 1990, pg.25).
Interpenetration can take place when buyers and their suppliers participate in each other’s dayto-day activities in an effort to jointly improve new product development performance. The
level of joint action firms engage in can extend from very basic exchanges such as phone calls, to
much more complex initiatives such as alliances and joint ventures.
Various other forms of joint action exist that fall between the extremes of this continuum.
Some of these include constructs oft studied in academic research including communication
(Prahinski & Benton, 2004), coordination (Sanders, 2008), cooperation (Kee-hung, 2009),
integration (Frohlich & Westbrook, 2001) and collaboration (Allred, Fawcett, Wallin, &
Magnan, 2011), to name a few. While these constructs have been investigated in supply chain
research, a lack of clarity exists regarding exact definitions and consistent measures of
operationalization that clearly demarcate one construct from the other. Where does the scope of
collaboration begin and end? What overlap and distinction exists between collaboration and
cooperation, for example? Our intent in this paper is not necessarily to delineate these overlaps

4

and distinctions, but to investigate the relevance of higher and lower levels of interfirm joint
action to explain differential new product development project performance. We therefore focus
on the more general construct of joint action in our research, thus capturing the essence of the
wide spectrum of interfirm activities while avoiding the unnecessary confusion in delineating
these constructs from each other.
Engaging in joint action and deciding what level of joint action to engage in, requires firms
to share risk, investments, and benefits. Risk sharing takes place along numerous dimensions,
such as risk of intellectual property leakage (Chopra & Sodhi, 2004), risk of disintermediation
(Mills & Camek, 2004) and risk of conceding internal data that the other party can leverage in
pricing or other rent-sharing negotiations. Further, joint action can also create shared operational
risk be it in a jointly owned manufacturing facility or in the market risk associated with a
particular new product development initiative. While both firms engage in some level of risk,
there is no assumption or requirement that joint action creates risk equally across both firms.
Engaging in joint action also requires some level of investment by both parties. Firms can invest
cash, time, capital equipment or other resources such as human capital and intellectual property.
Similar to risk, investments across both firms for a given level of joint action are not necessarily
equal. The purpose of joint action ultimately is to achieve some level of benefit to each of the
two firms involved in the joint action. Negotiation is the tool used to distribute both the costs
and the rents created through the joint action. The distribution of rents also need not be equal
and in fact is a topic of much research in the interfirm relationship literature.
Many different lenses have been applied to the study of interfirm relationships. The
relational view of the firm (Dyer & Singh, 1998), a subset of the resource-based view of the firm,
and transaction cost economics (Williamson, 2008) are research lenses commonly applied.

5

Other studies, however, have focused on behavioral issues such as trust (Zaheer & Venkatraman,
1995) and culture (McAfee et al., 2002), and how these constructs are relevant in explaining
differential performance in interfirm relationships. These behavioral issues are critical because it
recognizes that relationships are driven and maintained by not only economic considerations but
also by behavioral considerations (Hirsch et al., 1990).
1.2.2

Relationships and Interoperability

When we talk about interfirm relationships, we also must focus on how organizations
develop interfaces between themselves. This focus of building interfaces falls under the category
of interoperability. While numerous definitions for interoperability have been provided, the most
commonly accepted definition views interoperability as: “The ability of systems, units, or forces
to provide services to and accept services from other systems, units, or forces and to use the
services so exchanged to enable them to operate effectively together” (Ford, Colombi, Graham,
& Jacques, 2007, pg.6).
Past research in supply chain management has focused on material, process, and
informational flows where the focus is getting the right product to the right place at the right time
and at the right cost (Fisher, 1997; Lee, 2004). These are the building blocks of interfirm
interoperability. While much research focuses on synchronizing resources and process across
differing entities, less research has been devoted to exploring the behavioral aspects of
interoperability in the supply chain: a focus that is beginning to gain traction under the banner of
behavioral operations (Agarwal, Croson, & Mahoney, 2010; Siemsen, 2011). We posit that joint
action expectations are one of the behavioral variables that need to be considered in establishing
business-to-business interoperability in the supply chain.

6

1.2.3

Asymmetrical Joint Action Expectations

Extant research on the benefits of joint action has typically taken a focal firm approach
(Bercovitz, 2006). As such, an implicit assumption that arises from such research is that the
optimal desired level of interfirm joint action for the focal firm is also “optimal” for the partner
firm. We relax this assumption and explore the implications of doing so. When we treat
partnering firms as distinct actors, there is no guarantee that the two firms, even though they are
interdependent supply chain partners, will share the exact same theoretical optimal level of joint
action. In fact, we would expect that this would rarely be the case given the differing marginal
costs and marginal benefits associated with each firm’s engagement in joint action projects
(Nyaga, Whipple, & Lynch, 2010). For example, it has been shown in the bullwhip effect
research stream that the benefits of sharing point of sales data across a supply chain accrue to
upstream firms more so than to downstream firms (Croson & Donohue, 2003), despite the fact
that it is the downstream firm that must contribute much of the investment in sharing point of
sales data.
As long as the marginal benefits of joint action exceed the marginal costs for a given firm,
then that firm will most likely desire to engage in joint action. The existence of differing
marginal benefits and marginal costs associated with shared joint action renders likely that two
interdependent supply chain partners will in fact differ in their respective theoretical optimal
levels of joint action. By definition, however, both firms can only engage in one level of joint
action at a given point in time. We assume that each firms desires to operate at their respective
theoretical optimal level of joint action, and that these differing desires result in differing
expectations in the relationship. We label these differences as asymmetrical joint action
expectations. While studies have shown that the absolute level of joint action between two firms

7

has a direct impact on product innovation performance in the supply chain, this dissertation
explores the relationship that the asymmetry of joint action expectations has on behavioral
interoperability and on innovation performance in the supply chain.

1.3 Towards a Theoretical Framework of Asymmetrical Joint Action
Expectations
While we have conceptually argued the existence of asymmetrical joint action expectations,
there is further utility in modeling graphically this phenomenon. Doing so establishes a common
analytical framework that future research can build upon in investigating the variables that
contribute to the existence and size of gaps in joint action expectations. As highlighted
previously, we can model the choice of joint action engagement as a trade-off between marginal
costs and marginal benefits. In the figure below, we label the x-axis as the intensity of overall
joint action (J) between two firms for a given project, spanning from low intensity to high
intensity. We make no distinction in this model between the various forms of joint action, but
rather recognize that moving from low levels of overall joint action to higher levels of joint
action may in fact require the addition of different forms of joint action for that project. For
example, two firms engaged in a new product development project may have low levels of joint
action that only include infrequent face-to-face meetings, thus yielding a low J-score on the xaxis. Alternately, the two firms could engage in not only face-to-face meetings, but they could
also establish a resident engineer program where the supplier assigns one of their developers to
assume residence at the OEM’s engineering facility in order to expedite and facilitate
communication regarding the program. Additionally, the two firms could also jointly invest in
software and equipment related to the project where the costs for such investments are shared. In
this case, the resulting J-score would be higher on the scale and would be the result of numerous
forms of joint action initiatives simultaneously pursued.

8

We assume that all costs and benefits associated with a given level of joint action can be
monetarily
netarily expressed. As such, for each level of J, there is a net benefit curve (total benefit
minus total cost) that can be plotted against a yy-axis
axis expressed in $. In our base model, we
initially assume that both firms share the same net benefit curve, an assumption that we will later
relax. As argued in prior literature, we assert a curvilinear relationship where increasing overall
joint action intensity in a relationship initially increases net benefits, but at a decreasing rate such
that there is an optimal level of J where net benefits are maximized (Das, Narasimhan, & Talluri,
2006; Hoegl & Wagner, 2005; Uzzi, 1997; Villena, 2011)
2011).. After this point, the curve slopes
downward where high levels of joint action may expose the “dark side” referred to earlier where
costs exceed benefits, thus moving the curve towards zero. While the overall benefits of joint
action are still positive when surpassing the point of optimality, they decrease to levels lower
than what could be achieved with lower levels of joint action, thus creating an undesired state of
joint action.
Figure 1.1: Net Benefit Curve

We now decompose the net benefit curve into its two components of costs and benefits
where net benefits (NB) are equal to the total benefits (B) minus the total cost (C):

9

ࡺ࡮ ൌ ࡮ െ ࡯

Equation 1

Next we take the first derivative of NB with respect to J in order to define the relationship of
marginal costs and marginal benefits:

ࢊሺࡺ࡮ሻ
ࢊࡶ

ൌ

ࢊ࡮
ࢊࡶ

െ

ࢊ࡯
ࢊࡶ

Equation 2

NB is maximized when its slope equals zero. We therefore set the first derivative of net
benefits equal to zero and solve for the point of optimality:

૙ൌ

ࢊ࡮

ࢊ࡮

ࢊ࡯

ࢊࡶ

ൌ

ࢊࡶ

ࢊࡶ

െ

ࢊ࡯
ࢊࡶ

Equation 3

Equation 4

As a result, net benefits are maximized when the marginal benefits of joint action equal the
marginal costs of the same. The level of joint action where this occurs is defined as J*.
1.3.1

Model 1 – Simplifying Assumptions: Shared MB and MC curves

Marginal costs (MC) incurred along the curve could include incremental hardware and
software investments, personnel investments, time investments, financial capital investments,

10

assumed risks such as IP leakage risk (product, process and strategy plans), risk of dependency
or risk of decreased power balance/pricing power, to name a few examples. The marginal costs
associated with moving from the left-hand side of the joint action spectrum to the right-hand side
generally increase with J. This can be interpreted as stating that a one-unit increase in J at higher
levels of J results in a larger increase in C as compared to a one-unit increase in J at lower levels
of J. In practical terms, the incremental costs for engaging in increased lower levels of joint
action are less than the incremental costs for engaging in increased higher levels of joint action.
If two firms are not engaged in any joint action (J=0), then the incremental cost to begin basic
communications is minimal. On the other hand, if those firms are already engaged in high levels
of joint action, then to make the next incremental step in joint action could require that the firms
establish a formal joint venture business entity, for example. The incremental cost of doing so
could be significantly higher than the incremental costs of establishing basic communications as
highlighted previously.
Mirroring this relationship, the marginal benefits associated with moving from the left-hand
side of the joint action spectrum to the right-hand side generally decrease. In other words, firms
can reap large incremental benefits by engaging in lower-levels of joint action, however those
incremental benefits generally decrease as joint action levels increase. The concept of harvesting
low-hanging fruit applies here. We assume that high levels of joint action intensity approach a
saturation point in terms of marginal benefits. Marginal benefits could include quick notification
of disruptions, lower operating costs, improved market opportunities, reduced waste, etc.
We make another simplifying assumption that both marginal cost and marginal benefit
*

curves are linear. As shown earlier, the optimal level of joint action (J ) for a firm is the point at
which the marginal benefit and marginal cost curves intersect. Left of this point the marginal

11

benefits exceed the marginal costs thus justifying the investment to increase joint action. Right
of this point, the marginal costs exceed the marginal benefits thus discouraging further
investment to increase joint action. As a result, if we assume that both firms, whom we will call
alpha and beta, share identical marginal cost and marginal benefit curves, then we can conclude
that both alpha and beta will search for and desire to operate at the same optimal level of joint
*

action, J .
Figure 1.2: Alpha and Beta with shared MC and MB curves

1.3.2

Model 2 – Relaxed Assumptions: Differing MC and MB Curves

We now relax the assumption of shared marginal cost (MC) and shared marginal benefit
(MB) curves allowing alpha and beta to have their own unique curves. Differing curves will also
introduce unique terms for optimal joint action for each firm. While it is possible for the
differing curves to converge on the same optimal level of joint action for both firms, such an
occurrence would likely be a product of chance rather than strategic intent. We define optimal
joint action for alpha as the point J

*A

where alpha’s marginal benefit curve intersects
inters
its

respective cost curve, where:

12

ࢊ࡮࡭
ࢊࡶ

ൌ

ࢊ࡯࡭

Equation 5

ࢊࡶ

Similarly, the level of optimal joint action for beta is the point J

*B

where beta’s marginal benefit

curve intersects its respective cost curve, where:
ࢊ࡮࡮
ࢊࡶ

ൌ

ࢊ࡯࡮

Equation 6

ࢊࡶ

By nature of interfirm joint action, both parties share the same level of J since joint action is a
two-party endeavor. As a result, both alpha and beta may desire and pursue differing levels of
*

J (alpha pursues J

*A

while beta pursues J

*B

) yet by definition they can only jointly implement

one level of joint action, J.
In the figure below, we represent this very situation where alpha and beta have differing
marginal cost and benefit curves and therefore differing optimal levels of joint action.

13

Figure 1.3:: Alpha and Beta with differing MC and MB curves

In this case, both parties will be likely invest in joint action at a minimum up to point J

*B

.

Beyond this point, however, a series of differing situations may arise. The parties could invest in
joint action at point J

*A,

at point J

*B

or some point between J

dyad will invest in joint action beyond point J
two firms. If the dyad invests at point J

*A

*A

and J

*B

. It is unlikely that the

since this would not be desired by any of the

*B

, then beta will be satisfied, having optimized its

tradeoffs between marginal costs and marginal benefits. Alpha, however, will be left unsatisfied
since opportunity for further gain is forgone. Conversely, if the dyad invests in joint action at
level J

*A

, then alpha will be satisfied, having reached its level of optimality while beta will be in

the undesirable situation where its marginal costs surpass the marginal benefits received. If
I
investment is made at a level exactly midway between J

*A

and J

*B

, then both alpha and beta will

experience a suboptimal state that is equidistant from their respective points of optimality where
beta incurs excess marginal costs while alpha is left in a state where more joint action could be

14

beneficial.
Assuming that firms seek to maximize utility and desire to invest only in optimal levels of
joint action, then we can conclude that the resulting gap that emerges due to differing points of
optimality creates asymmetrical joint action expectations. We can analytically quantify the size
of this asymmetry for firms alpha and beta using the following expression where ∆‫ כܬ‬is defined
as the absolute value of the difference between alpha’s and beta’s respective levels of optimal
joint action.

∆ࡶ‫ כ‬ൌ ห ࡶ‫ ࡭כ‬െ ࡶ‫ ࡮כ‬ห

Equation 7

The model above can serve as a framework in future research concerning the causes and
nature of asymmetrical joint action expectations. For example, one could investigate to what
degree does increased trust influence the variables of the model, and therefore influence joint
action expectations? Does increased mutual trust shift the marginal cost curve of both parties
equally or differently? If so, in what directions do the curves shift and how does this impact the
gap between optimal joint action levels for the two firms? Perhaps increased trust doesn’t shift
the curves, but merely influences the slope of the curves; if so, this could have a different effect
on the size of the resulting asymmetry as compared to a simple shift. What if the increase in
trust is not mutual, but rather focused on one party versus the other?
In a similar manner, other constructs could be applied to the model investigating what
impact, if any, an increase/decrease of that construct has on the marginal cost and benefit curves
of each party, and subsequently how they impact asymmetry in joint action optima. Questions to
be investigated can include, but are not limited to, ‘what constructs contribute to shifts in the
marginal cost and benefit curves’ and ‘what constructs contribute to changes in slope of the
marginal cost and benefit curves’. Further, how can changes in these curves be managed to

15

reduce the asymmetry in joint action optima and ultimately the expectations of joint action by
both firms?
1.3.3

Model Limitations and Assumptions

The model above makes a few important simplifying assumptions. First, marginal cost and
marginal benefit curves are linear. This assumption supports the curvilinear relationship
between joint action and net benefits identified in prior research (Das et al., 2006) and is also an
assumption that has been used to model firm behavior in economics literature (Rosen, 2006).
Second, it is true that firms live in a dynamic environment and as such the concept of an optimal
point of joint action for either or both firms is a concept that can easily be challenged from a
dynamic perspective. We make the assumption, however, that such dynamism takes place over
a long-enough period of time, or is small enough such as to maintain the relevance of the
proposed static model in predicting firm behavior in the context of a single new product
development project.
In discussing the tradeoffs of making simplifying assumptions in economic models, Nobel
laureate Herbert Simon (1979) stated, “…decision makers can satisfice either by finding
optimum solutions for a simplified world, or by finding satisfactory solutions for a more realistic
world. Neither approach, in general, dominates the other, and both have continued to co-exist in
the world of management science” (Simon, 1979, pg.350). Our proposed model and its
simplifying assumptions follow the former category of finding an optimal solution for a
simplified world. Extensions of research into asymmetrical joint action expectations may find
interest in exploring deviations to these simplifying assumptions. One such deviation may be the
existence of an optimal range of joint action rather than an optimal point. In other words,
perhaps there is a tolerated gap in joint action expectations that is equivalent to no gap at all. If

16

such is the case, perhaps there is a “tipping point” after which the gap size becomes a significant
entity in the relationship equation.
Another assumption that this dissertation makes is that gap size in joint action expectations is
an appropriate surrogate for behavioral interoperability. Smaller gap sizes are a manifestation of
higher levels of behavioral interoperability. Conversely, larger gap sizes are a manifestation of
lower levels of behavioral interoperability. In short, two firms that share smaller gaps in joint
action expectations will share increased harmonizing behavioral norms that support
interoperability as compared to two firms that share larger gaps in joint action expectations. As a
result, we assume that gap size in joint action expectations serves as an adequate surrogate for
behavioral interoperability.

1.4 Research Propositions
My primary interest is in investigating what role behavioral interoperability, as represented
by size of gap in joint action expectations, has on innovation performance in a new product
development setting. The propositions outlined in this section build upon this basic inquiry.
Further research may confirm or disconfirm these propositions. In the course of this research, it
is plausible that rival theories may emerge. We may discover that joint action expectation gap
size isn’t the variable of interest, but rather other factor(s) related to gaps may emerge as being
more relevant. Such competing factors that the authors could conceive as being relevant are (to
name a few), 1) clarity of expectations, and therefore clarity of the existing gap, b) rate of
investments that one or the other firm typically makes in closing joint action expectation gaps
with partner firms and c) incentives associated with the project that may precipitate abnormal
attention and investment in the relationship. I begin, however, with a focus on gap size in joint
action expectations.

17

In order for firms to effectively work jointly, a basic level of resource interoperability is
required. Therefore, two firms with low levels of resource interoperability that desire to engage
in joint action will require varied levels of resource investments. This variation in investments
will most likely result in a greater difference in net benefit curves across the two firms for
engaging in joint action. Two firms, however, with high levels of resource interoperability have
the advantage of approaching the relationship with similar or harmonizing resources that allow
those firms to explore joint action from a more similar vantage point, where it is likely that they
will share more similar net benefit curves as compared to the two firms with lower levels of
resource interoperability. Since similar net benefit curves result in smaller gaps in optimal levels
of joint action, I make the following proposition:

PROPOSITION 1: Interfirm resource interoperability is inversely associated with gap size in
joint action expectations (higher levels of resource interoperability correlate with smaller gaps in
joint action expectations)

Process interoperability is achieved when two firms harmonize not only their resources, but
also their processes in order to effectively engage in joint action. In a similar argument to that
made for resource interoperability, we posit that firms that have higher levels of process
interoperability explore joint action from a more similar vantage point as compared to firms that
have lower levels of process interoperability. As a result firms with higher levels of process
interoperability most likely share more similar net benefit curves for joint action as compared to
firms with lower levels of process interoperability. The similarity in net benefit curves will
result in small gaps in optimal levels of joint action. I therefore make the following proposition:

18

PROPOSITION 2: Interfirm process interoperability is inversely associated with gap size in
joint action expectations (higher levels of process interoperability correlate with smaller gaps in
joint action expectations)

Interfirm relationships characterized by smaller gaps in joint action expectations will spend
less time and concern dealing with this asymmetry as compared to firms that have larger gaps in
joint action expectations. As a result, both the ideation and the problem solving processes
required during new product development will perform at higher levels as compared to those
firms that are constantly second guessing their own level of involvement in the project. I
therefore propose:

PROPOSITION 3: Gap size in joint action expectations is inversely associated with innovation
performance (smaller gaps in joint action expectations correlate with higher levels of innovation
performance)
Interfirm relationships characterized by higher levels of resource interoperability will be able to
more quickly and effectively identify opportunities and solutions during a new product
development project as compared to relationships characterized by lower levels of resource
interoperability. Lower resource interoperability may generate the situation where solutions and
opportunities identified by one partner may not be as feasible to execute by the other partner
since the solution or opportunity may be resource dependent. I therefore posit that higher
interfirm resource interoperability will be correlated with higher levels of innovation
performance.

19

PROPOSITION 4a: Interfirm resource interoperability is directly associated with innovation
performance (higher levels of resource interoperability correlate with higher levels of innovation
performance)

Further, resource interoperability, as argued in proposition 1, leads to reduced gaps in joint
action expectations. Ultimately, innovation performance in joint new product development
projects is achieved when firms make the decision to share ideas and invest resources in a timely
manner to solve the problems being addressed by the initiative. As such, it is expected that gap
size in joint action expectations will mediate the relationship between resource interoperability
and innovation performance.
A mediating, rather than a moderating relationship is theoretically asserted in the model.
Resource interoperability alone is not expected to be the key lever that drives innovation
performance. Rather, higher levels of resource interoperability lead to reduced AJAE gap size,
which in turn leads to improved innovation performance. Given the complexities of human
interaction involved in achieving innovation performance, it is anticipated that behavioral
interoperability (as operationalized by reduced AJAE gap size) is the key factor of success and
fully mediates the positive relationship between resource interoperability and innovation
performance. Arguing for a moderating relationship would place behavioral interoperability as a
secondary factor that simply amplifies or attenuates the direct relationship between resource
interoperability and innovation performance. This is clearly not the case since it is humans, not
resources that ultimately generate innovations.

20

PROPOSITION 4b: Gap size in joint action expectations mediates the relationship between
resource interoperability and innovation performance

Interfirm relationships characterized by higher levels of process interoperability will be able
to more quickly and effectively implement solutions during a new product development project
as compared to relationships characterized by lower levels of process interoperability. Lower
process interoperability may generate the situation where solutions and opportunities identified
by one partner may not be as efficiently or effectively implemented by the other partner given
the challenges that arise with coupling differing processes that may work against each other. I
therefore posit that higher interfirm process interoperability will be correlated with higher levels
of innovation performance.

PROPOSITION 5a: Interfirm process interoperability is directly associated with innovation
performance (higher levels of process interoperability correlate with higher levels of innovation
performance)

In a similar manner that AJAE gap size is expected to mediate the relationship between
resource interoperability and innovation performance, we also expect AJAE gap size to mediate
the relationship between process interoperability and innovation performance. The same
argument for mediation rather than moderation holds here as well. Both resource and process
interoperability are proposed to be ‘failure preventers’ while behavioral interoperability (AJAE
gap size) is proposed to be ‘success producers’ (Varadarajan, 1985). In short, the failure

21

preventers, resource and process interoperability are both necessary to start the relationship but
are not sufficient to assure success. The success producer is behavioral interoperability.

PROPOSITION 5b: Gap size in joint action expectations mediates the relationship between
process interoperability and innovation performance

22

Figure 1.4: Posited Model

Prop 5a, 5b

PROCESS
INTEROP

Prop 2

BEHAVIOR
INTEROP

Prop 3

(AJAE GAP SIZE)

Prop 1

RESOURCE
INTEROP

Prop 4a, 4b

23

INNOVATION
PERFORM

1.5 Conclusions and next steps
This paper has both theoretically and analytically argued that firms engaged or attempting to
engage in joint action will most likely differ in the optimal level of joint action for that
relationship. Assuming firms desire to operate at an optimal level of joint action, asymmetrical
joint action expectations will exist in the relationship, albeit the gaps between levels of
optimality may greatly differ across different relationships. This nature of this asymmetry is the
subject of interest in our research, in particular with its relation to innovation performance during
new product development projects. Further, we introduced the construct of interoperability and
extended it to the domain of interfirm relationships. We recognize that behavioral
interoperability is a broad construct impacted by numerous factors, however, our research takes a
focal approach on gaps in joint action expectations as a proxy for behavioral interoperability. In
particular we argue that the size of the gap is relevant to innovation performance.
While some prior research has recognized that firms may differ in expectations of the net
benefits associated with a resource (Barney, 1986), this paper is the first, to our knowledge, to
offer an analytical model of asymmetrical joint action expectations and to investigate more fully
the nature and impact that this gap has on innovation performance. Given the nascent nature of
this research, a qualitative study can both substantiate the key elements of the posited model, and
provide further insights to refine the model and its constructs. The next step of this dissertation
is to conduct this qualitative research in the form of a multi-case study (Eisenhardt, 1989). This
multi-case study will be the subject of essay two and will focus on identifying real-world
examples of gaps in joint action expectations and will seek to answer the following questions: 1)
are these gaps real? 2) if so, are they relevant? 3) does size of the gap in joint action expectations
really matter? 4) how do such gaps emerge? 5) in what ways do they impact performance . The

24

final and third essay in this dissertation will be a larger-scale investigation of the refined model
that results from essay two.

25

2 ESSAY TWO – ASYMMETRICAL JOINT ACTION
EXPECTATIONS AND INNOVATION PERFORMANCE: A
CASE STUDY ANALYSIS
2.1 Introduction
Essay one in this dissertation posited that gap size in joint action expectations has an inverse
relationship with innovation performance: in other words, smaller sized gaps are correlated with
increased innovation performance. Further, we placed the construct of asymmetrical joint action
expectations (AJAE) into the greater context of establishing behavioral interoperabi2lity between
two firms. We developed in essay one a theoretical framework based on past research. The past
research, however, does not adequately deal with the issue of asymmetrical joint action
expectations. Consequently, before proceeding to further empirical evaluation of the framework
and its associated propositions, this current essay seeks to verify and evaluate the existence of
AJAE and improve understanding of its nature and its relationship to interfirm innovation
performance in the supply chain.
Given the nascent nature of research on AJAE, we employ a qualitative research
methodology to help explore this construct and to better refine the a priori theoretical model
posited in essay one (Corbin & Strauss, 2008, pg. 12). The use of detailed case study data allows
us to determine the extent to which the posited theoretical framework can adequately explain
observed results and, if not, will help us identify appropriate refinements to the current
framework so that it provides a more accurate explanation of what is observed in actual complex
business situations (Yin, 2009, pg. 4). We conducted a multi-case study (Eisenhardt, 1989) of
six interfirm product development projects using a purposeful sample that spans both high/low
levels of success and high/low levels of product complexity.

26

Our findings from these case studies confirm the existence of AJAE and suggest that they do
have an impact on interfirm relationships. However, whereas we initially posited that the size of
the gap of AJAE was the key variable of interest, our case studies revealed another, potentially
more critical construct: that of the clarity of the gap. Clarity bespeaks an understanding that one
firm has of not only their own costs and benefits associated with varying levels of joint action,
but also an understanding of the costs and benefits associated with the partner firm’s engagement
in those same levels of joint action and the ensuing joint action expectation gaps that result.
Therefore, when clarity exists, a firm understands not only what level of joint action is beneficial
to their own firm, but also to that of their partner firm. Alternately, a lack of clarity is the result
of a firm only concerned with it’s own costs and benefits associated with joint action and thus
that firm lacks an awareness and appreciation of any asymmetry in joint action expectations that
may exist with its partner firm.
One observation we found interesting was that firms that had an established level of clarity of
AJAE still decided to enter or maintain existing business-to-business relationships with a partner
firm despite the existence of AJAE. Those firms appear willing to live with and accept some
level of asymmetry (gap) in joint action expectations. This is an important consideration for
managers who may be either hesitant to share with a partner firm their own costs and benefits
associated with a desired level of joint action, or may not be interested in what factors contribute
to their partner’s firm desired level of joint action. Further, this study suggests that alignment
between the respective optimal levels of joint action for both firms need not necessarily exist in
order for those two firms to succeed in joint innovation initiatives. On the contrary, it may be
that the existence of AJAE is an impetus for firms to establish clarity and therefore enhance
performance.

27

We next discuss the methodology used to conduct the qualitative research on AJAE. We then
present overviews of each of the six case studies conducted in the context of within-case
analyses. We then follow with a cross-case analysis identifying the themes and findings that
emerged across the six case studies. We lastly present a refined theoretical model that takes into
account our case study findings and present propositions that can be investigated with a future
confirmatory study.

2.2 Methodology
2.2.1

Analytic induction

Given the exploratory nature of this research, we apply an inductive theory-building
methodology using multiple case studies (Eisenhardt, 1989; Eisenhardt & Graebner, 2007) as
opposed to a single, deep case study (Dyer & Wilkins, 1991; Eisenhardt, 1991). A key
characteristic of a case study is that it “attempts to examine a contemporary phenomenon in its
real-life context…” (Yin, 1981, pg. 59) Our goal in this research is to verify and evaluate the
relevance of the theoretical model established in essay one using real-life cases involving
innovation in interfirm new product development projects. What we are doing is asking a simple
question: ‘to what extent does the data generated in the case study align with the model posited
in essay one?’ If it does not, why not and what changes need to be made? Investigating multiple
case studies enables the research to refine the theoretical model in preparation for a future
confirmatory study where deductive theory building can further establish the theories being
developed.
Inductive theory-building methodologies are categorized into two major groupings: grounded
theory or analytic induction (Bansal & Roth, 2000, pg. 719). Whereas grounded theory requires
researchers to explore case and field studies with little to no a priori hypothesized model

28

established, analytic induction allows for such. In particular, analytic induction requires
researchers to conduct case studies in search for disconfirming evidence that informs further
refinement of the model (Katz, 2001; Manning, 1982). Our study follows the analytic induction
approach and uses findings from the six case studies to refine the a priori model developed in
essay one.
2.2.2

A priori model

Essay one posits that the size of gaps in joint action expectations are directly and inversely
related with innovation performance (proposition 3). Specifically, smaller gaps correlate with
higher levels of innovation performance. Further, Essay one proposes three distinct levels of
interoperability (resource, process and behavior) between two firms and argues that gap size in
joint action expectations serves as a proxy for behavioral interoperability. While we posit that
higher levels of both resource and process interoperability are related with high levels of
innovation performance (propositions 5a and 5b) and higher levels of behavior interoperability
(propositions 1 and 2), we expect these relations to be mediated by behavioral interoperability, as
proxied by gap size in joint action expectations (propositions 4b and 5b). In this study,
however, we narrow the focus of our case studies to proposition 3. We do so in order to allow
the research to more deeply investigate the construct of AJAE and its relation to innovation
performance in the interfirm new product development setting.

29

Figure 2.1: Posited Model from Essay 1

Prop 5a, 5b

PROCESS
INTEROP

Prop 2

BEHAVIOR
INTEROP

Prop 3

(AJAE GAP SIZE)

Prop 1

RESOURCE
INTEROP

Prop 4a, 4b

30

INNOVATION
PERFORM

2.2.3

Multi-Case Sampling

Eisenhardt (1989) provides a framework for conducting multi-case study analysis that
supports analytic induction. This framework includes conducting both within-case and crosscase analyses using a purposeful sample of cases that adequately inform the researcher on the
phenomenon of interest. We followed this framework and conducted six case studies of
innovation in new product development initiatives involving both high and low levels of project
success across varying degrees of product complexity. While the data was collected from
organizations, the unit of analysis was the project.
Project success was initially categorized by the lead contact at the company that identified
what case would be included in the research at the company. Performance data were also
gathered during the actual interview process and was used to verify the performance rating
initially provided by the lead contacts. We also included product complexity as a dimension of
interest given prior research that has empirically validated a significant positive relationship
between product complexity and propensity for vertical integration (Novak & Eppinger, 2001).
If such a relationship does exist, then we wanted to ensure that our study included cases
involving both low and high levels of product complexity so as to diversify our exposure to
AJAE across both dimensions.
Novak and Eppinger (2001) operationalized product complexity as a multi-dimensional
measure of design complexity that varied according to the product. In one example, Novak and
Eppinger (2001) identified the relevant dimensions as 1) the newness of the technology being
used, 2) the number of moving parts in the product and 3) whether the system was active or
passive. We likewise identify the level of complexity of the products involved in each case
study (figure 2) using the three dimensions listed in Novak and Eppinger (2001).

31

While our hypothesized model does not necessarily focus on product complexity as a
construct of interest, we did want to control for this variable in our case selection criteria. As can
be seen in figure 2, our six case studies provide a sampling across all four quadrants of the
sampling matrix. Further, the six cases represent three distinct industries: the oil and gas
industry (Case A), the heavy vehicle manufacturing industry (Cases B, C, D and E) and the
telecommunications industry (Case F).

32

Figure 2.2: Case Sampling

33

The case studies explored in this paper come from a convenience sample generated from the
authors’ industry contacts of chief procurement officers and other executive-level supply chain
leaders across various industries (oil/gas, heavy vehicle manufacturing and telecommunications).
These industries were targeted due to the increased importance that innovation plays in
establishing a sustainable competitive advantage in their respective business environments.
Approximately 20 companies were contacted and sent one-page overviews of the research being
conducted and the type of case studies that would help further the research. Six companies
responded with interest yet only five were able to obtain company approval to proceed with the
studies.
Firms are generally hesitant to allow outside access to sensitive innovation projects. The
relatively low rate of participation, given that those invited already had previous contact with the
researcher, underscores the challenge of studying innovation in the supply chain. Confidentiality
agreements were required by all but one of the companies prior to allowing access to key
informants. Therefore, all names of companies, products or individuals in this paper are
pseudonyms in order to maintain the agreed to confidentiality. Table 1 provides an overview of
the six cases studied. A single case study was conducted at each company with the exception of
Construction Truck Corp where two separate projects were investigated with differing products
and suppliers (Cases D and E). Three of the companies, Armored Vehicles Inc., Blaze Inc. and
Construction Truck Corp (Cases B, C, D and E) are separate companies that reside under a
common parent company.

34

Table 2.1: Case Demographics

Company

Case A

Case B

Case C

Case D

Petroleum
Equipment
Corp (PEC)
Armored
Vehicles Inc
(AVI)
Blaze Inc.
(BI)
Construction
Truck Corp
(CTC)

Case E

Construction
Truck Corp
(CTC)

Case F

Electronics
Connect (EC)

Supplier

Metal Frame
Corp (MFC)
Shield-all
Technologies
(SAT)
Super Motors
Corp (SMC)
Gearbox Corp
(GBC)
Composite
Manufacturing
Incorporated
(CMI)
Design Service
Corp (DSC)

Industry

Project Focus

Project Length at
Time of Interview
(Years)

Total Length of
Relationship
(Years)

Oil and gas

Liquid dispenser

2

6

Heavy vehicle
manufacturing

Small armored
vehicle

4

13

Heavy vehicle
manufacturing

Fire truck

3

30+

Heavy vehicle
manufacturing

Mixer gearbox &
motor

3

20+

Heavy vehicle
manufacturing

Material chute

5

5

Telecom

Fiber optic
connection system

4

4

35

2.2.4

Interview protocol

The interview protocol was designed in accordance with established guidelines for a multicase study (Yin, 2009, pg.81) and was pre-tested with two supplier chain managers. Feedback
from those managers was used to clarify terms and other questions that otherwise might have
been confusing to the respondents. In addition to open-ended questions and Likert scale items,
the protocol also included graphical figures for respondents to use in quantifying responses
related to joint action in the relationship. A copy of this protocol instrument is included in the
appendix.
In this research, the project is the unit of analysis. Given the varied activities associated with
new product development, we asked respondents to describe the different stages of the
development project including ideation, idea development, application development, and launch.
After an overview of the different stages of the project was given, the interview then focused on
understanding both the actual joint action that took place as compared to the desired joint action
that existed for each stage. The respondents used figure 3 below to help them talk through these
items and to express any differences between expected and actual levels of joint action. A circle
represents a desired level of joint action while an X represents the actual level of joint action. A
circle and an X are then plotted for each of the four stages of the product development timeline.
In the case that the focal firm was not at all engaged in a specific phase, then that section was left
blank.
Whenever the respondent identified a gap, we asked for an explanation of why the gap
existed and what it would have taken to close the gap. While the data generated in this study is
used for inductive rather than deductive theory development, use of this assessment tool can also
be employed in future empirical studies with a large enough statistical sample in order to draw

36

deductive conclusions. Thus, this study not only helps shed light on the nature of AJAE, but it
also provides researchers with a common assessment tool that can be used for future
confirmatory studies.
Figure 2.3: Blank Assessment T
Tool

The final part of the interview focused on ascertaining project performance including
dimensions of innovation success. We asked respondents to identify performance using a 77
point Likert scale across
ross seven different dimensions including both innovation and project
performance criteria (financial and timing) as outlined in figure 4. The 77-point
point scale has been
shown to yield similar mean scores, once adjusted for scale, as compare to the 55--point scale
(Dawes, 2008).. Since description, rather than statistical properties of the performance scores
score
were of primary interest, we chose the 77-point
point scale in order to provide the respondent with the

37

opportunity to better differentiate performance across the numerous dimensions. In sum, the
interview protocol was carefully designed with the objective of providing a rich description of
each case that could uncover any inconsistencies in the a priori model, thus enabling the research
team to apply the analytic induction methodology in refining the model.
Figure 2.4: Performance Assessment Survey

2.3 Data collection and analysis
Initial contact with each of the five companies was made with a supply chain executive, who
identified both the new product development projects and the respondents that could best speak
to the topics outlined in the interview protocol. Respondents needed to be knowledgeable of the
events surrounding the chosen project, party to the interfirm relationship and able to provide
insights and informed comments on the processes and outcomes of the project. Further, it was
requested that at least two respondents be identified per case where one respondent could provide
a technical perspective while the other provide a financial perspective. All twelve of the key
respondents across the six case studies had job titles of manager or above: manager (N=6),

38

director (N=5) and vice president (N=1). The twelve respondents had a median seven years with
the company that employed them. The minimum was a purchasing manager with four years and
the maximum was an engineering manager with eighteen years. Where possible, respondents
were interviewed alone with the researcher and not in the presence of others in order to
encourage openness in the interview.
Copies of the semi-structured interview protocol, along with consent forms were sent to the
respondents beforehand in order to expedite the interview process. However, we did not request
that the protocol be completed beforehand; rather, we provided the protocol as a preparatory tool
so that the respondent would know what questions and topics would be discussed in the face-toface interview. The signed consent forms were collected at the time of the face-to-face interview
and are stored in a locked and protected file storage system per IRB standards.
Interviews for five out of the six case studies were conducted on-site and included tours of
manufacturing and engineering facilities. The interview for Case F, however, was conducted
over an international teleconference link and an on-site tour was not feasible. A hard copy of the
interview guide was used to direct the interviews and to capture responses to the two figures
above. Respondents were encouraged to elaborate or deviate from the protocol whenever they
felt relevant information on the topic needed to be mentioned. Two respondents were separately
interviewed for each case with the exception of cases E and F, where only one interview was
conducted yet with multiple respondents participating in the interview. All questions were
discussed with both participants simultaneously until a consensus was made on the response to
the topic at hand.
Interviews were conducted by a single researcher and were audio recorded with permission
from the respondents so that the researcher could focus on conducting the interview. Audio files

39

were subsequently transcribed and transcripts were emailed to the respondents to provide an
opportunity for correction. The transcripts were then edited to replace any actual names of
persons or companies mentioned with alternate names in order to maintain anonymity of both
participants and the companies involved. Given the competitive and sensitive nature of strategic
new product development projects, the research team assured the respondents and their
companies of anonymity in order to encourage openness in the interview and data gathering
process. Additionally, once the transcription and name replacement process was completed, the
original audio files were destroyed.
Sanitized transcripts were then imported into a qualitative analysis tool. For this project, we
used QSR International’s NVivo 9 qualitative data analysis software to manage and analyze the
data collected. The imported transcripts were coded according to themes that emerged, such as
joint action expectations, gaps in expectations, and project performance. The case studies were
conducted over a 9-month period of time starting late 2012. Coding protocols and results were
reviewed with the research team as the case studies progressed and new constructs were added to
the coding protocols as new themes emerged that were relevant to the study. Whenever new
themes emerged and were coded, prior cases were re-examined in search for the same new
themes. The codings were then used to conduct both within-case and cross-case analyses. We
first present the within-case analyses and then synthesize our findings in the cross-case analysis.

2.4 Within-Case Analysis
2.4.1

Case A: Petroleum Equipment Corp

Petroleum Equipment Corp (PEC) is a US-based company that provides fuel-dispensing
equipment to the petroleum industry. The company has over 3500 employees and annual
revenues of over US$1 billion. Products produced at PEC are exported around the globe.

40

Metal Frame Corp. (MFC) is a US-based company that provides metal-based components,
subassemblies and finished products to various industries including agriculture, heavy
construction, automotive and petroleum. The company has over 400 employees and annual
revenues of approximately US$150 million. MFC’s customer base resides mostly in the US.
2.4.1.1 Idea Conception
In 2010, PEC approached MFC, one of PEC’s key suppliers from a purchase value
1

standpoint, and invited them to participate in an on-site Kaizen event at PEC to analyze the
sheet metal body and design of a key fuel-dispensing product supplied by MFC.
MFC sent an engineer to participate in the kaizen event at the PEC facility. The MFC
engineer, together with personnel from PEC brainstormed ideas for design improvement that
could yield cost reductions in the products supplied by MFC. Multiple ideas were generated and
analyzed for both design and economic viability. The team converged on one idea that reduced
the amount of sheet metal used and simplified the overall design from both a manufacturing and
assembly standpoint.
The overall time spent in this idea conception phase lasted about one week. Both teams were
satisfied with the results. In particular PEC was pleased with the estimated pricing that MFC
quoted during the Kaizen event for the redesign.
2.4.1.2 Idea and Application Development
The idea development and application development stages are one in the same in this
particular project. While the idea conception phase was completed in a one-week time frame,
1

A Kaizen event is a continuous improvement initiative focused on a product or process.
Generally, individuals responsible for different parts of the product or process convene to
identify opportunities for improvement and to plan a process of implementation for those
improvements.

41

the development of that idea and ultimately the specific application design took over 1.5 years to
finalize. The key challenge the firms encountered in the development stage was determining the
appropriate manufacturing process to be used in implementing the new design. Prototypes of the
new design were fabricated using competing variations and manufacturing processes related to
conjoining the sheet metal assembly. Through this iterative learning process a final design and
process was decided upon. At this point in time revised quotations for the redesign were
presented at price levels much higher than what was originally quoted at the kaizen event.
Additionally, ownership of the tooling had not yet been determined. After further negotiation
including an agreement to a multi-your contract, MFC agreed to pay for and own the tooling.
While the tooling agreement removed some of the financial uncertainty for both sides, it also
created a situation where changes requested by PEC were often met with skepticism by MFC
since MFC was carrying the financial burden associated with change requests. PEC engineers
were focused on maximizing the technical performance of the redesigned component, while
MFC engineers pushed back on those items they felt were not necessary to achieve compliance
with the specifications. Disagreements, for example, arose regarding the number of joining
points needed between the two mating parts. Such design disagreements would be settled
through the production and testing of prototype samples.
Additional complications arose with respect to the speed of communication between the two
teams. What had initially been a shared sense of urgency amongst the two teams quickly turned
into a one-sided focus where MFC would respond in a slow, careful and well-measured cadence
that left PEC frustrated from a timing standpoint. The delayed implementation timeline,
combined with the decreased promise of savings became a point of consternation for PEC.

42

2.4.1.3 Launch
The redesigned product finally launched two years after the initial Kaizen event. The launch
itself could have taken place slightly sooner, however there were delays due to managing excess
raw material specific to the older design that was already committed to. While this issue could
have caused an even greater launch delay, PEC purchasing and engineering identified an
alternate application that could consume the obsolete material and committed to purchase the
material from MFC accordingly.

43

Figure 2.5:: Actual vs. Desired Joint Action - Case A
Respondent 1

Respondent 2

44

Table 2.2: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case A

Respondent 1

Respondent 2

Idea conception

“They were very actively participating, they
sent three different people to come participate
with us, they were varying engaged with us. I
would say that the expectation was definitely
met”

“The guy that was here for the Kaizen event… that was
a really good event… but he was here only for the
event. But once we got out of the event the idea
seemed to stall.”

Idea development

NA

NA

Application
development

“I think it would've been more helpful had
they been more on-site, because many of the
things that we were coming into, where the
actual idea needed to be adjusted was because
of something that one of the engineers here
observed on the line that they can’t really see
when they are not interacting completely”

“It felt like they were always wanting to push us
towards what they wanted to do in stamping and with
their processes in stamping rather than taking an open
box approach… and so that idea development seemed
to fall into just one area, where they wanted it to go. So
I felt like we got boxed in with some of the ideas we
had. So we could have had a better joint effort to
develop whatever works…”

Launch

“…they got better again… and I think that
was because that new project manager kind of
came on board and was the one that helped
get this pushed through to completion… he's
here a lot more often, and is a lot more
participative. So I would say that that
definitely helped during the implementation
phase”

“I would have wanted them to be a little bit more
involved in the field in trials and in production runs
here. I think the guy came here maybe one time, or
maybe two times. We ran field trials for probably a
month, and it would have been good to have him here
for a couple of those”

45

Figure 2.6: Performance - Case A
Respondent 1

Respondent 2

46

2.4.1.4 Performance
It can be noted in the performance scores that both respondents appeared to agree that the
project did not achieve the highest scores of either 6 or 7 with respect to innovation (the first four
categories). Both were in agreement on the lowest score possible on timing and slightly below
average financial (budget performance) and market acceptance. We categorize this case as one
with an overall lower level of performance.
2.4.1.5 Key points
While both firms shared an initial vision of what would be achieved, execution during the
idea and application development phases slowed down significantly due to lapses on
communication. Both respondents suggested a lack of trust contributed to the slow and
measured rate of response to communications and to problems that arose.
2.4.2

Case B: Armored Vehicles Inc.

Armored Vehicles Inc. (AVI) is a US manufacturer of medium and heavy-duty vehicles used
for military logistics. While AVI’s key customer is the US military, it also sells its transportation
solutions to other local and national governments around the globe. AVI’s vehicles are often
used for transporting military personnel and equipment in both on-road and off-road
environments.
Shield-All Technologies (SAT) is an armor solutions company that is headquartered in the
Asian continent. Its key engineering and manufacturing facilities are located near its
headquarters facility. SAT also has a business presence, including manufacturing, on the North
American and European continents and market their armor solutions to companies and
governments across the globe.

47

2.4.2.1 Idea Conception / Idea Development
SAT approached AVI in 2008 with a business proposition. SAT had identified a mediumduty chassis made by a large US OEM that could be used to integrate an SAT armor solution in
order to produce an armored vehicle larger than a typical SUV but smaller than a Hummer-sized
vehicle. This unique-sized armored vehicle segment had not yet been exploited.

The chassis is

similar in size to those used by tow trucks.
SAT had already produced a concept-evaluation prototype to prove the concept. What SAT
needed was a vehicle manufacturer that could commercialize and produce the vehicle in larger
quantities than what it had ability to produce. While it was able to make a prototype concept
vehicle, SAT did not have the vehicle manufacturing skills needed to handle the project
themselves. AVI could provide the necessary skills to make the project succeed. The project
was named Desert Fox and the target market was local governments and border patrol agencies
as opposed to the main branches of national military groups.
2.4.2.2 Application Development
Executives from both companies recognized the potential that existed in the new segment
and agreed to purse the Desert Fox project. SAT agreed to relinquish all vehicle and
vehicle/armor integration responsibilities to AVI and SAT became solely responsible for the
armor solution. In fact, any changes to the armor solution would need to be agreed to by AVI
before implementation since any changes could impact the integration of the armor solution. In
effect, AVI became the gatekeeper of the project. In return AVI agreed to use SAT as the sole
supplier for the armor.

48

AVI and SAT had an already established executive steering committee that met regularly to
discuss joint projects and to provide assistance where needed. One respondent described the
steering committees as:
‘…anyone would anticipate there's going to be some conflict. So there was a
mechanism built into this to have an executive level steering committee, because there
was a desire to have a close working relationship tween the two organizations, Shield-All
Technologies and Armored Vehicles Inc; it had existed before Desert Fox came along.
There was an interest in continuing that relationship… and not just for this one product
but for lots of different product initiatives’

Application development for this project required AVI to learn how the concept vehicle was
made and to establish a supply chain and manufacturing process that could repeat the design in
an economically feasible and reliable manner. Doing so was no small task and required much
coordination between SAT, AVI and BuildSmith, a chassis modifier located near the OEM
chassis manufacturing location. SAT and BuildSmith had established a business relationship
before AVI came into the picture. Their sole purpose was to help SAT build vehicle prototypes
that SAT would use in marketing their product.
While BuildSmith was initially engaged by SAT for the initial concept vehicle build, they
were only guaranteed work for an initial 10-vehicle pilot build and did not have a business
relationship with AVI. The uncertainty surrounding BuildSmith’s future involvement with AVI
was a complicating factor in this three-party initiative for the pilot build. On the one hand, they
were desirous for the future business and wanted to put their best foot forward in order to
showcase their capabilities to AVI. On the other hand, the risk existed that their efforts would
yield no returns. AVI ultimately decided to use the in-house manufacturing facility of a sister
business group to manufacture the vehicle, leaving BuildSmith without any portion of the
business.

49

2.4.2.3 Launch
This case was interesting in that the idea creation and development stages were done 100%
by SAT, the upstream supplier, with no input from AVI, the downstream customer. AVI,
however, took over the application development and launch stages and quickly assumed
approximately 90% of the workload required to execute these last stages. According to an AVI
manager,
‘…the basic product we were not involved at all (sic), but now the sustainment of the
product and the adaptation of it to meet customer requirements in the US and our
international markets, we control the lion share of that. So I would say that is 90%… We
are 100% leading it, but we do have to farm work back to Shield-All Technologies.’

Frequent communication between the two firms was cited as a key ingredient in enabling the
success of this project. Regular meetings were held at both executive and working levels. The
executive-level meetings were established prior and independent to the Desert Fox project and
were attended by presidents of both companies. This executive steering committee discussed
issues pertaining to any of the joint projects, including the Desert Fox project and would resolve
issues that bubbled up from the working level.
At the working level, employees from both companies engaged in weekly meetings that took
place whether or not there was much content to cover. In the words of an AVI manager,
‘…we had to set up a regularly scheduled time, a weekly team meeting whether or not
anybody thought that we had something to talk about, we were going to have a weekly
event where people from both sides got together on the phone and talk through a project
agenda… This is what's going on, this is what's going to need to happen from an
Armored Vehicles Inc perspective, this is what support we need from Shield-All
Technologies, this is when we need it. And once we got that going on a formal schedule
things really worked well. So solving the communication problem, getting the right
people talking from both sides to each other was really the key, for my judgment, to the
success of the project.’

50

In addition to the manufacturing and supply chain design and setup work that was done, both
companies also engaged in global marketing efforts. These efforts, however, were not as well
coordinated as were the manufacturing and supply chain aspects of the project. Situations
occurred where AVI marketing personnel contacted a government organization to only find out
that SAT marketing personnel were already in talks with that same government entity. While
there seemed to lack an overall coordination in marketing, this was not viewed as a key
roadblock to the success of the project. In fact, AVI admitted that SAT was able to bring
customers to the table that AVI did not have access to and that the joint coverage of the two
marketing teams yielded strong success.

51

Figure 2.7:: Actual vs. Desired Joint Action - Case B
Respondent 1

Respondent 2

52

Table 2.3: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case B

Respondent 1

Respondent 2

Idea conception

NA

NA

Idea development

NA

NA

Application
development

“There was cooperation with regard to customer
development… the communication was frequent and
regular… both parties were motivated to see the program
succeed.”

NA

Launch

“When we needed Shield-All Technologies to send people
here, they did. You know, they stepped up to that. Again our
expectation for how far along they probably should've been,
or where we would have liked them to have been with
documentation may or may not have been realistic… maybe
the people to sign the agreement didn't realize how much
work needed to be done when they stepped into this thing, but
they really did step up"

“… I think they did an outstanding job
supporting us…they met our
expectations”

53

Figure 2.8: Performance - Case B
Respondent 1

Respondent 2

54

2.4.2.4 Performance
Overall, the Desert Fox project was deemed a success. This case scored above average
scores on the innovation performance dimensions of challenging existing ideas, offering new
ideas and creativity. Scores were in the average range for value creation, market acceptance and
project execution dimensions of timing and cost. We categorize this case as one with an overall
higher level of success.
2.4.2.5 Key points
This is an interesting case where two supply chain partners engage partway through an
existing product development project rather than at the idea conception phase. This added the
complication of melding two existing supply chains where BuildSmith eventually needed to be
disintermediated for the new relationship to work. Carefully managing this process was
important to the success of the project.
2.4.3

Case C: Blaze Inc.

Blaze Inc. is a US manufacturer of fire trucks. Their customer base comprises mostly of
local governments, in particular, fire departments that are established and controlled at the city or
township level. Fire trucks are expensive investments for these fire departments and are highly
customized in terms of fire fighting accessories and capabilities. Often the fire chiefs that run
these fire departments have significant input into the customization of the order. Blaze Inc.
purchases heavy-duty engines from engine suppliers and assembles the engines to the chassis.
They then install the required optional equipment (water tanks, pumps, auxiliary power units,
etc.) and then paint and finish the trucks in preparation for final delivery to the customer.
Super Motors Corp. (SMC) is a heavy-duty diesel engine manufacturer. SMC designs
engines for multiple applications including light-duty, medium-duty and heavy-duty trucks.

55

SMC sells their engines across the globe and to numerous industries. In fact, SMC engines are
one of the leading choices of engine options in fire truck manufacturing. While SMC is a
leading engine supplier for fire trucks, they are not without competition. Orion and Cruzada are
both strong competitors with similar global reach and technical capabilities that are considered
viable options to SMC.
The performance of an engine in a fire truck has a significant impact on the performance of a
fire fighting team. Not only must the engine reliably and quickly move the heavy fire truck from
the station to the point of need, it must also provide the electrical and hydraulic needs of the lifesaving accessories installed on the fire truck. For example, many fire trucks carry a water
reservoir and pump that can be used on a limited basis until a fire fighter is able to tap into the
water supply and pressure of a nearby fire hydrant. Additionally, the engine must reliably
provide the power used to lift a rescue ladder to its needed position. It is no understatement to
highlight the fact that lives often depend on the performance of the fire truck engine.
2.4.3.1 Idea Conception
SMC’s position as one of the leading fire truck engine suppliers was not sufficient enough to
keep them from attempting to leave the business. The highly fragmented fire truck market meant
that SMC needed to deal with numerous different customers, each attempting to differentiate
their product from that of their competition, thus driving complexity into the application phase of
engine manufacturing. This complexity, coupled with low order volumes made the fire truck
market an unattractive business to operate in. New emissions standards, set to become effective
in 2010, would necessitate that most engines be redesigned in order to ensure compliance.
When the global financial crises of 2007 began to take shape, SMC evaluated the multiple
market forces at play and decided to exit the business altogether.

56

SMC’s announcement to exit the market sent shockwaves throughout the industry and
manufacturers quickly began aligning future business with the two remaining engine
manufacturing giants in the industry: Orion and Cruzada. In the midst of this situation, Blaze
Inc. conceived a strategic alternative involving SMC in an attempt to keep SMC in the business.
Blaze Inc. proposed an exclusivity agreement with SMC for the supply of fire truck engines.
Blaze Inc. would agree to use only SMC engines for 100% of their fire truck sales, and SMC in
turn would sell their engines only to Blaze Inc. It was Blaze Inc.’s hope that providing SMC
with a combination of their entire volume coupled with a single customer interface would present
an attractive business proposition to SMC. Further, Blaze Inc. anticipated that the risks inherent
with any single-source agreement would be outweighed by the strategic positioning that Blaze
Inc. would have in the marketplace as the sole source of fire engines using SMC diesel engines.
The value that customers placed on SMC engines was key to the potential success of the
strategy.
2.4.3.2 Idea Development
Convincing Blaze Inc. to sign up to the exclusivity agreement was not an easy task. For one,
the two companies, despite having a history of cooperation, were going through a rough patch in
their relationship. A few years previous there had been a falling out amongst the two firms’
leadership. While those same leaders were no longer in their same positions, the pall of mistrust
they had cast over the relationship remained. However, what also remained was the recollection
by employees at both companies regarding how strong the relationship was prior to the falling
out. Ironically, the combination of these recollections with the current situation served as an
impetus to leverage the opportunity as a means to repair the relationship. In the words of a
manager close to the situation:

57

‘There was interest on both sides in figuring out how to get back to working well
together. We had some key individuals and some emotions that had gotten in the way
and it had pushed us apart. We recognized it, our customers recognized it and we wanted
to bring it back together, because we know that there was an opportunity. A lot of people
saw this as a stepping-stone in getting us back into the right direction. So having the
history, and having the incentive of ‘this is where we want to get back to’, I think helped
us plow through this agreement. Even though we had a lot of skeptics and weren’t quite
sure if we were going to be able to pull it off, there was that incentive to get back to
where we once used to be. That has carried us through.’
SMC and Blaze Inc. eventually penned the exclusivity agreement; this created a significant
win for both sides. Blaze Inc. would be the sole distributor of fire trucks using SMC diesel
engines. SMC, in turn was able to reduce the complexity in part proliferation by dealing with
just one customer while ensuring their volume forecast by aligning with one of the top fire truck
manufacturers in the nation. It was anticipated that the net sales impact of this strategic move
would be positive, where increased sales through Blaze Inc. would offset and surpass the sales
they had anticipated with other smaller players in the market.
SMC thus began designing the new engine that would meet the 2010 regulatory
requirements. Blaze Inc. participated in the idea development phase mainly by providing input
to the engine performance specifications from a functionality, durability and regulation
compliance standpoint. The electronics interface between the fire truck’s many electrical
accessories and components and the engine itself was particularly complex. Blaze Inc.
acknowledged that most of the effort and contribution of the idea development stage was born by
SMC.
2.4.3.3 Application Development
Once the engine design was developed, the next stage in the project was to finalize the
physical integration of the engine onto the truck and to validate the integrated truck. By all

58

accounts, this process ran smoothly and Blaze Inc. was satisfied with the degree of joint action
engaged by both parties.
2.4.3.4 Launch
One area that could have seen better alignment with joint action expectations on launch was
joint marketing efforts. One respondent cited missed opportunities for using the engine in
adjacent truck products made by a sister company (not necessarily fire truck applications) while
another respondent cited the need for SMC to have more strongly defended and marketed their
technological approach to engine design when another competitor announced to the public that it
was pursuing a different technology for future engines. Despite these identified shortcomings,
however, the Blaze launched on time, on budget and with strong sales.

59

Figure 2.9:: Actual vs. Desired Joint Action - Case C
Respondent 1

Respondent 2

60

Table 2.4: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case C

Idea conception

Respondent 1

Respondent 2

“I would say in the idea conception, we were
probably pushing Super Motors Corp. more than
maybe what they wanted. They had just made the
announcement that they were not selling to any
OEMs, and here we come and say, ‘come be our
partner’…”

“…in the idea conception, we would have liked to
have more input; we had higher expectations than
what we actually got… we felt that there was more
opportunity than what Super Motors Corp. actually
wanted to support”

“I think it took a little time for them to warm up
to the idea. But once they got on board, like I said
here during the idea development, we got pretty
close to our expectations versus their
involvement”

“…in the idea development, we did a lot of joint
testing to ensure that their actual design was robust."

Application
development

“…the application development, I would say we
were in lockstep”

“…we worked quite well together… getting those
programs together and then actually developing a
main application, developing the programming so
that it would support the pump and interface with our
systems”

Launch

“…at the launch I would separate us again a little
bit. I think we would have liked to have had more
of their involvement and resources to market the
product, and to really push the fire industry and
make this a little more of a marketing splash and
a joint development.”

“… the production launch delivering on time and
everything else, but actually spending time out in the
market together, marketing and selling the product
together”

Idea development

61

Figure 2.10: Performance - Case C
Respondent 1

Respondent 2

62

2.4.3.5 Performance

We categorized this case as one with a high level of success. The market reception to the
redesigned fire truck using SMC’s diesel engine was very positive, with sales surpassing initial
expectations. The product launched on time and on budget. Both respondents considered the
project an innovation success from not only a product standpoint, but from a business strategy
standpoint as well, where the new exclusivity agreement with SMS provided both firms with a
sustainable competitive advantage that will be difficult for competitors to challenge into the
future. Not only was the project itself successful, but the overall relationship between the two
companies improved as a result.
“Given some of the difficult history that both companies experienced in the past with
each other, I think initially both sides thought that this was something that would never
happen. But when we did pull it off, it served as a catalyst. Almost overnight the
relationship switched. So now we are to the point where our relationship with Super
Motors Corp. is one of the least contentious. Things are pretty easy going in that
relationship. Other than occasionally they are not really responsive, they’re very
supportive, they are really easy to work with on commercial terms that are typically a
challenge with other suppliers: vending agreements, material cost price adjustments,
exclusivity… they are pretty easy-going.”
2.4.3.6 Key points
This is a case where the value added innovation was not only in the product development
initiative, but also in the innovation of the business model. The shared vision of the new
business model became the foundation and driving force behind the success of the new product
development initiative.

63

2.4.4

Case D: Construction Truck Corp. and Gear Box Corp.

Construction Truck Corporation (CTC) is a North America-based company that assembles
mixer trucks for the construction industry. CTC assembles the truck chassis to the mixer barrel
and other required accessories and then sells the completed truck through a dealership network.
Gearbox Corporation (GBC) is a European-based company that designs and manufactures
driveline systems for the vehicle industry. Their products span both commercial and consumer
vehicles. GBC for many years was the sole supplier of the gearbox used by CTC trucks on their
mixer barrels. CTC combines the GBC gearbox with one of two different motors supplied by
other suppliers. This combined gearbox/motor system is responsible for turning the barrel while
the mixed product is being transported for use at a construction site.
2.4.4.1 Idea Conception
In 2009, GBC notified CTC that it was developing a new generation design that would
integrate the gearbox and motor into one assembly using a MasterMotors Inc. (MMI) designed
and built engine. Despite the brand recognition that MMI carries in other industries for
designing and building motors, this would be MMI’s first time building motors for CTC mixer
trucks. The Gen-2 design, as the integrated design was called by GBC, would be approximately
350 pounds lighter than the prior gearbox + motor system. It would reduce the overall package
size and would also reduce operating noise with the use of an isolation barrier between the
assembly and the chassis mount. While these achievements were considered definite
improvements, they came at a cost from both a purchase price and a serviceability standpoint.
The price for the integrated system was significantly more than the combined stand-alone prices
of the original gearbox and motor. Additionally, the integrated system offered less warranty than
what was being offered with the original system and would cost more to repair and service.

64

2.4.4.2 Idea Development
GBC did not solicit CTC’s input or approval prior to proceeding with the redesign and
assumed all development costs and risks associated. The initial reaction by CTC when notified
of the development project was guarded; nevertheless GBC forged forward with idea
development and did not involve CTC at all in the development process. As the project
developed and more details of the design emerged, CTC became greatly concerned regarding
both the serviceability and cost aspect of the redesign. If the motor failed on the newly
integrated design, the entire module would need to be replaced: gearbox and motor included. In
the original system, if the motor failed, then just the motor itself could be repaired or replaced.
The service infrastructure and know-how to manage the current system was well established and
would be deemed irrelevant with the Gen-2 system. All repairs for the Gen-2 system would need
to be handled by GBC, and not by the existing service infrastructure.
2.4.4.3 Application Development
Despite objections from CTC regarding serviceability, GBC proceeded with developing the
Gen-2 system. Given GBC’s lack of alignment with CTC’s expectations on this product, CTC
realized that maintaining GBC as a sole supplier for the gearbox was not in the best long-term
interest of the company. They therefore resolved to introduce competition for this product and
began working with an alternate gearbox supplier who would not only manufacture gearboxes
with the desired warranty level, but would also brand the product with CTC’s private label.
CTC resolved to attempt once more to realign the GBC relationship. They assembled their
entire leadership team of not only CTC leaders involved with GBC, but also those leaders of
other businesses under their parent company that also did business with GBC. This combined

65

leadership contingency traveled to GBC’s European headquarters and communicated their
concerns to the GBC leadership team.
Pricing and warranty were key topics of negotiation, however warranty was the overriding
concern. The warranty terms offered for the Gen-2 system were actually less than what GBC
was currently providing on their existing stand-alone gearbox. The constraining factor on the
Gen-2 system warranty terms was MMI’s motor. Since MMI was not willing to match the more
generous warranty terms of the gearbox, GBC could only warrant the Gen-2 system at the level
provided by MMI. In the meantime, the alternate supplier that CTC developed offered an
improved warranty on the private label gearbox, even beyond that which was currently being
provided by GBC.
Only marginal improvements to warranty terms and pricing were achieved through the CTC
leadership contingency visit to GBC’s European headquarters. The team therefore resolved to
push forward with launching the alternative supplier’s gearbox with the CTC private label in
addition to the Gen-2 system provided by GBC.
From an application development standpoint, CTC needed to make alterations to their
existing truck pedestal that holds the mixer in order to accommodate the Gen-2 system design.
CTC proceeded with this redesign and tested the new product. In essence, CTC proliferated the
number of pedestal designs in order to accommodate not only the new Gen-2 system, but also the
new alternative supplier’s gearbox design so that the final customer could choose between the
available options.
2.4.4.4 Launch
CTC has a customer catalogue that details what options are available for order. Customers
use this catalogue in customizing their truck orders. Flexibility, however, exists for customers to

66

request adaptations not specifically detailed in the catalogue. One key consequence of CTC’s
dissatisfaction with GBC’s Gen-2 system was that the Gen-2 system was not included in the
customer catalogue. If customer’s wanted the Gen-2 system, they would need to specifically
request the product. CTC instead included in the catalogue the new private label gearbox that
provides a lower price and greater warranty than even the current GBC system. The updated
catalogue offered customers the option to either order the new private label option or the
traditional GBC option, but did not explicitly advertise the Gen-2 system. If a customer,
however, was aware of and desired to specify GBC's new Gen-2 system, they could do so and
CTC would be able to accommodate the request into the order.
Since CTC did not include the Gen-2 system in their catalogue, GBC embarked on a
marketing campaign of their own in order to market the new product directly to end-users.
GBC’s marketing campaign was done on their own and was not planned nor executed jointly
with CTC. CTC forecasts projected no more than 10% -20% market penetration for the new
Gen-2 system (as a percent of CTC total sales) – a far cry from the previously held 100% market
penetration that GBC previously held with CTC assembled trucks.

67

Figure 2.11:: Actual vs. Desired Joint Action - Case D
Respondent 1

Respondent 2

68

Table 2.5: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case D

Respondent 1

Respondent 2

Idea conception

"Idea conception was theirs…I'm kind of
"…early on during the concept development phase we
satisfied with how it went… we’re not going to would have liked to have done joint development. We
start designing our own mixer drives"
would have liked to have sat down with them and said,
'look, your current product is good, but we struggle with
bearing issues and leaks and seals. So there are areas of
opportunity, and if we could get it smaller, lighter… So we
would have liked to have been part of it, instead of them
developing a solution that was actually counterintuitive to a
solution that we would have asked for."

Idea development

NA

NA

Application
development

"Application development… there was some
interaction where they sent us a few and we
tested them. We do this every day, not a big
deal"

"…they provided the required information for the vehicle
application development"

Launch

"…on the launch, if there were negative numbers I would
give it to them because they are undermining everything we
"…and then a little more communication as we are doing. They are trying to go to the end customer and
got closer to launch to make sure parts were
incent the product and everything else… Their actions are
available"
going to force us to deal with the product that we don't want
to. They are making matters worse for us."

69

Figure 2.12: Performance - Case D
Respondent 1

Respondent 2

70

2.4.4.5 Performance
We note that the respondents interviewed offered markedly differing views on expected
levels of joint action across the different stages of the project while generally agreeing on the
actual levels of joint action experienced. This difference highlights that gaps in joint action
expectations can exist even within a single firm. While our research focuses on the firm as a
single actor and seeks informed participants that can speak for the firm regarding a specific
project, divergence as we observed is a real possibility and can provide rich insights into the
dynamics of the relationship-building process. In this particular case, inspection of the interview
transcript reveals that respondent 1 was pleased with the outcome of introducing a competing
supplier for the component. In fact, little was mentioned regarding the reduced serviceability of
the module while much was mentioned regarding the positive outcome of having two competing
suppliers for the product. The two respondents viewed the same events through different lenses.
2.4.4.6 Key points
What provide interesting insight into this case is the performance scores assigned by both
respondents. Scores across all dimensions match within two units on the Likert scale with the
exception of 'offering superior value to the customer'. Respondent 1, who reported small gaps in
AJAE, assigned a '5' to this performance dimension while respondent 2, who reported very large
gaps in AJAE, assigned a '1'. Both respondents agree that the product developed was innovative
from the perspective of challenging existing ideas, offering new ideas, and being creative, yet in
the context of the supply chain, the product was late, over budget and market acceptance was
below average. Due to these ratings, we categorize this case as one with an overall lower level
of success.

71

2.4.5

Case E: Construction Truck Corp. and Composite Manufacturing

Construction Truck Corporation (CTC) is a North America-based company that assembles
mixer trucks for the construction industry. CTC assembles the truck chassis to the mixer barrel
and other required accessories and then sells the completed truck through a dealership network.
Composite Manufacturing Incorporated (CMI) is a North America-based company that deals
in rubber and composite components for the construction, defense, agriculture and recreational
vehicle industries.
2.4.5.1 Idea Conception
A key product on the CTC mixer truck is the chute used to pour construction material from
the barrel on the truck to the desired location of use. This chute is comprised of three sections,
which are all manually hung separately on the side of the truck and are subsequently assembled
into place at point of usage. The chute is traditionally made of steel and each section weighs
approximately 50 pounds, dry. After usage throughout the day, however, the construction
material that sticks to the surface of the chute can cause the total weight of each section to
increase up to 60-70 pounds, each. Given the need to manually assemble and disassemble the
chute at the beginning and end of each project, the weight of each section becomes an important
design consideration that impacts worker safety.
Aluminum is a more expensive material alternative that saves weight. One drawback of
aluminum, however, is its reduced resistance to abrasion. As a result, the lifespan of aluminum
is significantly less than that of steel. CTC decided to search for and develop a technical solution
that would match the weight of aluminum while also achieving the life-span of steel chutes.
CTC found a solution in composite materials. In particular, it identified CMI as one of the few

72

suppliers in the world that could mass produce a specific type of composite that had the
structural strength needed for use in the chute design.
2.4.5.2 Idea Development
A unique aspect of CMI’s design was a proprietary manufacturing process where the
manufacturing tools moved down the line versus the traditional static molding process. While
CMI had never applied this new technology to the chute product line, it was concurrently
developing the technology to service not only CTC, but to also service a major recreational
vehicle company (RVC) for use in recreational watercraft hulls. RVC was moving from a
composite hand lay-up process to the more automated process being developed by CMI. CTC's
portion of the business was less than one-tenth of RVC's; however, CMI's new manufacturing
line would service both RVC and CTC. From a revenue perspective, CTC and their new
composite chute was by far secondary in importance (to CMI) as compared to RVC and their
recreational watercraft hull business.
2.4.5.3 Application Development
The overall project took approximately 3.5 years from start to finish and was characterized
by the respondents as a successful joint effort throughout the development phase. Given the new
technology involved in the project, many surprises and cost overruns were encountered. These
issues did not dissuade CTC from the project and price adjustments were agreed to by CTC to
help cover the unforeseen costs. Both parties made significant monetary investments into the
project. CTC owned the intellectual property of the design while CMI owned the intellectual
property of the manufacturing process and the physical tools needed to make the chute
components. Both design and manufacturing process were key sources of product value in the
project.

73

Since composites are not abrasion resistant, a special coating was required to protect the
chute surface. The initial coating formulation, however, did not perform to expectations and
alternative coatings were developed and tested until the team refined an adequate solution.
Another issue that arose was related to the hooks used to hang the chute segments on the side of
the truck during transport. Initially, the hooks were designed of composite materials like the
chute themselves. This design, however, was not robust enough for the application and the team
developed a steel hook design that then needed to be integrated with the composite chute. The
joint CTC/CMI team eventually developed a solution that integrated these two parts successfully.
2.4.5.4 Launch
After 3.5 years of development, the product launched and thousands of composite chutes
were sold in the first year alone. The product was lauded as a success and orders continued to
come in. Despite the price premium of the composite chute as compared to the aluminum and
steel options, customers were willing to pay for the weight savings that the composite chute
provided. The lighter chute not only enabled a larger construction material payload, but also
helped reduce workman’s comp liability by reducing the lifting forces that employees needed to
exert to assemble and disassemble the chute.
The first year of launch was a complete success. The success, however, was short-lived
when CMI suddenly notified CTC that it would no longer be able to manufacture the chutes.
CMI explained that they needed to focus all their efforts on the RVC business. While they did
not provide details to CTC, it became evident that CMI was encountering difficulties with
sustaining the business that they had developed with RVC, thus placing their entire operations in
jeopardy. CTC was forced to remove the successful composite chute product offering from their
catalogue. Later, CTC discovered that RVC pulled out of their agreement with CMI and

74

abandoned the new technology altogether. While the reasons for RVC’s abandonment of the
technology were not made know to CTC, it was shared that RVC was reverting to the composite
hand lay-up process and that their manufacturing would be shifted to a low labor-cost country.
Since a traditional lay-up process was not compatible with the composite technology needed for
a construction material chute, the chain of events left CTC with no supply of composite chutes at
all and what was initially deemed a strategic breakthrough in composite manufacturing quickly
disappeared into the graveyard of unsuccessful innovation projects.

75

Figure 2.13:: Actual vs. Desired Joint Action - Case E
Respondent 1

76

Table 2.6: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case E

Idea conception

Respondent 1

Respondent 2

"…in the idea conception phase, Construction Truck Corp
came up with the idea of the chute. …I would say that that
portion of the interaction was pretty low. We had the concept
already done when we started our discussions with
Composite Manufacturing Inc."

NA

Idea
development

NA
NA

Application
development

"As we got closer and closer to launch, and as we learned
more and more… in terms of joint action along this entire
period of time, there were weekly phone calls, multiple site
visits from our engineering and supply chain group to their
facility…"
"I think we realized that the quality of the joint action was
deteriorating. But I don't think we realized or understood
until afterwards… we didn't know why... hindsight being 20
– 20, we now understand much more of what was going on
inside Composite Manufacturing Inc., but we didn't know it
at the time."

NA

Launch

NA

"You have tons of communication, doesn't mean you're
getting the project any further. It depends on what the actual
communication looks like…quantity went up, but quality
went down as we got to the end "

77

Figure 2.14: Performance - Case E
Respondent 1

78

2.4.5.5 Performance
While the project launched later than originally scheduled and was over budget, it scored
very high ratings on creativity and idea quality. Additionally, market acceptance was stronger
than anticipated. The respondent hesitated, however, to give the highest marks on 'offered
superior value to the customer'. The respondent explained "It does offer superior value to the
customer in terms of performance. It is lighter, it does last, but we've got to get the price right
though. I couldn't say that that's a seven until we get the price to hit the target".
2.4.5.6 Key points
From a product only standpoint, the project was a success. From an overall new product
development standpoint, however, the project failed since the launch never completed and was
altogether cancelled. (Yin, 2009, pg.81) operationalize innovation performance as the fraction of
revenue at a firm that comes from products that are either new to the world market (radical
innovation), or new to that firm (incremental innovation). Implicit in this operationalization is
the fact that only projects that result in revenue can be considered a success. We therefore
recognize that while CTC and CMI jointly developed an innovative product that the market
initially accepted, failure to sustain the launch relegated the project as one with a low-level of
success (figure 2).
2.4.6

Case F: Electronics Connect and Design Service Corp

Electronics Connect is a large European-based company that supplies electrical and
communication connection and device systems to industrial and commercial applications across
various industries. Electronics Connect has a global customer base and competes with some of
the world’s largest electrical and communication device manufacturers.

79

Design Service Corp is a small engineering design company based in Europe that provides
design services to companies in the electronics industry. They currently have fewer than 200
employees and follow a business model that is focused solely on engineering services. In other
words, Design Service Corp does not produce or manufacture any tangible products, but rather
helps firms in their product development initiatives
2.4.6.1 Idea Conception
Electronics Connect developed a concept for a new fiber optics connection system that would
enable the customer to assemble the connection system in the field themselves using a special
tool set. While Electronics Connect has expertise in both fiber optics and connectors, they didn't
have the know-how to develop both the connection assembly protocol and the specialized tools
necessary to launch this product. As a result, Electronics Connect initiated a search to find the
appropriate supply chain partner with the mechatronics background needed to support the
project.
2.4.6.2 Idea Development
Using an external consultant familiar with such type of engineering services companies,
Electronics Connect identified a short-list of companies and then pursued a series of audits and
site-visits necessary to narrow down the list to two companies. Throughout this search,
Electronics Connect made clear to all companies the expectation that they would own 100% of
all intellectual property generated with the project. Any company that could not accept these
terms was not considered. Ultimately, two companies were chosen that would participate in a
six-month competition for the business; Design Corp was one of the two companies. In the
words of one respondent,

80

"We told both of them that they were competing for the job and that after six months, we
would make up our minds with whom we would work together. Basically we would
select the one with not only the best ideas, but also the most cooperative one, the most
open one, most willing to work together towards a solution".
Given the small size of many of these engineering services companies, Electronics Connect
resolved to pay each company for their time in participating in the design competition. At the
end of the competition, Design Corp was chosen as the winning partner. Given Design Corp’s
business model of engineering services, there was no desire on their part to own any intellectual
property, thus this topic was not a debate for the two firms.
2.4.6.3 Application Development
Key elements of the application development phase were 1) to define and prove the
necessary steps needed to make a fiber-optic connection using the system, and 2) design and
develop the necessary tools that the customer would need to make the connection. The majority
of the time elapsed during the overall project timeline was dedicated to this application
development phase of the project.
Design Corp’s CEO was personally involved during the design competition defining the
framework of the overall solution and ensuring appropriate support was being provided to the
project. Upon securing the business, however, working-level engineers at both companies took
over the project. These engineers would meet every two weeks for status update meetings.
Additionally, once a month, management from both companies would meet as a steering
committee to discuss high level issues, key milestones and budget performance. Given the
uncertain nature of the project, a flexible 'time plus material' contract was used as opposed to a
'fixed-price' contract.
Electronics Connect was pleased with the frequency and the openness of communication
with Design Corp. In this case, communication was referred to as the timely sharing of

81

information and feedback relevant to the project. In fact, such communication was cited as a key
decision criterion in awarding them the business in the first place. According to an Electronics
Connect manager,
“I think very positively was the openness of that company – that they are willing to share
all their successes, but most of all they were also willing to share their failures. The thing
that was more the key in getting to the point where we are right now was a really open and
honest communication”.

Part of the upfront agreement between the two companies recognized the possibility and
ability of Electronics Connect to request a personnel change in the project team should lack of
individual performance necessitate such a change. The Electronics Connect manager credited
success in implementing this contract term with the openness in communication that existed
between the two firms, where both firms freely shared feedback with each other. Regarding
exercising the personnel change clause in the contract, one Electronics Connect manager
explained:
‘In our country they would say, ‘maybe it’s good if you leave the room for a moment’,
because we don’t want to have that discussion in front of that guy; but Dutch guys would
probably just say, ‘this guy is not performing on this and this and this…’ and they would
have no issue with that at all, which makes it very easy as a customer to discuss
performance with them. By the way, they also do it with us and our performance!’

The culture of open communication without offense was a key element in aligning
expectations to reality throughout the project.
2.4.6.4 Launch
While Design Corp was brought in for their expertise in designing and developing the
connection protocol and related mechatronic tools, they do not have large-scale manufacturing
expertise. As a result, Electronics Connect partnered with another firm for the launch phase of

82

the project. Design Corp was still involved to a small degree, however, ensuring that any
ensuing changes to the design did not violate critical performance assumptions.
Electronics Connect initiated and ended the relationship with Design Corp, within the scope
of the single project. In other words, there was no attempt to establish a long-term supply chain
partnership with expectations of multiple projects into the future. Rather, both firms understood
that their efforts, risks and rewards were confined to the context of the single project.

83

Figure 2.15: Actual vs. Desired Joint Action - Case F
Respondent 1 (group)

84

Table 2.7: Respondent Quotes Regarding Actual vs. Desired Joint Action - Case F

Respondent 1 (group)

Respondent 2

Idea conception

"Idea conception, that was fully on our side. We
came up with the idea."

NA

Idea
development

"…we meet each other every two weeks where we
discuss the results of the past two weeks and where
we failed on both sides"… regarding actual and
desired joint action, "it was aligned"

NA

NA

Application
development

researcher: "OK… what would have closed that
gap?"
respondent: "Competences on our side. We did not
have the resources, nor the competencies to review
the part of their work because we stayed too far out
of their domain. We should have been more
involved in reviews. I think we should have done
more from our side. It’s not their mistake, I think it
is our own, we should have managed this better."

NA

Launch

"For the launch we are going for the mass
production sets and designs, and that’s not their
expertise… we really need to go and partner with
somebody else. So they will still be involved to
review what we are doing and that we do not break
any rules that we established in the past, but the
majority of the work will not be executed by them
but by someone else."

85

Figure 2.16: Performance - Case F
Respondent 1 (group)

86

2.4.6.5 Performance
Overall the project was reported as being highly successful, with 10 patents filed for the
tooling portion of the project alone. Electronics Connect made clear in the interview that even
though they owned the IP for the product, they included the names of the contributing engineers
at Design Corp as co-inventors on the patent. Electronics Connect felt that the project strongly
challenged existing ideas, offered new ideas and offered superior value to the customer. While
the project was slightly late and over budget, this was reported as relative to the unknown and
was not seen as deterioration in performance of the project. We categorize this case in the high
level of success category.
2.4.6.6 Key points
This is a unique case where the supply chain relationship is created with an understanding
that the relationship will dissolve upon completion of the project. Whereas other supply chain
relationships are created with a long-term relationship horizon in mind, this was purposefully
designed to be a short-term relationship. This project highlights that successful interfirm
innovation performance can occur without the need to commit to a long-term relationship.

2.5 Cross-Case Analysis
2.5.1

AJAE Gap Size

A key objective of this study was to empirically validate the existence of AJAE and
investigate its impact on performance. As qualitatively highlighted in the previous section, there
in fact does exist a difference between actual levels of JA and desired levels of JA in the
interfirm new product development projects investigated in this essay. Further, we observed that
the levels of both desired JA and actual JA can vary across the different stages of the project (eg.
cases A, C and E). As a result, the degree of the asymmetry (AJAE gap size) between desired

87

and actual levels of JA also can vary across the different stages of the project (figure 5, figure 9
and figure 13). Given this observation, we recommend future studies control for the stage of the
innovation project when studying AJAE since AJAE may be more prominent in some phases of
the development project than in others.
While the interview protocol did not have a numeric scale identified on the figure, the figure
is depicted with ten equally spaced lines that the respondent used in quantifying the actual versus
desired levels of JA from low to high (figure 3). We subsequently assigned a scale of 0 (bottom
line) to 100 (top line) to assist in our analysis of the respondent scores. It is interesting to note
that in all cases where a gap exists, the respondents reported higher average desired levels of JA
as compared to the actual levels achieved. In no case did a respondent identify a level of joint
action that was higher than desired. Either the respondent felt that actual and desired levels were
aligned, or he felt that a higher level of JA was desired.
It would be of interest to also assess the suppliers’ responses to the same instrument and see
if a similar pattern persisted or if the reverse pattern is observed where suppliers felt that the
actual level of joint action was not necessary and that a lower level would have been more
appropriate. This behavior has been noted in prior research where more powerful supply chain
members can use their leverage to coerce less powerful partner firms to engage in joint action
investments where the net benefit is more favorable to the powerful partner (Hart & Saunders,
1997; Weinstein, 2005). Exploring the supplier view, unfortunately, was not possible in this
study since in only three cases were supplier contact information provided for the dyad; of those
three supplier contacts, none agreed to participate in the research. In the other three cases where
no contact information was provided, the participants hesitated to facilitate supplier participation
citing concerns that doing so could adversely impact the ongoing business relationship by

88

signaling to that supplier that something was wrong and the relationship need to be studied by an
academician. This highlights the challenging nature of obtaining dyadic empirical data from
informed respondents regarding buyer/supplier relationships. A possible extension, however, to
this research could be a replication of this study using supplier respondents rather than OEM
respondents that are not matched to the original research.
For sake of comparison in this cross-case analysis we average the scores for a given JA
dimension across all stages of the project and across both respondents. For example, respondent
one in Case A reported desired levels of JA of 50, 50, 60 and 45 across the respective stages of
idea conception, idea development, application development and launch. The average of
respondent one’s four scores for desired JA is therefore 51.25. For the same stages, respondent 2
reported desired levels of JA of 60, 70, 60 and 35 with an average of 56.25. Averaging the two
respondent scores together yields an overall desired JA score of 53.75. The same process
applied to actual levels of JA results in a combined average score of 35.00. The difference
between the averaged levels of actual JA and desired JA is reported as the average AJAE gap
size, and in this case (Case A) equals to 18.75. Table 8 reports these values for all cases and for
all respondents. Note that values above both median and average scores are highlighted with an
asterisk.
Using both average and median scores, we categorized desired JA, actual JA and AJAE gap
size into the three groupings of ‘high’, ‘low’ and ‘medium’ (table 9). A score is categorized as
‘high’ if it is above both the average and median scores. It is categorized as ‘low’ if it is below
both the average and median scores. It is categorized as ‘medium’ if it is equal to, or bounded by
the average and median scores.

89

Table 9 also includes two additional columns: ‘AJAE gap clarity’ and ‘management performance
assessment’. The scoring methodology for ‘AJAE gap clarity’ was qualitatively derived from
the interview transcript codings and is discussed in the following section. The ‘management
performance assessment’ is the performance category assigned to the project by the key
executive contact at each firm at the onset of the interview process when the original projects
were selected for the research study. Later in this essay we conduct a validation analysis (table
10) where the survey performance scores reported by key respondents during the interview
process are compared to the a priori management performance assessments listed in table 9.

90

Table 2.8: AJAE Gap Size

Respondent 1
Project

Desired
JA

Actual
JA

Case A - PEC

51.3

42.5

Case B - AVI

46.3

Case C - BI

Respondent 2
Desired
JA

Actual
JA

AJAE
Gap
Size

Desired
JA

8.8

56.3

27.5

28.8

53.8*

35.0

18.8*

45.0

1.3

21.3

21.3

0.0

33.8

33.1

0.6

67.5

50.0

17.5

72.5

67.5

5.0

70.0*

58.8*

11.3

Case D CTC/GBC

25.0

20.0

5.0

65.0

3.8

61.3

45.0

11.9

33.1*

Case E CTC/CM

67.5

43.8

23.8

NA

NA

NA

67.5*

43.8*

23.8*

42.5

40.0

2.5

NA

NA

NA

42.5

40.0*

2.5

49.4

37.5

15.0

52.1

37.1

15.0

Case F - EC

AJAE
Gap Size

Average

Median
Average
* Score is above both the median and the average for that dimension

91

Actual
JA

AJAE
Gap Size

Table 2.9: AJAE Gap Size, Gap Clarity and Performance

Project

Desired JA

Actual JA

AJAE Gap
Size

AJAE Gap
Clarity

Case A - PEC

High

Low

High

Low

Low

Case B - AVI

Low

Low

Low

High

High

Case C - BI

High

High

Low

High

High

Case D - CTC/GBC

Low

Low

High

Low

Low

Case E - CTC/CM

High

High

High

Low

Low

Case F - EC

Low

High

Low

High

High

92

Management
Performance
Assessment

Table 2.10: Performance Score Validation

Challeng
ed
Offered
Existing
New
Ideas
Ideas

Creati
ve

Value to
the
Customer

Market
Acceptan
ce

Timing

Budget

Total
Average

Survey
Performance
Score

Management
Performance
Assessment

Case A - PEC

3.5

3.5

4.0

3.5

3.5

1.0

2.5

3.1

Low

Low

Case B - AVI

5.5

6.0

5.5

4.0

3.5

4.0*

3.0*

4.5

Med

High

Case C - BI

6.5*

5.5

6.0*

7.0*

6.5*

4.0*

3.5*

5.6*

High

High

Case D CTC/GBC

6.0

6.5*

6.5*

3.0

2.5

1.0

1.5

3.9

Low

Low

Case E CTC/CM

7.0*

7.0*

7.0*

4.0

6.0*

1.0

1.0

4.7*

High

Low

Case F - EC

7.0*

7.0*

5.0

7.0*

6.0*

2.0

2.0

5.1*

High

High

Average

5.9

5.9

5.7

4.8

4.7

2.2

2.3

4.5

Median

6.3

6.3

5.8

4.0

4.8

1.5

2.3

4.6

* Score is above both the median and the average for that dimension

93

2.5.2

AJAE Gap Clarity

Despite the challenge to obtain the dyadic view on a single innovation project, we were still
able to qualitatively identify asymmetry between desired and actual levels of joint action for a
focal firm through the descriptions obtained and reported previously. Another construct that
emerged in the course of this research related to AJAE was the clarity of the gap. As highlighted
in essay one of this dissertation, AJAE arise when two firms engage in joint action and those two
firms bring with them differing marginal cost and benefit curves associated with varying levels
of shared joint action. This results in those two firms most likely differing in the optimal level of
joint action desired, thus leading to asymmetrical joint action expectations (AJAE).
As researchers, we can graphically conceptualize the gap between the respective optimal
levels of joint action when we plot both firms’ marginal cost and benefit curves; thus for the
researcher the gap can be clearly comprehended from a theoretical perspective. Applied in a
managerial context, however, the gap may not be as readily clear to those involved. A manager
with limited information may not be able to identify the existence of such gaps to the extent that
the marginal costs and benefits for engaging in joint action are unclear from either his own firm’s
perspective or from the partner firm’s perspective. If a manager has a high level of knowledge of
his own firm’s marginal cost and benefit curves yet for various reasons has little to know
knowledge of the partner firm’s marginal cost and benefit curves, then gap clarity will be low
and the manager may be unaware that the partner firm desires a differing level of joint action as
compared to what the manager’s firm desires. The greater the knowledge a manager has of these
factors, the greater the gap clarity will exist. If a manager is informed of both his own firm’s
marginal cost and benefit curves and that of the partner firm, then gap clarity can exist for that
manager.

94

We observed the construct of gap clarity in our transcripts. In some cases, it was evident that
the respondent was interested in understanding the viewpoint of the partner firm. In other cases,
lack of clarity was also evident. In case A (PEC), for example, the respondent demonstrated
frustration in his inability to establish clarity when he stated “I just wish that they would have
been a little more forthright about it and just been direct. I just felt like there was an elephant in
the room at all times, and they would still come together with us and still be nice and still talk
and still work together and still go forward… but there was still this underlying feeling of ‘we
don't trust you… we don't know what you're doing’ ”. Conversely, case B (AVI) exhibited
evidence of establishing clarity in their relationship through regular meetings at both the
operational level and at the leadership level where an executive steering committee convened to
discuss issues and opportunities. In this environment of communication, both sides were able to
help each other understand and appreciate their respective marginal costs and benefits associated
with joint action, thus establishing clarity in AJAE.
In case C (BI), the entire premise of the joint project was based on Blaze Inc’s willingness to
think out of the box and identify a business arrangement (exclusivity contract) that would be
beneficial not just to them, but also to Super Motors Corp. Much of the effort that went into
establishing the project built upon this foundation of clarity. Case D is another example with a
lack of clarity in the relationship. One respondent summarized this lack of clarity when he
stated,
‘They went off on their own to develop the new technology for that market and they get a
little bullish assuming that they… dictate the technology because they owned it, they
owned the market. They developed some unique technology that would be beneficial to
increase their revenue stream, to increase their business case and quite a few benefits for
them. They failed to think of the impact to Construction Truck Corp. and to the
customer… assumed that that would just be necessary pains for us because we didn't
have any other options, so we went out and found options.’

95

Case E appeared in all aspects to have a high degree of clarity and success associated with it
until Composite Mfg (CM) left the business after one year of production with little to no warning
leaving CTC with no option but to revert to the old technology. It became apparent to CTC that
there had been much more to the story they were not aware of pertaining to the true costs and
benefits that CM incurred for their joint product development project. In short, CM assumed
clarity existed when in fact it did not. Case F exhibited a high degree of clarity in the
relationship, partly due to the level of openness in communication referred to in the within-case
analysis of the case. We summarize in table 9 the level of gap clarity associated with each case
along with the gap size previously reported and performance for all six cases.
A gap clarity score of ‘high’ was assigned to those cases with statements coded in NVivo 9
under the category of gap clarity. A gap clarity score of ‘low’ was assigned to those cases with
no evidence of gap clarity in the coding. The overview in the preceding paragraphs summarizes
this analysis. In summary, while desired JA, actual JA and AJAE gap size were categorized as
high/low using graphical feedback generated by the respondents, ‘AJAE gap clarity’ and
‘management performance assessment’ were categorized using coded verbal evidence from the
transcripts.
2.5.3

Performance

How innovation performance is operationalized varies greatly across innovation research
literature. Some studies measure innovation performance at the firm level (Ahuja & Katila,
2001; Laursen & Salter, 2006; Luca & Atuahene-Gima, 2007) while others measure it at the
product or project level (Arora, Gambardella, Magazzini, & Pammolli, 2009; Wagner, 2012).
Studies also vary in the items used to measure innovation performance. Performance can be
measured by percent of revenue attributed to innovative products (Laursen & Salter, 2006), the

96

number of patents produced (Ahuja & Katila, 2001), or financial performance derived from
innovative products (Luca & Atuahene-Gima, 2007). Goodale, Kuratko, Hornsby, and Covin
(2011) used a multi-dimensional approach that combined multiple factors into a single
importance-weighted score.
In a like manner, we measure innovation performance across three key dimensions relevant
to the innovation project: 1) innovativeness (challenged existing ideas, offered new ideas and
creative), 2) market performance (value to the customer, market acceptance) and 3) project
execution (timing, budget). We average the scores across all seven items into a single composite
score for innovation performance. Whereas Goodale et al. (2011) employed a weighting scheme
using respondent-provided weights, we assume equal weighting for all items and average the
scores for each case (table 10). Additionally, those scores above both median and average are
highlighted with an asterisk. Finally, the ‘total average’ scores are categorized as either high or
low given the respective score’s position as either above or below both average and median
scores. In the one case (Case B) where the ‘total average’ score is exactly equal to the overall
average, we categorized it as ‘medium’.
Given the qualitative nature of this research, we do not assert any statistical significance to
this analysis but use the scoring mechanism simply to assist the research team to identify both
themes across cases and outliers that emerge. In particular, we were interested to establish some
level of validation to the a priori qualitative assessment of performance provided by the
executive contact when cases were initially identified. This a priori assessment is reported in the
column titled ‘management performance assessment’.
We now compare this a priori assessment to the performance scores collected from the key
informants during the interviews. Cases A, C, D and F all align in the two respective

97

performance assessments (table 10); namely cases A and D are considered as low-performing
while cases D and F are considered as high-performing. Case B average survey performance
score matches the overall average, thus we assign it a ‘medium’ rating for performance. This is
not a concern for validity since the project still scores above both average and median scores on
the execution dimensions of timing and budget while scoring near the average/median scores for
the remaining dimensions.
The performance data from Case E, however, does provide us initial cause for concern. The
project was identified by the executive contact as a lower-performing example while the survey
responses state otherwise (table 10). Review of the transcripts for this case highlights that the
project did, in fact, develop and launch as a success. Had the research team conducted the
interview within the first few months after the initial launch, most likely the executive contact
would have categorized case E as a high-performing example of success. Yet the subsequent
collapse of the supply chain for the product after the first year of production rendered the overall
project a lower performance category.
It is important to note that the measures used in the survey instrument focus only on the
project development and initial launch stages of case E and capture the higher degree of success
achieved during this time. The measures do not, however, capture the persistence of
performance beyond launch. Our research specifically bounded the domain of applicability to
the project development and initial launch phase. In fact, the initial request for participation
stated that the research team was interested in studying projects that had recently launched.
However, given this discrepancy observed in case E, follow-up emails were sent to all other
firms that participated in the study to ascertain if any material differences in performance

98

emerged over the previous 12 to 21 months since the interviews were conducted. No notable
discrepancies were reported.
Another limitation in this study is potential common method bias of the reported
performance dimensions. Unfortunately, due to the limited nature of published performance data
at the project level, our study relied solely on self-reported measures of project performance.
Ideally, one would want to verify project performance against an independent data source, but
this was not possible with the cases and performance dimensions studied in this project. At a
minimum, however, we relied on multiple sources within the firms in order to generate an
acceptable level of reliability on the data gathered. Also, the two methods of collecting
performance data as just discussed previously provided an increased level of reliability as
opposed to only using the initial management assessment or relying solely on the survey
response scores from the interview protocol.

2.6 Propositions and Refined Model
Inspection of table 9 reveals a few important patterns across the constructs. First, as
expected, gap size appears to be inversely related to performance. Large gaps are associated
with lower levels of performance (cases A, D and E) while smaller gaps are associated with
higher levels of performance (cases B, C and F). This relationship was conceptually predicted in
the a-priori theoretical model and is supported by this qualitative empirical study. Second, gap
clarity appears to be positively related to overall performance. When interfirm relationships
establish clarity of AJAE, they appear to achieve higher levels of performance (cases B, C and F)
as compared to relationships that have lower levels of clarity (cases A, D and E). This finding
was not predicted in the original conceptual model and is subsequently added as an additional
dimension relevant to establishing behavioral interoperability. Third, absolute JA level also

99

appears to be weakly related to overall performance. Cases A, B, C and D exhibit matching
levels of performance with absolute levels of JA. Cases E and F, however, are the exceptions
where an inverse relationship exists between absolute level of JA and the level of performance.
In case E, both firms engaged in very high levels of JA yet the project ultimately failed. Case F,
on the other hand, had lower levels of JA yet was seen as a highly successful joint development
project. This finding is interesting in that it supports the premise that while JA in interfirm
relationships (as often operationalized as the level of collaboration, integration or cooperation
between two firms) is an import factor of success, perhaps even more relevant to predicting
performance success are the behavioral factors related to expectations that give context to the
relationship engaged in JA. In this research we identify both AJAE gap size and AJAE gap
clarity as import behavioral factors. This leads us to modify the original proposition 3 as
follows:
Original Proposition
Proposition 3: Gap size in joint action expectations is inversely associated with
innovation performance (smaller gaps in joint action expectations correlate with higher
levels of innovation performance)
Revised Proposition
Proposition 3a: Interfirm behavioral interoperability is significantly and directly
associated with innovation performance. (higher levels of behavioral interoperability are
directly associated with higher levels of innovation performance)

Proposition 3b: Both AJAE gap size and AJAE gap clarity are reflective indicators of the
behavioral interoperability construct

100

Figure 2.17: Revised Model

Prop 5a, 5b

PROCESS
INTEROP

Prop 2

BEHAVIOR
INTEROP
(AJAE GAP SIZE)
(AJAE GAP
CLARITY)

Prop 3b
Prop 1

RESOURCE
INTEROP

Prop 4a, 4b

101

Prop 3a

INNOVATION
PERFORM

2.7 Limitations and Conclusions
It is important to note that the size of the gaps identified in this study may or may not be
similar to the size of the theoretical gaps between the two firms’ optimal levels of JA as depicted
*

by ∆J in figure 18.
Figure 2.18:: Graphical Model of Asymmetrical Joint Action Expectations

MC - marginal cost curve
MB - marginal benefit curve
J - Level of joint action
*

J - Optimal level of joint action where marginal costs equal marginal benefits for a given firm

In this study, we observe gaps between actual and desired levels of JA from a focal
firm/OEM perspective. We make the assumption tthat
hat these firms seek to maximize utility and
therefore engage in a level of joint action that is either equal to, or bounded by the respective

102

*

*

theoretically optimal levels of joint action (J B and J A). Referring to figure 18, we expect the
firms to engage in a level of joint action (J) that is defined by:

*

*

J B≤J≤J A

Equation 8

Further, since all gaps identified in the case studies reported a desired level of joint action
(J*) that is higher than the actual level of joint action experienced, we can assume that the focal
firm interviewed is represented by firm ‘A’ (figure 18), whose marginal cost and benefit curves
intersect at a higher level of joint action (J*A) than that for firm ‘B’ (J*B). The challenge,
however, is that we have only uncovered the gap from a single perspective. Two alternative
situations therefore exist if we were to obtain the dyadic perspective: 1) the actual level of joint
action engaged in is in fact optimal for the supplier and the supplier is not willing to engage in
any higher levels of JA, in other words:

*

J=J B

Equation 9

or 2) the actual level of joint action engaged in is greater than the optimal and desired level of JA
for the supplier and that the actual level of JA engaged in is somewhere between the two optimal
levels of JA for the two firms:

*

*

J B<J<J A

Equation 10

103

Either case is feasible and while we know a gap exists from the perspective of one firm, we don’t
yet have the complete picture that allows us to accurately assign a graphical representation of the
gap in each case.
One could argue that it is possible that two firms report AJAE gaps where both parties desire
higher levels of JA. This situation could arise in the case that both firms share differing views on
the forms of higher levels of JA. For example, one firm could desire a greater number of
resident engineers assigned by both firms to the project while the other firm desires an increase
in joint marketing efforts by both firms. Each firm desires their form of joint action while
possibly not being aware of the other form desired by their partner. This hypothetical case is a
clear example of a relationship that has not yet achieved consensus on the forms of joint action
available to their partnership; in short, the definition of the J axis in figure 18 has not yet been
clearly established. The graphical model, therefore helps us identify another important aspect to
understanding clarity in AJAE: namely, firms need to establish a mutual understanding of what
forms of JA are possible and feasible in the relationship in order to better establish clarity in joint
action expectations.
The construct of asymmetrical joint action expectations (AJAE) clarity captures the extent to
which both firms have visibility and understanding of each others’ respective costs and benefits
for the shared joint action. Establishing clarity is not an easy endeavor, especially if the two
firms do not share key underlying assumptions that guide day-to-day decision-making that define
their respective corporate culture (Schein, 1985). In the presence of clarity, however, firms will
be more likely to identify and embrace forms of joint action that are beneficial to both firms,
rather than to simply maximize benefits to itself. Additionally, gap clarity will enable a firm to
more easily identify investment opportunities that may beneficially shift the cost and benefit

104

curves of both partners, or allow the partners to increase the overall level of JA to a more
economically optimal level for both parties.
A further study of AJAE can be aided by a few extant constructs already researched in the
supply chain literature. Team orientation is one such construct that seeks to capture the extent
that a team spirit pervades, commonality of purpose exists and visions are shared (Hult, Ketchen
Jr, & Nichols Jr, 2003). Additionally, supply chain orientation is another similar construct that
emphasizes the need for a firm to reach beyond its own paradigm (Mentzer et al., 2001) and
adopt a systemic/holistic view of the supply chain (Esper, Defee, & Mentzer, 2010). Another
consideration for future studies is to control for the length and history of the relationship. Such a
control variable can help segregate any potential variance in performance due to established
relationships.
This qualitative research has uncovered numerous apparently relevant aspects of buyer
supplier relationships that impact innovation performance in joint new product development
initiatives. Whereas prior research focused on the absolute level of joint action engaged between
firms, this essay elevates the importance of the behavioral dimension and role that joint action
expectations have on performance. In particular we identified AJAE gap size and AJAE gap
clarity as relevant. We have only begun to understand the role that understanding and managing
joint action expectations has in establishing effective interfirm relations. Whether or not a gap
exists, whether or not that gap is transparent to both parties, and whether or not (and how) firms
seek to close such gaps are all separate, yet important concepts that can influence the
performance of joint action efforts. These questions offer excellent opportunities for further
research that will enrich our understanding of effective supply chain management.

105

3 ESSAY THREE - ASYMMETRICAL JOINT ACTION
EXPECTATIONS, SUPPLY CHAIN INTEROPERABILITY
AND INNOVATION PERFORMANCE: AN EMPIRICAL
ANALYSIS OF INTERFIRM NPD POWERTRAIN PROJECTS
IN THE AUTOMOTIVE INDUSTRY
3.1 Introduction
Using literature review and case study research, the first two essays of this dissertation
have established a case for asymmetrical joint action expectations (AJAE) as a valid
phenomenon in inter firm relations and as relevant to innovation performance in the supply
chain. Essay one laid the theoretical foundation of understanding how differing marginal cost
and benefit curves associated with joint action between two firms can lead to differing optimal
levels and therefore differing expectations regarding the desired level of joint action across those
firms. Essay two investigated this phenomenon in the context of six case studies of interfirm
new product development projects and found that indeed, gaps between actual and desired levels
of joint action do exist. The case studies support the premise that AJAE exist and are relevant to
project innovation performance from a financial, design and timing perspective. Further, the
case studies reveal that not only is the size of the gap in AJAE relevant, but so also is the clarity
of those expectations.
While research has already been done on how firms can successfully integrate in
interdependent supply chain relationships, the research has mostly focused on the effective
management of resources and processes across supply chain members (Frohlic and Westbrook,
2001). This research, however, highlights that behavioral considerations, in particular joint
action expectations, are also relevant in establishing effective supply chain relationships in the
context of innovation performance.

106

Firms have the choice to go it alone (make), outsource completely (buy), or engage in a
partnership (ally) in pursing innovation initiatives. Transaction cost economics broaches the
question of how firms define their boundaries and associated governance structures in terms of
the make/buy/ally decision (Coase, 1937; Williamson, 1979). The choice between these options
is largely dependent on the respective costs and benefits a firm could derive by pursuing each of
the three options. The ability (or inability) a firm has to effectively manage an ally relationship
has a critical bearing on the strategic options that firm can pursue. Lack of ability to manage
interfirm relationships effectively renders the ally option unavailable and limits the strategic
toolbox available to that firm. Worse yet, firms could overestimate their ability to pursue an ally
option, invest time and resources accordingly to only come up empty handed.
Research has reported that approximately 50% of alliances do not achieve desired outcomes
(Parise & Casher, 2003; Park & Ungson, 2001). This high incidence rate of failed alliances is a
testament to the prevalence of this strategic fallacy. Alternately, competence in managing
interfirm ally relationships can expand the strategic toolbox available to a firm and provide that
firm with options and strategic avenues not easily accessible by its competitors, thus creating a
potential competitive advantage. This research focuses on the ally scenario where a firm engages
an external business partner that is directly associated with the focal firm in the supply chain
(Williamson, 2008). The research question driving this inquiry is: what role does AJAE play in
achieving innovation performance within supply chain partnerships (ally)?
Interoperability was previously introduced as a construct to describe the ability of entities
to effectively interface with each other over time. Originally explored in both military and IT
literature, interoperability is extended in this dissertation to the field of supply chain management
in order to examine the research question. This dissertation categorizes supply chain

107

interoperability into three categories: resource interoperability, process interoperability and
behavioral interoperability, and hypothesizes that managing AJAE is both relevant to
establishing behavioral interoperability within supply chain relationships and impacts innovation
performance.
In this third essay of the dissertation, I empirically evaluate the hypothesized model, as
presented at the end of Essay Two, using a secondary data set collected on 67 powertrain product
R&D development projects. The model posits that all three categories of interoperability are
relevant to innovation performance and that behavioral interoperability mediates the influence of
resource and process interoperability on innovation performance. Using structural equation
modeling, we find statistical support for the hypothesis that behavioral interoperability (as
operationalized by AJAE gap size and clarity) has a statistically significant positive relationship
with innovation performance. In particular, smaller sized gaps in AJAE and higher levels of
AJAE clarity are both associated with increased levels of innovation performance.
Clarity, a construct not originally considered in Essay One, emerged in Essay Two as a
relevant dimension of AJAE; it is confirmed in this study as having a statistical relationship with
innovation performance. While the analysis in this Essay does not confirm the hypothesized
mediating relationship of behavioral interoperability, it does open the possibility that these
distinct dimensions of interoperability may not be as interdependent as originally hypothesized.
This finding is significant in that it suggest that firms can invest in behavioral interoperability
independent of the state of the other dimensions of interoperability. The analysis also finds that
resource interoperability has a statistically significant positive relationship with behavioral
interoperability.

108

These results are relevant to firms and managers seeking to maximize the innovation
potential that resides within their supply chain. Analysis of the secondary data set supports the
argument that firms that invest time and effort in establishing clarity of expectations by
explicating the costs and benefits of joint action from both their own point of view and from that
of the partner firm can achieve a higher level of interoperability and therefore an improved level
innovation performance as compared to those firms that don’t. These findings support the notion
that effective supply chain management for innovation requires that firms invest in
synchronizing not only resources and processes, but also expectations. Doing so enhances the
feasibility of the ally option and expands the strategic innovation toolbox available to a firm.
This research contributes to both the interfirm relationship literature and the innovation
literature. While much of the current R&D literature explores innovation performance at the
firm level (Calantone, Harmancioglu, & Droge, 2010), this dissertation approaches innovation
performance at the interfirm supply chain dyad level. Given the significant pressures that exist
in the automotive industry to improve fuel economy, the use of powertrain-related R&D project
data provides a salient context from which these hypotheses can be tested.

Additionally, the

trend of R&D intensity shifting from OEM’s to Tier 1 suppliers increases the impetus to
understand effective interfirm innovation performance management.
The following section expands on the theoretical construct of interoperability and its
relation with AJAE. First, a summary of the hypothesized model developed in the first two
essays is presented. Then the statistical analysis of the secondary data set used for this study is
summarized. Finally, this essay concludes with a results discussion section and highlights both
limitations to the study and opportunities for extension.

109

3.2 Interoperability
Firms that engage in innovation have a strategic choice to keep the work in-house,
leveraging their own resources, or to leverage the resources of external entities in what is
sometimes called open innovation (Chesbrough, 2003). These external sources can be single or
multiple entities and can be found as either existing members of the supply chain, or as nonsupply chain related entities such as a universities or government agencies. Our study centers on
the configuration of two interdependent partner firms (buyer-supplier dyad) that share a common
supply chain for which the innovation is intended. Understanding the nature of how such firms
can successfully position themselves to interact in order to achieve the desired innovation
outcomes is of critical importance to managers. Interoperability is an existing construct already
used in other fields of study (military and IT) that addresses the effectiveness of how
interdependent entities prepare themselves to successfully interact over time (Ford, Colombi,
Jacques, & Graham, 2009; LaVean, 1980; Otjacques, Hitzelberger, & Fernand, 2007). We adopt
this concept as a starting point and extend it to the field of supply chain management. To the
best of our knowledge, this is the first study to extend the concept of interoperability to predict
innovative outcomes in buyer-supplier supply chain relationships.
3.2.1

Interoperability in military and IT fields of research

The discussion on interoperability first emerged in the military after experiencing
coordination failures between branches of the armed services during the Vietnam War (Ford et
al., 2007; LaVean, 1980). The discussion continued over the ensuing decades as a key topic of
planning and strategy in military organizations around the world (Ford et al., 2009). While
numerous definitions for interoperability emerged over this time, the most common definition
that researchers converged on defines interoperability as: “The ability of systems, units, or forces

110

to provide services to and accept services from other systems, units, or forces and to use the
services so exchanged to enable them to operate effectively together” (Ford et al., 2007, pg.6).
In military applications, interoperability is concerned with enabling different groups (e.g.
branches of the military) to effectively combine efforts in achieving specific military objectives.
Such efforts require a focus on both the resources (physical systems) and the procedures used to
integrate those resources. In fact, these two dimensions, resources and processes are a common
theme across much of the extant literature on interoperability. In recent military research,
interoperability is operationalized using a measure of sameness across both friendly and hostile
interactions (Ford et al., 2009). This concept of sameness underscores the importance of
considering the matching qualities of the respective sides’ systems and processes.
Interoperability subsequently emerged in the IT literature where the ability for commercial IT
systems across firms to ‘provide services to and accept services from other systems’ is a
fundamental determinant of success (Mouzakitis, Sourouni, & Askounis, 2009; Otjacques et al.,
2007). The establishment of clearly defined and adopted communication protocols such as
Transmission Control Protocol (TCP), Internet Protocol (IP) and HypterText Markup Language
(HTML) has been instrumental in the establishment of the Internet as an agent of IT
interoperability and the ensuing fundamental changes that emerged from a global society linked
by the Internet (Friedman, 2005). This example alone underscores the transformative power that
resides within the construct of interoperability.
IT interoperability emerges when both resource and process dimensions are purposefully
designed and implemented to allow two or more firms to interact effectively from an information
technology standpoint. Again, the dimensions of both resources and processes are instrumental
in establishing IT interoperability. The resource dimension includes computing and

111

communication hardware and software while the process dimension relates to business processes
needed to allow those hardware and software systems to interact with the systems of other
entities. Not only must firms use compatible files and software to exchange electronic data, but
they must also know when to send what information and how that information needs to be used
for decision-making.
3.2.2

Interoperability in supply chain management

Supply chain management also is concerned with effectively managing resources and
processes across entity boundaries in order to achieve operational success (Frohlich &
Westbrook, 2001). IT interoperability plays a critical role in establishing interoperability across
firms in the supply chain (communication, scheduling and tracking systems) (Mouzakitis et al.,
2009). However IT interoperability alone is not sufficient to ensure that parts and data arrive at
the right place, at the right time and at the right price. Effective operational integration with
supply chain partners requires the interoperability of other resource categories including
transportation equipment and infrastructure, packaging design and material, testing equipment
and compatible part interface designs. With respect to the process dimension of interoperability,
supply chain management needs to synchronize long-term capacity planning, sales forecasting,
production scheduling, materials requirement planning, supply chain risk contingency planning,
and new product development (NPD) initiatives. Achieving resource and process
interoperability, even amongst a buyer-supplier dyad in a supply chain, is no small feat!
The importance of the resource and process dimensions is further highlighted by Frohlic and
Westbrook (2001), who operationalized supply chain integration with eight key activities
(Frohlich & Westbrook, 2001). Referring to the listed integration activities, Frohlic and
Westbrook (2001, p. 197) asked informants to rate “to what extent do you organizationally

112

integrate activities with your suppliers and customers?” The eight factors are listed below in
Table 1 along with the relevant interoperability dimension. Upon inspection of Table 1, it can be
seen that these eight factors reflect both resource and process interoperability. The resource
interoperability dimension is reflected in the activities of integrating joint EDI access/networks,
packaging customization, using common logistical equipment/containers and common thirdparty logistics services; while the process dimension is reflected in the activities of integrating
shared access to planning systems, sharing production plans, integrating knowledge of inventory
mix/levels and integrating delivery frequencies.

113

Table 3.1: Supply Chain Interoperability
Relevant
Frohlic and Westbrook (2001)

Interoperabiity

Integration Activity

Dimension

Access to planning systems

Process

Sharing production plans

Process

Joint EDI access/networks

Resource

Knowledge of inventory mix/levels

Process

Packaging customization

Resource

Delivery frequencies

Process

Common logistical equipment/containers Resource
Common use of third-party logistics

Resource

First, we note that integration and interoperability are distinct, yet related constructs.
Interoperability refers to the ability of systems to interact with other systems (Ford et al., 2007,
pg.6), while integration, however, refers to the extent that firms actually coordinate with each
other in order to act as a single company (Frohlich & Westbrook, 2001, pg. 187). Two firms that
have established a high degree of interoperability simply have the potential to achieve a high
level of integration; in essence interoperability refers to the latent integration capacity of two
systems. Firms that invest in integration, it can be argued, should first address interoperability
before they engage in integration activities; otherwise, those integration efforts may not yield the
desired results. As such, the categories relevant to interoperability are also relevant to
integration.
The activities explored in Frohlic and Westbrook (2001) relate mostly to a product
manufacturing setting where the product has already launched and is in a state of ongoing
production. In the context of the Supply Chain Operations Reference (SCOR®) model (Council,
114

2008), a widely used framework that organizes key supply chain activities (Zhou, Benton,
Schilling, & Milligan, 2011), the Frohlic and Westbrook (2001) activities relate mostly to the
“make” and “deliver” categories. Yet, effective supply chain management also calls for
interfirm participation in the “source” and “plan” categories, both of which are not addressed by
Frohlic and Westbrook (2001).
If the interfirm relationship is strategically limited to only an operational context, then the
Frohlic and Westbrook (2001) framework may be sufficient. However, if the context of the
interfirm relationship extends into the strategic realm of interfirm joint action involving new
product development, then a richer model of interoperability is needed that accounts for the
complex human interactions involved in the ideation and problem-solving activities involved in
the work of innovation (Mohr & Spekman, 1994). Behavioral interoperability, we posit, is a
third interoperability dimension that should be considered in establishing a more complete model
of the effective supply chain interoperability needed to achieve innovation performance.
The question that then arises is ‘what entails behavioral interoperability’? In other words,
what behavioral aspects of the interfirm relationship either enhance or hinder performance? Are
there basic beliefs and norms of doing business that, when matched together across two firms,
facilitate the efficient and effective flow of ideas, information and solutions? Behavioral
interoperability may in fact require the blending of a broad set of beliefs and norms as opposed
to any single dimension. Research already exists regarding specific behavioral aspects of B2B
relationships that can be synthesized and built upon in order to investigate the question of ‘what
entails behavioral interoperability’. These potential areas of research are listed in Table 2.

115

Table 3.2: Research Streams Potentially Related to Interoperability
Research Streams
Shared understanding of procedural and distributive justice

Shared vision of the desired outcomes of the relationship
Financial transparency in the relationship
Corporate culture
Risk profile
Trust

Joint action

Citation(s)
(Griffith, Harvey, & Lusch,
2006; Hofer, Knemeyer, &
Murphy, 2012)
(Melnyk, Davis, Spekman, &
Sandor, 2010)
(Lipparini, Lorenzoni, &
Ferriani, 2014)
(Ettlie, 2007; Oliver, 1997;
Schein, 1985)
(Lee & Johnson, 2010)
(Fawcett, Jones, & Fawcett,
2012; Moldoveanu & Baum,
2011)
(Gulati & Sytch, 2007; Heide
& John, 1990)

Each of the items listed above merits extensive discussion and research in the context of
interfirm relations and may be considered as potential indicators of behavioral interoperability as
a reflective construct. However, for this study we narrow our focus to that of the final item: the
role of joint action expectations and its relation to innovation performance.

3.3 Hypothesized model
Essay One of this dissertation established a theoretical model for understanding how joint
action expectations emerge in interfirm relationships given the costs and benefits associated with
that joint action. Since each firm brings with it a unique history of past investments, resources
and capabilities, it is likely that those firms have different marginal cost and benefit curves
associated with various levels of joint action within the partnership. This is likely to result in
two differing optimal levels of joint action for the two firms thus leading to asymmetrical joint
action expectations (AJAE). The distance between the respective desired levels of joint action

116

across two firms is labeled as the AJAE gap size. Essay One posited that AJAE gap size is
relevant to predicting innovation performance in interfirm NPD projects.
Essay Two confirmed the existence and relevance of AJAE gap size in the context of six case
studies of NPD projects involving a buyer-supplier relationship. In addition to AJAE gap size,
the case studies also revealed the construct of clarity as relevant to understanding AJAE. In
short, some cases exhibited a higher level of clarity of joint action expectations as compared to
others. The firms involved with higher levels of clarity appeared to understand not only their
own company’s costs and benefits associated with joint action, but they also sought to
understand the partner firm’s costs and benefits as well, thus raising an awareness of the
existence of AJAE in the relationship. Partnerships where clarity was evident exhibited higher
levels of performance as opposed to those that did not. We place AJAE in the broader context of
behavioral interoperability and hypothesize that AJAE gap size and AJAE gap clarity are both
reflective indicators of the broader behavioral interoperability construct.

HYPOTHESIS 1: Both AJAE gap size and AJAE gap clarity are reflective indicators of the
interfirm behavioral interoperability construct

As reflective indicators of behavioral interoperability, we posit that AJAE gap size will be
inversely related to performance. In other words, smaller gap size is expected to correlate with
higher innovation performance. AJAE gap clarity, however, is expected to correlate directly
with innovation performance. Given that these two measures are reflective indicators of
behavioral interoperability, we anticipate that higher levels of behavioral interoperability will be
positively related to improved innovation performance.

117

Buyer-supplier partnerships that engage in joint NPD initiatives take their business
relationship to a higher level of human interaction as compared to an operationally-focused
relationship where make and deliver are the sole focus of activity. Operational activities
generally focus on the effective and efficient production and delivery of goods and services,
where roles, responsibilities and the sharing of rents are contractually defined. Joint NPD
initiatives, however, require numerous additional problem-solving activities that are not so
readily anticipated and therefore not contractually defined. Navigating successfully through this
oft times ambiguous process requires a higher degree of assumptions and trust regarding the
roles, responsibilities and behavior of the other partner firm. Partner firms that establish higher
levels of behavioral interoperability are those that have aligned assumptions and behavioral
norms that facilitate effective and efficient problem solving so that the supply chain can offer
products and services that satisfy customer needs profitably.

HYPOTHESIS 2: Interfirm behavioral interoperability is significantly and directly associated
with innovation performance. (higher levels of behavioral interoperability are directly associated
with higher levels of innovation performance)

Resource interoperability is achieved when the resources of both firms enable those firms to
successfully interact. In the absence of resource interoperability, there is less likelihood that a
supply chain partnership can generate rents sufficient to justify the relationship. The relational
view of the firm (Dyer & Singh, 1998), a theory related to the resource-based view of the firm
(Barney, 1991; Wernerfelt, 1984), argues that supply chain partners can jointly create or manage
rent-generating resources across firm boundaries thus creating an interorganizational competitive

118

advantage. We posit that resource interoperability therefore has a positive impact on interfirm
innovation performance.

HYPOTHESIS 3: Interfirm resource interoperability is significantly and directly associated with
innovation performance (higher levels of resource interoperability are directly associated with
higher levels of innovation performance)

Firms that invest in interoperable resources have a foundation on which more complex
interactions can take place. As stated previously, behavioral considerations become a more
prominent determinant of success as complexity of interactions increase. We therefore postulate
that interfirm relationships that share interoperable resources are also more likely to share
interoperable behavioral norms as well.

HYPOTHESIS 4: Interfirm resource interoperability is significantly and directly associated with
behavioral interoperability (higher levels of resource interoperability are directly associated with
higher levels of behavioral interoperability)

Process interoperability reflects the ability for interfirm relationships to successfully integrate
processes across firm boundaries. Prior research has supported the premise that interfirm
relations that so do, generally achieve improved performance (Das & Narasimhan, 2001; Das et
al., 2006; Perols, Zimmermann, & Kortmann, 2013) including improved innovation performance
(Perols et al., 2013). We therefore hypothesize that increased process interoperability will be
associated with increased levels of innovation performance.

119

HYPOTHESIS 5: Interfirm process interoperability is significantly and directly associated with
innovation performance (higher levels of process interoperability are directly associated with
higher levels of innovation performance)

In a similar manner to firms with higher levels of resource interoperability, firms that have a
higher level of process interoperability also have a stronger foundation upon which more
complex interactions can build upon. Firms that achieve higher levels of process interoperability
can focus further relationship enhancing initiatives on behavioral dimensions of the relationship,
rather than focusing on operational short falls.

HYPOTHESIS 6: Interfirm process interoperability is significantly and directly associated with
behavioral interoperability (higher levels of process interoperability are directly associated with
higher levels of behavioral interoperability)

Given the role that human interaction has in successful problem solving initiatives, we
anticipate that behavioral interoperability holds the key to achieving interfirm innovation
performance. In consequence, we anticipate that the positive relationship that both process and
resource interoperability have with improved innovation performance will be mediated by
behavioral interoperability.

HYPOTHESIS 7: Interfirm behavioral interoperability significantly mediates the relationship
between resource interoperability and innovation performance.

120

HYPOTHESIS 8: Interfirm behavioral interoperability significantly mediates the relationship
between process interoperability and innovation performance

121

Figure 3.1: Hypothesized Model

H: 5
H: 8
PROCESS
INTEROP

H: 6
BEHAVIOR
INTEROP
(AJAE GAP SIZE)
(AJAE GAP
CLARITY)

H: 2

H: 1
H: 4

RESOURCE
INTEROP
H: 3
H: 7

122

INNOVATION
PERFORM

The option to co-develop with a supply chain partner (ally) introduces numerous
complicating factors at the boundaries of the firm that would not be encountered in either the
make or buy governance structures (Coase, 1937; Geyskens, Steenkamp, & Kumar, 2006;
Williamson, 1979). The hypothesized model categorizes these relational complexities into the
groupings of resources, processes and behavior. The model presented herein offers a possible
explanation for how such interfirm relationships can maximize the potential of working together,
thus enabling the ally option as a feasible and valuable strategic tool for supply chain managers.

3.4 Research methodology and data analysis
The US automotive industry is undergoing significant transformations in powertrain
technology due to the simultaneous pressures of increasing global competition, improving
alternative fuels technology, and increasing governmental efficiency standards (Berggren,
Magnusson, & Sushandoyo, 2009). These converging pressures result in a unique business
environment where automotive companies must constantly engage in innovation initiatives.
Given limited resources, OEM’s increasingly rely on tier one suppliers for innovation. Evidence
of this exists in the shifting locus of innovation (as measured by R&D intensity) from OEMs to
tier one suppliers in the automotive industry (Rothenberg & Ettlie, 2011).
These factors make innovation in powertrain technologies a salient context to test the
theoretical model presented in this essay. Dr. John Ettlie, co-chair of this dissertation, is also a
member of a separate research team that conducted a survey of R&D powertrain projects. Dr.
Ettlie provided access to a non-published portion of this data set for the purpose of this present
study. This access was provided given the relevance of the non-published portion to the
theoretical interoperability constructs presented in the prior section of this essay.

123

3.4.1

CB-SEM (with bootstrapping)

Interoperability cannot be directly measured. We therefore must rely on reflective measures
that can indicate the presence of interoperability. Testing the research hypotheses therefore
requires a methodology that enables the simultaneous analyses of multiple constructs as
measured using reflective indicators. Covariance-based structural equation modeling (CB-SEM)
is a methodology that enables this type of analysis and is therefore used for this study. This
methodology is appropriate for two reasons. First, our model comprises numerous latent
constructs that are neither directly observable nor measurable. Instead, we use observable and
measureable reflective indicators that allow us to test the relationship between the latent
constructs in our model. CB-SEM allows for such a situation. Second, CB-SEM allows for
testing competing models where the fit indices of each model can be compared. This feature is
necessary for testing the hypothesized mediating relationships according to the Baron and Kenny
(1986) mediation test guidelines. The bootstrap analysis allows testing the statistical power of
the data set.
Partial least squares structural equation modeling (PLS-SEM) is another competing
methodology that was also considered for this study. While PLS-SEM is often used when
sample sizes are small, as is the case in this study, the methodology does not provide the needed
fit indices required to adequately test our hypotheses, in particular the mediation tests. Rather,
PLS-SEM requires a high degree of theoretical confidence in the structure of the hypothesized
model and therefore focuses its explanatory power in the strength and statistical significance of
the paths in the model. Given the nascent nature of the theories presented in this dissertation and
the necessary importance of the fit indices needed to assess our model, CB-SEM becomes the
preferred approach for this study.

124

3.4.2

Secondary data set

The secondary data set used in this analysis comes from a survey conducted by a separate
research team on powertrain-related development projects in the US automotive industry. The
data was collected in late 2010 and 2011 and was part of a larger study comparing the U.S. and
China across numerous industries.
The powertrain portion of the study was chosen since it focused on the interfirm innovation
initiatives that are of interest to our research question. Further, only the U.S. portion of the data
was analyzed since, according to the original research team, the China portion of the data
suffered from a lack of validity given the unwillingness of the informants to elaborate on openended questions. Additionally, the uniformity of answers across items and informants added to
further validity concerns regarding the China data. The U.S. portion of the study, however, did
not suffer from these issues2.
The survey items included in the powertrain study combine both five-point Likert scale items
and open-ended questions to gather data at both the company level and at the project level. Since
our research focuses on the interfirm relationship and its impact on innovation performance at
the project level, we narrow our focus of the data to just those survey items where the specific
interfirm relationship and project are the units of analysis. In all, the data set described 76
unique powertrain-related projects in the automobile industry.
Informants of the study included engineering and R&D managers, directors and a few vice
presidents. While the original data set included measures for 50 items, 11 of these items were
identified as relevant to our hypothesized model and as potential reflective indicators for the

2

The University of Michigan Transportation Research Institute – Automotive Analysis
Division, led collection of the powertrain data set and the survey was partially funded by NSF
grant number 0725056.
125

constructs of innovation performance, resource interoperability, process interoperability and
behavioral interoperability.
The sample size in this study of n=76 is not ideal for testing our model. An oft-cited
guideline for conducting CB-SEM is a minimum of n=10 for every measured variable in the
model (Nunnally, Bernstein, & Berge, 1967). Anderson and Gerbing (1988) reported unbiased
estimates at sample sizes as low as n=50, however, recommended a sample size of n=150 or
more in order to avoid large error estimates. Bentler and Chou (1987) reported that the ratio of
sample size to free parameters could be as low as 5:1, but that higher sample sizes are
recommended to achieve trustworthy significance testing of the estimated parameters. The
sample size of this study is above the minimum guidelines reported by both (Bentler & Chou,
1987) and Anderson and Gerbing (1988) yet is low enough to cause concern regarding the
statistical power of any results. Bootstrapping can be used to help assess statistical power and is
therefore applied to this study as a post-test to the CB-SEM analysis (Mooney & Duval, 1993).
3.4.3

Operationalizing the constructs

Given the nature of this secondary data analysis, the original survey was not specifically
designed with the purpose of testing interoperability constructs, yet sufficient scale items were
identified that were relevant to the constructs of our hypothesized model. This enables us to test
our hypothesis using this existing data set in an industry (automotive) where innovation in
powertrain technologies has become a de facto differentiator in competitive advantage. Effective
buyer-supplier relationship management in this setting is key in determining success, especially
since the locus of innovation, as measured by R&D intensity, has systemically been shifting from
OEMs to tier-one suppliers (Rothenberg & Ettlie, 2011). As a result, we expect this data set to
be highly relevant in the context of our research question.

126

We operationalize the constructs of our model using these data and conduct a confirmatory
factor analysis before proceeding to test the hypothesized path model. While parts of the original
data set have been used in other research currently under review for publication, the items used
to operationalize the three dimensions of interoperability in this study are not part of any other
study and are newly considered in this research. The only items used in this study that have also
been used in other research are the innovation performance items, which are the dependent
variables in our study.
Behavioral interoperability is operationalized using coded qualitative responses from the
survey. Findings from the case study portion of this dissertation (Essay Two) were initially
shared with members of the powertrain study research team and also presented at the 2013
Decision Sciences Institute conference. In particular, the relevance of both AJAE gap size and
AJAE gap clarity as measures of behavioral interoperability were highlighted as key emerging
topics. The powertrain study research team subsequently coded all open-ended responses from
the interviews for these two concepts and made their coded results available to our current study
for this dissertation. Inter rater reliability correlation scores for the coding of AJAE gap size and
AJAE gap clarity were reported as .932 and .827, respectively. We expect that these two
measures be negatively correlated to each other and that both are appropriate indicators of
behavioral interoperability. Smaller gap size reflects increased behavior interoperability while
higher levels of clarity reflect increased behavior interoperability. This will be assessed in the
construct reliability scores obtained in analysis of the measurement model.
Process interoperability is operationalized with three items from the survey: 1) difficulty of
coordinating the development of the project with customers, 2) difficulty of coordinating the
development of the project with suppliers and 3) sophistication of project management tools and

127

methods used to manage the project. We expect relationships that are identified as more difficult
to coordinate to be indicative of higher levels of process interoperability. Difficulty to
coordinate is a sign that much effort (investment) is being placed into managing the project. As
a result of the high level of effort, we expect there to be an increased level of process
interoperability. Similarly, we expect those projects using more complex management tools and
methods to be indicative of higher levels of process interoperability. If a project has a higher
level of process interoperability, then it can be argued that more complex management tools are
needed to ensure an effective and consistent procedural interface between the two firms.
Resource interoperability is operationalized with three items from the survey: 1) the level of
company resources mobilized for the project, 2) how often alliances and partnerships are used
for external technology networking and 3) how often joint ventures are used for external
technology networking. Resource interoperability is more likely to exist in the presence of
resource munificence as opposed to an environment of resource scarcity. Similarly, a company’s
propensity for using alliances, partnerships and joint ventures could indicate a corresponding
level of maturity in establishing interfirm linkages to share rent-generating resources across firm
boundaries (Dyer & Singh, 1998).
Our unit of analysis is the interfirm new product development R&D project. The survey
includes three items relevant to innovation performance at the project level: 1) prospect of
overall success (technical and commercial) for the project, 2) relative importance of the project
to the company and 3) belief that the technological solution at the core of the project will remain
as the dominant approach in the next five years. Successful innovation performance not only
meets the financial and technical objectives of the new product development project, but it does
so in a manner that is both impactful at the company level and has the prospects of generating

128

rents for a prolonged period of time. These three measures therefore capture the essence of
innovation performance at the project level.
As mentioned previously, all measures used for this study have not been used in any other
study with the exception of ‘prospects for project success’, one of three measures used as
reflective indicators of our dependent variable, innovation performance. The research team that
originally collected the data validated this measure using a three-step approach. I quote with
permission a direct communication received by Dr. John Ettlie regarding this validation:
First we predicted that the stage of the project and the odds of success being known
would be correlated, so we correlated project phase (early=1, middle =2, or late
phase=3) with prospects of success. For the U.S. data, we can report significant
results in the predicted direction, r=.334, p=.008, n=63.
Second, we correlated the prospects for success with longitudinal data on revenues of
these firms given the assumption that these power train R&D projects were central to
their overall technology strategy and that they were central to new vehicle
compliance with increasing environmental standards in both countries. Most of these
firms launch global cars and trucks, so the impact of environmental regulation is
trans-border for the typical case we studied, whether originating in the U.S., China,
or Europe (Teece, 2009, p.127). The results for the longitudinal, criterion validation
of the dependent variable were very encouraging.
For the U.S. cases, we also can report very significant correlations between what
managers estimate for R&D project success and percentage increases in sales
resulting after these data were collected. We collected archival revenue data for all
of the firms conducting R&D in the U.S., and after eliminating outliers we found the
following: the correlations (all using two-tailed tests) between self-reported project
success and annual revenue changes are, for revenue changes between 2007 and
2008, r= .334 (p=.095, n=29); for revenue changes between 2007 and 2009, r=.564
(p=.003, n=29); for revenue changes between 2007 and 2010, r=.255 (p=.208, n=26);
and for revenue changes between 2008 and 2009, r=.544 (p=.004, n=26). We
concluded that the self-reported success levels for these R&D projects were a valid
indicator of the true underlying performance of these R&D organizations.
Another normed measure that validates the data is the R&D ratio (R&D investment
as a percentage of sales) in the US sample. The average R&D ratio of the firms that
participated in the survey is 3.98%, while archival NSF data for this industry has a
mean 4% ratio.

129

3.5 Results
The data analysis looks at both the measurement model and path model separately and in
succession (Anderson & Gerbing, 1988). Following this two-step approach, a bootstrapping
analysis is conducted to assess the statistical power of the results (Mooney & Duval, 1993).
Given the importance of the data normality assumption, we first begin with a data properties
analysis and then proceed to the SEM analysis.
3.5.1

Data properties

The data was first analyzed for normality. Skewness, kurtosis and distributional pareto
charts are reported in the appendix. A ladders-of-power analysis using chi square-probabilities
to maximize the likelihood that a given sample is representative of a normal distribution resulted
in five transformation recommendations. The first three were square transformations for the
three innovation performance variables. The fourth was a log transformation of the clarity
measure, while the fifth was a square root transformation of the management tools sophistication
measure. All five transformations were accordingly implemented. The ladders-of-power
analyses, including all transformation recommendations and related chi-square values, are also
reported in the appendix.
Analysis of the data resulted in the deletion of nine cases where missing data was systematic
across multiple measures relevant to this study. This brought our overall sample size to n=67.
Analysis of non-related measures of the nine deleted cases revealed no systematic bias in the
deleted group. The gap size and gap clarity measures had a higher incidence of missing data as
compared to other measures since not all informants provided the in-depth qualitative responses
needed to score these measures. As a result, the sample size for gap size was n=35 while for gap
clarity was n=39. Again, there was no statistical indication that the informants that did not

130

provide qualitative responses were different than those that did, thus leading us to assume no bias
in non-respondents.
The pairwise covariance matrix is therefore used in the SEM analysis along with maximum
likelihood estimators in order to make more full use of all the data set. No imputation of missing
data is recommended when using SEM/ML since doing so can bias the results; rather, the use of
SEM/ML with a pairwise covariance matrix makes the best use of the available data and is the
recommended approach least likely to generate biased results (Peugh & Enders, 2004). The
resulting pairwise covariance is reported below:

131

Table 3.3: Variance/Covariance Matrix
QQ16
QQ16
0.853
QQ18
-0.212
QQ19
-0.190
QQQ9C
-0.294
QQQ9D
-0.373
GAP
-0.082
QQ5SQ
1.300
CLARITYL 0.058
QQ20SQRT 0.001
QQ8SQ
0.862

QQ18

QQ19

QQQ9C QQQ9D GAP

QQ5SQ CLARITYL QQ20SQRT QQ8SQ

0.845
0.309
0.223
0.127
0.000
-1.511
-0.026
0.044
-0.867

0.747
0.198
0.203
0.074
-0.895
-0.061
0.123
-0.792

0.849
0.593
-0.243
-1.614
0.046
0.075
-2.256

29.964
-0.076
-0.066
18.189

1.382
-0.216
-1.884
0.077
0.108
-1.172

1.659
-1.106
-0.491
-0.020
-0.414

132

0.248
0.010
0.262

0.120
0.076

48.430

3.5.2

Measurement model

Using EQS 6.2 for Windows (Bentler, 2002), a confirmatory factor analysis was conducted
to assess the appropriateness of the chosen measures as reflective indicators of the latent
constructs in our model. Not only can we test the fit of the measurement model, but results also
provide indicators of both reliability and validity. Our initial measurement model included the
project’s relative importance to the firm (QQ6) as a reflective measure of innovation
performance, however this measure did not load well (.355) on the innovation performance
construct and was subsequently dropped.
Appropriate fit can be assessed using the comparative fit index (CFI) with desired scores of
0.9 or higher and root mean-square of approximation (RMSEA) with desired scores of .08 or less
(Fawcett et al., 2014). The results from the confirmatory factor analysis yielded a CFI of .965
and an RMSEA of .052, both acceptable values. Key model statistics are reported in tables 3 and
4.

133

Table 3.4: Key Model Statistics

Construct/Item
Innovation Performance
QQ5SQ: Prospect of overall success of the project
QQ8SQ: Technology will be dominant approach in 5
years
Behavioral Interoperability
GAP: Size of gap in JAE
CLARITYL: Clarity in JAE
Process Interoperability
QQ18: Difficulty coordinating project with customers
QQ19: Difficulty coordinating project with suppliers
QQ20SQRT: Sophisitication of project mgmt
tools/methods
Resource Interoperability
QQ16: Level of resources mobilized for the project
QQQ9C: How often company utilize
alliances/partnerships
QQQ9D: How often does company utilize joint ventures

Mean (SD)

Standardized
loadings

15.5079 (5.4739)

0.800

16.5909 (6.9592)

0.597

1.4000 (1.2880)
0.9632 (0.4984)

Average
variance Composite
extracted reliability
0.465
0.634

Mean
shared
variance
0.1395

0.946

0.889

0.0247

0.413

0.628

0.1011

0.441

0.426

0.1811

0.765
-1.000

3.3514 (0.9194)
3.1417 (0.8640)

0.477
0.807

1.6074 (0.3458)

0.482

2.8333 (0.9236)

-0.445

3.1311 (0.9215)
2.4264 (1.1756)

0.818
0.671

Notes: χ2(d.f.) = 33.055 (28); CFI = .965; IFI = .969; RSMEA (90% CI) = .052 (.000-.112). All loadings significant at p < .01.

134

Table 3.5: Squared Correlation Matrix - AVE on Diagonal

PERF
BI
PI
RI

PERF
-------0.465
0.028
0.095
0.295

BI
--------

PI
--------

RI
--------

0.946
0.003
0.043

0.413
0.205

0.441

135

Figure 3.2: Measurement Model

136

3.5.2.1 Reliability
Composite reliability scores have a lower cutoff level of .50 (Hair, Hult, Ringle, & Sarstedt,
2014). A higher score of .70 or greater (.60 for exploratory constructs) is generally
recommended for establishing internal consistency (Fawcett et al., 2014; Nunnally, 1978 ). Our
study was both exploratory and used secondary data that was not initially designed to measure
the constructs of interest in this study. While three of the four constructs exhibited satisfactory
composite reliability scores, the fourth (resource interoperability), fell just below the
recommended threshold of .50 with a composite reliability of .426. As such, we recommend that
a future construct development study be conducted where measures for these constructs can be
further developed and refined.
3.5.2.2 Convergent validity
We can assert convergent validity when standardized loadings on the intended construct are
greater than .50 and the average variance extracted (AVE) exceeds .50 (Fornell & Larcker,
1981a). In our analysis, the AVE for behavioral interoperability is .946 while the other three
AVE values ranged from .413 to .465. Standardized loadings were also near or above .50 (table
3). The lowest standardized loading was -.445. While there is a strong level of convergent
validity for behavioral interoperability, the other three constructs have marginal levels of
convergent validity that are near the threshold guidelines.
3.5.2.3 Discriminant validity
In order to assert discriminant validity, the mean shared variance should be below .50
(Fornell & Larcker, 1981a, b). Alternatively, the AVE for each construct should be greater than
any covariance value with another construct (Fornell & Larcker, 1981a, b). In our analysis, all
mean variance values are less than .50. Additionally, AVE values for all constructs are greater

137

than any single squared correlation value for that construct with the other constructs (table 4).
We therefore have support for discriminant validity.
3.5.2.4 Statistical power analysis – bootstrapping
Given the lower sample size of this data set, a bootstrap analysis is conducted to assess the
statistical power of our measurement model and underlying data set. Using EQS 6.1 for
Windows (Bentler, 2002), bootstrapping conducts an iterative and random sampling of the data
set used in this analysis, conducts the SEM analysis for our posited model, and then calculates
key fit statistics for each iteration. The key statistics are then averaged across all iterations and
reported accordingly. A data set has a high level of statistical power if the bootstrap results are
consistent with the SEM results. Our bootstrap analysis randomly takes 67 samples (with
replacement) from our dataset of 67 cases and does so for 1000 iterations. Of those 1000
iterations only 815 instances actually converged; the remaining 185 instances did not. The
ultimate sample size for the bootstrapping analysis was 815 times 67, or 54,605. While our
initial SEM analysis generated a CFI of .965, the bootstrap average over the 815 iterations that
converged reported an average CFI of .692. While the initial SEM analysis reported an RMSEA
of .052 (.000-.112), the bootstrap analysis reported an average RMSEA of .202 with an average
upper bound of .159 and an average lower bound of .243). In both cases the bootstrap result is
significantly weaker than the original SEM result. Since the original SEM results report
acceptable statistics, we can conclude that our model is a legitimate explanatory model for the
data set at hand. Alternatively stated, with high statistical certainty we cannot reject the null
hypothesis that the data does fit the model. The weak bootstrap results, however, suggest that the
data, not the model, may be a weak link in our analysis. Further replication with a larger sample
size is therefore recommended before we can assert statistical power in the results of this study.

138

3.5.3

Path model

With a measurement model that has acceptable fit, reliability and validity statistics, the
analysis next proceeds to testing the hypothesized relationships with a path model analysis. In
order to test the posited mediating relationship that behavioral interoperability has on both
process and resource interoperability, we take a two-step approach where the first step tests the
relationship between both process and resource interoperability and innovation performance with
no mediator in the model (figure 3). The second step then introduces behavioral interoperability
as the mediating construct while still maintaining direct paths between both process and resource
interoperability and innovation performance (figure 4). Evidence of mediation exists if 1) step
one reveals a statistically significant relationship between the independent variables (process and
resource interoperability) and innovation performance, and 2) the strength of the direct
relationship attenuates in the presence of a mediation variable. Full mediation occurs if the
direct relationship attenuates to zero while partial mediation occurs when the direct relationship
attenuates, yet to a value greater than zero (Baron & Kenny, 1986).
As discussed in Essay One, a mediating rather than a moderating relationship is hypothesized
given the critical role that human interaction plays in the actual innovation process. Ultimately,
it is humans, not resources or processes that generate the solutions and insights needed to achieve
innovation performance. As such the positive influence of both resource and process
interoperability are anticipated to flow through behavioral interoperability in assessing the
relationship between resource/process interoperability and innovation performance. For this
reason, a moderating model is not theoretically justified. Further, any moderation analysis would
require a two-sample invariance test (James & Brett, 1984), which would require splitting the

139

current sample size of n=67 into two groups of n=33 and n=34, a prohibitively small sample size
for CB-SEM. This analysis therefore only tests the mediation model.
Results of the mediation analysis reveal that process interoperability has no statistical
influence on either behavioral interoperability or innovation performance. Resource
interoperability, as predicted, is positively related to behavioral interoperability but is inversely
related to innovation performance, which is opposite from what was initially hypothesized.
Behavioral interoperability, as predicted, is positively related to innovation performance. There
is no attenuation of the relationship between resource interoperability and innovation
performance when behavioral interoperability is introduced as a mediator, therefore the data does
not support a mediating relationship. Both step 1 and step 2 models exhibited acceptable fit
statistics with CFI scores of .904 and .913, and RMSEA scores of .084 and .084, respectively.

140

Figure 3.3: Path Model Step 1 – Mediation Test

PROCESS
INTEROP

-.111 (NS)

INNOVATION
PERFORM

-.533 *
RESOURCE
INTEROP

CFI = .904
RMSEA = .084
* Significant at .01 level

141

Figure 3.4: Path Model Step 2 –Mediation Test

H: 5
H: 8
PROCESS
INTEROP

-.074 (NS)

-.126 (NS)
H: 6
BEHAVIOR
INTEROP
(AJAE GAP SIZE)
(AJAE GAP
CLARITY)

.296*
H: 2

INNOVATION
PERFORM

H: 1
H: 4

.248*
RESOURCE
INTEROP

-.591*
H: 3
H: 7

142

CFI = .913
RMSEA = .084
* Significant at .01 level

3.6 Discussion
3.6.1

Results

In summary, three of the eight hypotheses are supported (Table 5). These results suggest that
while the overall model may not accurately capture the relationship between the constructs of
interoperability, the impact that these constructs have on innovation performance is statistically
significant. The following section discusses the results of each hypothesis.

143

Table 3.6: Hypotheses Overview
H1 Both AJAE gap size and AJAE gap clarity are reflective indicators
of the interfirm behavioral interoperability construct
H2 Interfirm behavioral interoperability is significantly and directly
associated with innovation performance.
H3 Interfirm resource interoperability is significantly and directly
associated with innovation performance
H4 Interfirm resource interoperability is significantly and directly
associated with behavioral interoperability
H5 Interfirm process interoperability is significantly and directly
associated with innovation performance
H6 Interfirm process interoperability is significantly and directly
associated with behavioral interoperability
H7 Interfirm behavioral interoperability significantly mediates the
relationship between resource interoperability and innovation
performance.
H8 Interfirm behavioral interoperability significantly mediates the
relationship between process interoperability and innovation
performance

SUPPORTED
SUPPORTED
NOT SUPPORTED
SUPPORTED
NOT SUPPORTED
NOT SUPPORTED
NOT SUPPORTED

NOT SUPPORTED

Analysis of the data set supports our first hypothesis (H:1) that AJAE gap size and AJAE gap
clarity are both relevant dimensions of behavioral interoperability. It also supports the second
hypothesis (H:2) that behavioral interoperability is positively related to innovation performance.
These two findings alone provide new insight into how differing expectations of joint action can
influence innovation performance in the context of interfirm new product development projects.
Further, they highlight that managing both AJAE gap size and clarity within supply chain
partnerships merits inclusion in the innovation discussion.
A counterintuitive and interesting finding of the analysis is that resource interoperability and
innovation performance are negatively related (H:3) while resource interoperability and
behavioral interoperability are positively related (H:4). Inspection of the factor loadings for the
resource interoperability construct in the measurement model reveals that resource munificence
is negatively correlated to resource interoperability while utilization of partnerships, alliances

144

and joint ventures are positively related to resource interoperability. It appears that a scarcity of
resources (as opposed to resource munificence) is correlated with a higher incidence of formal
external collaboration initiatives. This has a logical basis in that firms that find themselves with
limited resources seek external collaboration opportunities where they can leverage the resources
of other firms to achieve that which they could not achieve on their own (Dyer & Singh, 1998).
A higher level of resource interoperability appears to be evident in these situations. The same
scarcity of resources that leads firms to collaborate with external parties appears to also have an
adverse impact on innovation performance. If we consider the resource-based view of the firm
theory (Barney, 1991; Harmancioglu, Droge, & Calantone, 2009; Wernerfelt, 1984), that a firm
can establish a competitive advantage in the marketplace leveraging its resources, then these
opposing results are understandable.
Another potential explanation of this counterintuitive finding is that the sample used in this
dataset surveyed companies that typically have a long-term relationship with their supply chain
partners such that resource interoperability is already established at a high level. Further
increases in investment in resources may experience diminishing returns as evidenced by the
data. The results of the negative relationship in the analysis may be indicating such a situation.
A third potential explanation is that innovation performance is more likely to be achieved
when firms combine a diversity of resources, rather than redundant resources. Sampson (2007),
for example, argues that technological diversity can have a positive (yet curvilinear) relationship
with innovation performance. The data observed in this dataset may in fact be reflecting this
very phenomenon where lower levels of resource interoperability may actually be contributing to
higher levels of innovation performance.

145

Diversity of resources can be considered as both the form of the resource and the criticality
of the resources committed to the relationship. If one partner commits their prime resources to
the relationship while the other partner does not (and reserves their prime resources for another
supply chain partner), then issues in resource interoperability can emerge. This is an area with
potential for future study.
Another unexpected finding is that process interoperability has no statistical relationship with
either innovation performance (H:5) or behavioral interoperability (H:6). Process
interoperability was operationalized with three measures: two of them captured the difficulty of
coordinating with customers and supplier, while the third measure captured the sophistication of
project management tools and techniques used to manage the project. Give the extant supply
chain literature on the benefits of process integration across supply chain partners (Frohlich &
Westbrook, 2001; Schoenherr & Swink, 2012) we recommend further construct development
research be done to strengthen the measurement of the process interoperability construct.
Finally, our mediation analysis did not yield any results that suggest that behavioral
interoperability acts as a mediator. Since process interoperability did not exhibit a direct
relationship with innovation performance, the mediation analysis related to this dimension was a
moot point (H:8). The resource interoperability dimension, however, was a different case. Step
one of the mediation analysis showed that resource interoperability is negatively related to
innovation performance with a path loading of -.533 (figure 3). When we introduce behavioral
interoperability in step two, the path loading remains relatively unchanged at -.591 (figure 4).
No attenuation of the relationship occurs in the presence of behavioral interoperability, therefore
a mediation model is not supported (H:7).

146

3.6.2

Managerial implications

The lack of a mediating relationship, while not aligned with the initial hypothesis, is
interesting in and of itself and has important managerial implications. Namely, this suggests that
resource interoperability does not influence the ability for behavioral interoperability to
positively impact innovation performance. Managers need not wait to establish high levels of
resource interoperability before they can reap the benefits of establishing behavioral
interoperability. Similarly, managers need not be dissuaded to pursue behavioral interoperability
if the relationship currently lacks desired levels of resource interoperability. The benefits
achieved by having lower sized gaps in AJAE and higher levels of clarity of those expectations
can generate a positive impact on innovation performance independent of the state of other
interoperability dimensions. This finding is significant.
The original research question of ‘what role does AJAE play in achieving innovation
performance within supply chain partnerships (ally)’? is answered with this analysis. Both
smaller AJAE gap size and increased levels of AJAE gap clarity are reflective of higher levels of
behavioral interoperability. Additionally, higher levels of behavioral interoperability are
statistically and directly related with improved innovation performance. In short, understanding
and managing AJAE is relevant to achieving improved innovation performance. Therefore firms
that can understand and manager AJAE with supply chain partners will have the ‘ally’ option
available as a strategic alternative to the ‘make’ or ‘buy’ options.
3.6.3

Limitations

Interoperability needs further construct development given its nascent nature in supply chain
literature. The nature of the secondary data set used in this study limits the ability of this
research to explore a more full treatment of the construct, yet provides a starting point from

147

which future research can build upon. Future studies that expand upon relevant indicators and
establish more robust levels of reliability of the three dimensions of interoperability will
strengthen the foundation for this research stream going forward.
Additionally, studies with larger samples sizes will likewise enhance the statistical power of
the hypothesized model and potentially uncover alternative theories that refine that which has
been presented in this essay. One inherent limitation in this research stream regarding increased
sample size is the guarded nature of firms with respect to sharing both dyadic relationship data
and information regarding innovation initiatives. Sharing dyadic relationship data with a third
party carries risks for the sharing firm, especially if the data suggests strained relations and the
other parties finds out that their partner has shared such data with a third party. Knowledge of
this could further strain the relationship.
Innovation initiatives represent a source of competitive advantage and any risk of related
information leaking to competitors through third parties renders this a difficult topic to fully
explore with firms. There are remedies to these issues, but such remedies often require legal
agreements (non-disclosure agreements) and other processes that limit the ability of a researcher
to gather large numbers of data points for analyses and often constrain the ability to publish
results.
3.6.4

Extensions

Findings in this study provide numerous seeds from which unique yet relevant lines of
inquiry can emerge and develop. One potential competing theory could be that behavioral
interoperability is endogenous to resource and process interoperability rather than the other way
around as is presented in this study. There is much that can be explored with respect to the
nature of the relationship between the three dimensions of interoperability. Another alternative

148

theory is that interoperability is a second-order formative construct with the three dimensions of
behavioral, resource and process interoperability acting as the underlying constructs. These
alternative models merit discussion, debate and further research.
This study has focused on the role that AJAE have on behavioral interoperability. Earlier in
this essay, we listed numerous other potential factors that may also impact behavioral
interoperability. These included firms’ respective views on distributive/procedural justice,
appropriate levels of financial transparency, risk profiles and corporate culture. One aspect of
interfirm supply chain partnerships that can profoundly impact these factors is the respective
business model (desired outcome) that each firm pursues. Focus on a given business model may
require a certain set of priorities and views that are not shared when pursuing other business
models. These differences may become apparent when both firms attempt to solve a common
problem differently and in ways that suboptimize the desired outcome for the other firm. When
two firms engage in a relationship, the match or mismatch of their business models may largely
impact the interoperability of those two firms.
This research used the interfirm project as the unit of analysis. Extensions could include
expanding the unit of analysis to the business unit level, plant level or company level. The
research could also expand the unit of analysis to include a triad supply chain relationship.
While the dyad is the basic unit of the supply chain, emerging research suggest that triads may
also be a relevant unit of analysis in supply chain research (Choi & Wu, 2009). In addition to
considering alternate units of analyses, extensions could also include studying longitudinal or
panel data so that the impact of time could be included in the model.
Another area of research that may benefit from the lens of interoperability and AJAE is
virtual engineering (Montoya, Massey, Hung, & Crisp, 2009), where firms collaborate

149

engineering design activities over virtual communication networks and tools. Ettlie and Pavlou
(2006) used virtual engineering cases of NPD to test the relationship between dynamic
capabilities and NPD performance. Interoperability may be an additional construct to their
model that can more fully describe the relationship between dynamic capabilities and
performance. On the one hand, interoperability may be a moderating construct while on the
other hand interoperability may be endogenous such that dynamic capabilities mediates the
relationship between interoperability and performance. In short, there are multiple research
extension opportunities for the constructs presented in this essay.

3.7 Conclusion
While the importance of good supplier relations is not new to supply chain literature,
exploring the nuanced behavioral elements of good relations remains a fertile area with much
research opportunity. In this essay, we have provided empirical support for the theory that AJAE
in interfirm relations has a statistically significant relationship with innovation performance.
These results argue that managers desiring improved innovation performance in the context of
interfirm NPD initiatives should focus on two dimensions related to joint action expectations: 1)
the size of the gap in AJAE and 2) clarity of those expectations. Clarity arises when firms take
time to understand the costs and benefits associated with various levels of joint action from both
their own perspective and that of the partner firm. Further, these two dimensions related to
AJAE are related to a firm’s ability to establish interoperability with its supply chain partners.
Interoperability is a construct that captures the ability of two entities to provide services to
and receive services from each other. Whereas integration and collaboration are measures of
actual exchange between two entities, interoperability measures the readiness or capability for
those entities to so engage in successful exchanges. Firms that have established certain levels of

150

interoperability will be able to engage in collaborative and integrative initiatives (ally) both more
efficiently and more effectively than those that do not. Those firms that continue to struggle
with executing an ally strategy will waste both scarce time and resources and will find
themselves at a disadvantage to those firms that are able to successfully pursue an ally strategy.
Our research highlighted three dimensions of interoperability: resource, process and
behavioral. Of these three dimensions, only behavioral interoperability was found to have a
positive impact on innovation performance in the context of a joint interfirm NPD project. In
today’s environment of ever-increasing clockspeeds of business, establishing behavioral
interoperability thus becomes an important concept that managers should consider when
managing supply chain relationships, especially for those seeking to maximize the innovation
potential that resides within.

151

4 Conclusion
Supply chain relationships are becoming increasingly important, especially for those supply
chains where deliberate innovation is a critical outcome and a determinant of success.
Innovation today increasingly takes place across firm boundaries where improved
telecommunication and computing tools have facilitated the shift from a centralized innovation
model to one that is more dispersed across the supply chain. The shift in the locus of innovation
from automotive OEMs towards their tier one suppliers is an example of this trend (Rothenberg
& Ettlie, 2011). Increasingly, the interfirm relationship has become an important unit of analysis
with respect to understanding innovation and successful new product development initiatives.
Interoperability is a construct traditionally used both in military and information technology
fields of research to describe the ability for different groups to successfully interact (LaVean,
1980). This dissertation extended the construct of interoperability to the field of supply chain
management and in particular to the study of interfirm relationships. Three distinct levels of
interoperability were proposed: resource interoperability, process interoperability and behavioral
interoperability. While much supply chain research traditionally focuses on the factors that
support resource and process interoperability, research on the behavioral aspects related to
effective supply chain interoperability is still nascent and was the domain for the present
research.
This dissertation argued that economic models alone are insufficient to explain differential
success in interfirm innovation performance given the complex human interactions involved.
Rather, behavioral interoperability need also be considered in tandem with economic models.
Asymmetrical joint action expectations (AJAE) is posited as relevant to and indicative of
behavioral interoperability. Essay one in this dissertation built upon a traditional economic

152

analysis, where firms engage in activities as a function of balancing the marginal costs and
marginal benefits of those activities. Joint action between interfirm supply chain partners is
applied to this analysis where both firms engage in a shared level of joint action, however, bring
with them differing marginal cost and benefit curves. As a result, both firms likely experience
differing optimal levels of joint action, which therefore can lead to differing expectations
regarding the desired level of joint action for the interfirm relationship. Essay one proposed that
larger gaps in AJAE represent lower levels of behavioral interoperability while smaller gaps in
AJAE represent higher levels. Further, essay one proposed that higher levels of behavioral
interoperability, along with higher levels of resource and process interoperability, are all directly
related to higher levels of interfirm innovation performance.
Given the nascent nature of research in AJAE, a multi-case field study was conducted to
confirm the existence of AJAE and understand its impact on innovation performance. Essay two
reported the results of the case study and confirmed both the existence and the relevance of
AJAE to innovation performance. Further, essay two revealed that not only is the size of the gap
of AJAE important, but so also is the clarity of the expectations related to AJAE. In other words,
an important factor related to behavioral interoperability and innovation performance is the
understanding that one firm has of not only their own marginal costs and benefits associated with
joint action, but also that of the partner firm. As a result, the proposed model was refined to
include both AJAE gap size and AJAE clarity as reflective indicators of behavioral
interoperability.
Essay three applied structural equation modeling to test the hypothesized model using a
secondary dataset of survey responses regarding automotive powertrain R&D projects. Results
confirmed the relevance of behavioral interoperability to innovation performance and also

153

confirmed both AJAE gap size and AJAE clarity as reflective indicators of behavioral
interoperability.
This dissertation provides an analytical framework by which other behavioral-related
research can be enriched and furthered. Trust-related research, for example, can be applied to
the analytical model presented in essay one to understand how increased levels of trust with one
or both parties can impact AJAE. Other behavioral-related topics such as risk profiles can also
be explored in this context. Interesting questions include what factors can shift the marginal cost
and benefit curves associated with joint action and what factors can change the slope of these
curves? How do these changes therefore impact the size of the gap in AJAE? When an interfirm
relationship is characterized by a large gap in AJAE, how do the firms deal with such gaps?
Assuming AJAE clarity is present, do relationships persist despite the gap, do they attempt to
close the gap, or do they dissolve the relationship? How do firms close gaps in AJAE and who is
able to capture the benefits from so doing? Additionally, are economic rents the only benefits to
be negotiated or are there other beneficial rents to be earned, such as learning. It is anticipated
that topics such as relational power, citizenship behavior and trust, to name a few, may provide
insights into these questions.
Another topic of interest that can extend from this research is a better understanding of the
creation, persistence and dissolution of supply chain relationships. Traditional purchasing
literature advocates for deep and long-term relationships with one’s suppliers. Anecdotal
evidence (Babcock, 2011) including Case F in essay two, however, suggest that effective supply
chain relationships may not in fact require long-term commitments in order to achieve high
levels of relationship quality and performance. Some firms engage in a series of short-term
relationships that generate success for both participating firms without the need for the

154

relationship to be long-term in nature. Applying the theoretical lens of managing AJAE
presented in this dissertation may help explain and predict why such relationships can still be
successful.
SEM analysis assumes linear relationships between the constructs. An interesting question to
investigate in future research would be whether or not the relationship between smaller
gaps/greater clarity in AJAE is in fact linear with innovation performance. Alternatively, the
relationship could be curvilinear where diminishing returns occur with higher levels of
behavioral interoperability.
Another line of inquiry could apply this research to a broader unit of analysis. This
dissertation focused on the supply chain dyad. However, in complex projects, innovation could
come from multiple tiers of the supply chain. Studying the innovation of entire supply chains as
opposed to dyads may provide interesting insights as to how firms can participate in and help
create an innovation ecosystem that leads to differential performance.
Another dimension that can be explored is the intent of the relationship entered into. An
observation in the case studies that was not elaborated on in Essay Two was the concept of how
some firms’ primary focus was on the outcome of the project at hand while others firms focused
more on the long-term benefit potentially derived from the relationship. In Case D, CTC
tolerated a significant gap in AJAE with GBC due to the longevity of the established relationship
and the future potential rents to be gained in both the product category being studied and in
adjacent product categories provided by GCM to other business units. In short, CTC tolerated
the gap due to their focus on the relationship as opposed to the outcome of the project at hand.
In Case F, however, the relationship was created for a single project where both parties
understood that the relationship would dissolve upon project completion. In that setting, gaps in

155

AJAE were not tolerated and investments in communication of expectations were made up front.
This was done because the relationship had only ‘one shot’ at achieving the performance needed
and extracting the benefits needed to justify investment in creating the relationship to begin with.
Culture (Hofstede, 1984) is another construct that can be introduced in future research of
behavioral interoperability and AJAE. Differences in national culture may have an impact on
respective marginal cost and benefit curves of partner firms. How dyads close the AJAE gap
may also be impacted by culture.
This dissertation has approached optimality as if it were a single point. What if we take a
more broad approach and consider a range of AJAE gap size that is acceptable? If so, then
where is the tipping point where the size of the asymmetry does in fact matter?
Essay three argued that in the context of transaction cost economics (Williamson, 1979), the
ability for a firm to establish behavioral interoperability with a supply chain partner is critical to
that firm’s ability to consider and implement ‘ally’ as a feasible governance structure as opposed
to the ‘make’ or ‘buy’ structures of governance (Geyskens et al., 2006; Williamson, 2008).
Absent this ability, the firm will be left with a more limited strategic toolbox. The high
incidence rate of alliance failures, reported as high as 50% (Kale & Singh, 2009; Nidumolu et al.,
2014; Park & Ungson, 2001), again suggests that economic models alone may not be sufficient
to explain differential interfirm innovation performance. The ability of firms to manage AJAE
with their supply chain partners was shown in this dissertation to be relevant to both behavioral
interoperability and innovation performance. As such, this dissertation presents an important
step forward in better understanding how firms can create a supply chain ecosystem where
innovation can flourish.

156

APPENDICES

157

APPENDIX A: CASE STUDY INTERVIEW PROTOCOL
Figure A.1: Interview protocol
INTERVIEW GUIDE: Interviewing the Buying Firm
1. For this particular project <RESTATE NAME OF PROJECT> that we are focusing
on, who has the ownership of the value-adding intellectual property?
Mostly the
Buying Firm
Supplier Firm
O

Shared Equally
Across Both
O

O

O

Mostly the

O

2. Do you have a technology agreement with the partner firm? If yes, please briefly
describe.

158

Figure A.1 (cont’d)
SAMPLE: Innovation/Contribution Timelin
Timeline

YOUR COMPANY – PLEASE FILL IN: Innovation/Contribution Timeline

Initiation Stage
3. Please fill in the timeline above showing your company’s % contribution to the
innovation project for each of the periods of the timeline. Please describe each phase

159

Figure A.1 (cont’d)
4. When exactly did the relationship with <INSERT SUPPLIER NAME> form for this
particular project?
5. Who initiated the relationship?
Implementation Stage
6. Do you believe that your supplier has an “A” team and a “B” team regarding product
produc
development? If so, then how would you classify the resources that <INSERT
SUPPLIER NAME> has made available to your firm for this particular project? Why
do you think this was so?
7. Were there any factors that positively or negatively impacted the ability
abili to bring the
innovation to market? Please explain..
8. What, if anything, do you wish <INSERT SUPPLIER NAME> would have done
differently?
Joint Action Expectations
The level of joint action between two firms can be described as how intensely two firms interact
inte
for a common purpose. We are interested in both the actual level of joint action that existed and
the desired level of joint action that existed. Below is a sample figure that depicts three different
scenarios:
A) The actual level of joint action for the project is the same as the desired level
B) The actual level of joint action for the project is lower than the desired level
C) The actual level of joint action for the project is higher than the desired level
SAMPLE: Innovation/Joint Action Timeline

160

Figure A.1 (cont’d)

YOUR COMPANY – PLEASE FILL IN: Innovation/Joint Action Timeline

9. Please fill in the figure above with the actual and desired levels of joint action for
each stage of the project development cycle and discuss the following for each stage:
a. Describe the nature of the joint action
b. Are you satisfied with <INSERT SUPPLIER NAME> level of joint action in
the project? Too much, too little?
i. If too much, why? How has your partner gone “too far”
ii. If too little, why? What opportunities exist?
c. Aree you satisfied with your company’s level of joint action in the project?
Too much, too little?
iii. If too much, why? How has your company gone “too far”
iv. If too little, why? What opportunities exist?
d. If a gap exists between actual and desired levels of joint action, did this gap
impact the outcome of the project? If yes, in what ways?

161

Figure A.1 (cont’d)

Performance of the Innovation Project
10. How did this project perform along the following dimensions?
Did not challenge
existing ideas

1

2

3

4

5

6

7

Challenged existing
ideas

Did not offer
new ideas

1

2

3

4

5

6

7

Offered new
ideas

Not Creative

1

2

3

4

5

6

7

Creative

Did not offer
superior value
to the customer

1

2

3

4

5

6

7

Offered
superior value
to the customer

Late

1

2

3

4

5

6

7

Ahead of Schedule

Over Budget

1

2

3

4

5

6

7

Under Budget

7

MarketAcceptance
Stronger than
Expected

Market Acceptance
Worse than
Expected

1

2

3

4

162

5

6

APPENDIX B: SURVEY MEASURES FROM US POWERTRAIN R&D
STUDY
Figure B.1: R&D study survey protocol
Innovation performance
QQ5) How do feel about the prospects of the overall success (technical &
commercial) for the project?
1=not likely to succeed but we will learn a lot; 2=less than 50/50 chance of success;
3=50/50 chance of success; 4=good chance of success; 5=a sure thing, very good odds
of success.
QQ6) How does this project compare to others you are familiar with in terms of its relative
importance to the company?
1=Not Important, 2=Marginally Important, 3=Somewhat Important, 4=Important,
5=Very Important
QQ8) How much do you believe that the technological solution that is the core of the
project will remain as the dominant approach in the next 5 years where you are:
1=Very unsure; 2=Somewhat unsure; 3=Neither certain nor unsure; 4=Somewhat
certain; 5= Very certain
Process interoperability
QQ18) (For b-to-b suppliers) How would you rate the difficulty of coordinating the
development of this project with your customers where….
1=Very Easy, 2=Easy, 3=Neither Easy Nor Difficult, 4=Difficult, 5=Very Difficult
(if difficult, please explain_______________________________________)
QQ19) (For manufacturers and suppliers) How would you rate the difficulty of
coordinating the development of this project with your suppliers where….
1=Very Easy, 2=Easy, 3=Neither Easy Nor Difficult, 4=Difficult, 5=Very Difficult
(if difficult, please explain______________________________________)
QQ20) What project management tools or methods (if any) were used for this project (e.g.,Gantt
charts)?_______________________________________

163

Figure B.1 (cont’d)

Resource interoperability
QQ16) How would you describe your company’s level of resources mobilized for this project?
1=Extremely Scarce Resources 2= Somewhat Scarce Resources, 3=Neither Scarce nor
Ample Resources, 4=Ample Resources, 5=More than Ample Resources
QQQ9) How often do you utilize the following mechanisms for external technology
networking?
C. Alliances and Partnerships
1=not at all
2=not very often
3=somewhat often 4=often
5=very often
D. Joint-ventures
1=not at all
2=not very often
3=somewhat often 4=often
5=very often

164

APPENDIX C: DATA
TA NORMALITY AND TRANSFORMATION
ANALYSIS
Figure C.1: Normality analysis and data transformation analysis

.
Transformation
cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

ladder
formula
QQ5^3
QQ5^2
QQ5
sqrt(QQ5)
log(QQ5)
1/sqrt(QQ5)
1/QQ5
1/(QQ5^2)
1/(QQ5^3)

QQ5
chi2(2)
3.1
3.3
16.41
26.46
38.23
50.81
62.84
.
.

Recommendation: Use ’square’ transformation.

165

P(chi2)
0.212
0.192
0
0
0
0
0
0
0

Figure C.1 (cont’d)

. ladder QQ6
Transformation
----------------------cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
---------------QQ6^3
QQ6^2
QQ6
sqrt(QQ6)
log(QQ6)
1/sqrt(QQ6)
1/QQ6
1/(QQ6^2)
1/(QQ6^3)

chi2(2)
-------------------18.82
5.29
7.46
11.98
17.51
23.35
28.99
38.38
44.64

Recommendation: Use ’square’ transformation.

166

P(chi2)
------0
0.071
0.024
0.003
0
0
0
0
0

Figure C.1 (cont’d)

. ladder QQ8
Transformation
----------------------cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
---------------QQ8^3
QQ8^2
QQ8
sqrt(QQ8)
log(QQ8)
1/sqrt(QQ8)
1/QQ8
1/(QQ8^2)
1/(QQ8^3)

chi2(2)
-------------------14.1
3.81
10.29
20.85
33.75
46.47
56.97
69.08
73.47

Recommendation: use the ‘square’ transformation

167

P(chi2)
------0.001
0.148
0.006
0
0
0
0
0
0

Figure C.1 (cont’d)

. ladder GapSh
Transformation
----------------------cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
----------------GapSh^3
GapSh^2
GapSh
sqrt(GapSh)
log(GapSh)
1/sqrt(GapSh)
1/GapSh
1/(GapSh^2)
1/(GapSh^3)

chi2(2)
------------------5.69
5.5
2.78
4.47
8.24
12.52
16.12
20.25
21.72

P(chi2)
------0.058
0.064
0.249
0.107
0.016
0.002
0
0
0

Note: Because of the negative values, I first did a shift in the data creating variable GapSh = Gap
+2. This enabled me to feasible investigate all transformations without running into the negative
number issue.
Recommendation: Do nothing - maintain identify
168

Figure C.1 (cont’d)

. ladder Clarity
Transformation
----------------------cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
-----------------Clarity^3
Clarity^2
Clarity
sqrt(Clarity)
log(Clarity)
1/sqrt(Clarity)
Clarity)
1/Clarity
1/(Clarity^2)
1/(Clarity^3)

chi2(2)
-----------------5.54
5.07
5.91
4.17
3.14
5.61
9.11
14.24
16.29

Recommendation: transform data using the ‘log’ function

169

P(chi2)
------0.063
0.079
0.052
0.125
0.208
0.061
0.011
0.001
0

Figure C.1 (cont’d)

Transformation
cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
QQ18^3
QQ18^2
QQ18
sqrt(QQ18)
log(QQ18)
1/sqrt(QQ18)
1/QQ18
1/(QQ18^2)
1/(QQ18^3)

chi2(2)
7.96
3.32
0.33
5.28
14.17
25.14
35.39
48.84
54.24

Recommendation: Maintain identity - no transformation

170

P(chi2)
0.019
0.19
0.847
0.071
0.001
0
0
0
0

Figure C.1 (cont’d)

. ladder QQ19
Transformation
----------------------cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
---------------QQ19^3
QQ19^2
QQ19
sqrt(QQ19)
log(QQ19)
1/sqrt(QQ19)
1/QQ19
1/(QQ19^2)
1/(QQ19^3)

chi2(2)
-------------------10.95
3.16
0.66
2.23
9.49
21.86
37.05
64.64
.

Recommendation: maintain identity - no transformation

171

P(chi2)
------0.004
0.206
0.718
0.327
0.009
0
0
0
0

Figure C.1 (cont’d)

. ladder QQ20
Transformation
----------------------cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
---------------QQ20^3
QQ20^2
QQ20
sqrt(QQ20)
log(QQ20)
1/sqrt(QQ20)
1/QQ20
1/(QQ20^2)
1/(QQ20^3)

chi2(2)
-------------------17.96
8.51
4.41
2.91
3.22
8.82
15.81
25.71
29.65

Recommendation: Transform data using ’sqrt’ function

172

P(chi2)
------0
0.014
0.11
0.234
0.2
0.012
0
0
0

Figure C.1 (cont’d)

. ladder QQQ9C
Transformation
----------------------cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
----------------QQQ9C^3
QQQ9C^2
QQQ9C
sqrt(QQQ9C)
log(QQQ9C)
1/sqrt(QQQ9C)
1/QQQ9C
1/(QQQ9C^2)
1/(QQQ9C^3)

chi2(2)
------------------17.84
10.15
1.36
3.41
13.7
27.69
40.91
58.37
65.57

Recommendation: No transformation - maintain identity

173

P(chi2)
------0
0.006
0.506
0.182
0.001
0
0
0
0

Figure C.1 (cont’d)

. ladder QQQ9D
Transformation
cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
QQQ9D^3
QQQ9D^2
QQQ9D
sqrt(QQQ9D)
log(QQQ9D)
1/sqrt(QQQ9D)
1/QQQ9D
1/(QQQ9D^2)
1/(QQQ9D^3)

chi2(2)
11.25
7.15
12.9
14.53
12.45
10.06
9.28
9.98
10.61

P(chi2)
0.004
0.028
0.002
0.001
0.002
0.007
0.01
0.007
0.005

Recommendation: No transformation - maintain identity.
Note: Since QQQ9C is similar in measure to QQQ9D, and since QQQ9C is normally distributed,
then I recommend maintaining QQQ9D per identity, even though the square function would

174

Figure C.1 (cont’d)
provide marginal improvement in normality. This variab
variable
le may be problematic in our analysis
and I may want to consider dropping the variable.

. ladder QQ16
Transformation
cubic
square
identity
square root
log
1/(square root)
inverse
1/square
1/cubic

formula
QQ16^3
QQ16^2
QQ16
sqrt(QQ16)
log(QQ16)
1/sqrt(QQ16)
1/QQ16
1/(QQ16^2)
1/(QQ16^3)

chi2(2)
13.81
4.16
1
2.54
9
18.06
27.05
39.39
44.61

Recommendation: no transformation - maintain identity

175

P(chi2)
0.001
0.125
0.606
0.281
0.011
0
0
0
0

REFERENCES

176

REFERENCES
Adobor, H., & McMullen, R. 2007. Supplier Diversity and Supply Chain Management: A
Strategic Approach. Business Horizons, 50(3): 219-229.
Agarwal, R., Croson, R., & Mahoney, J. T. 2010. The Role of Incentives and Communication in
Strategic Alliances: An Experimental Investigation. Strategic Management Journal,
31(4): 413-437.
Ahuja, G., & Katila, R. 2001. Technological Acquisitions and the Innovation Performance of
Acquiring Firms: A Longitudinal Study. Strategic Management Journal, 22(3): 197220.
Allred, C. R., Fawcett, S. E., Wallin, C., & Magnan, G. M. 2011. A Dynamic Collaboration
Capability as a Source of Competitive Advantage. Decision Sciences, 42(1): 129-161.
Anderson, J. C., & Gerbing, D. W. 1988. Structural Equation Modeling in Practice: A Review
and Recommended Two-Step Approach. Psychological Bulletin, 103(3): 411.
Arora, A., Gambardella, A., Magazzini, L., & Pammolli, F. 2009. A Breath of Fresh Air? Firm
Type, Scale, Scope, and Selection Effects in Drug Development. Management Science,
55(10): 1638-1653.
Babcock, C. 2011. Heavy Truck Maker Revs Up IT Innovation Amid A Brutal Recession.
Informationweek(1309): 21-24.
Bansal, P., & Roth, K. 2000. Why Companies Go Green: A Model of Ecological
Responsiveness. Academy of Management Journal, 43(4): 717-736.
Barney, J. 1991. Firm Resources and Sustained Competitive Advantage. Journal of
Management, 17(1): 99.
Barney, J. B. 1986. Strategic Factor Markets: Expectations, Luck, and Business Strategy.
Management Science, 32(10): 1231-1241.
Baron, R. M., & Kenny, D. A. 1986. The Moderator-Mediator Variable Distinction in Social
Psychological Research: Conceptual, Strategic, and Statistical Considerations. Journal of
Personality and Social Psychology, 51(6): 1173.
Bentler, P. M. 2002. Structural Equations Program Manual. Encino, CA: Multivariate
Software, Inc.
Bentler, P. M., & Chou, C.-P. 1987. Practical Issues in Structural Modeling. Sociological
Methods & Research, 16(1): 78-117.
Bercovitz, J. J., Sandy D.; Nickerson, Jack A. 2006. The Antecedents and Performance
Implications of Cooperative Exchange Norms. Organization Science, 17(6): 724-740.
177

Berggren, C., Magnusson, T., & Sushandoyo, D. 2009. Hybrids, Diesel or Both? The Forgotten
Technological Competition for Sustainable Solutions in the Global Automotive Industry.
International Journal of Automotive Technology and Management, 9(2): 148-173.
Calantone, R. J., Harmancioglu, N., & Droge, C. 2010. Inconclusive Innovation “Returns”: A
Meta-Analysis of Research on Innovation in New Product Development*. Journal of
Product Innovation Management, 27(7): 1065-1081.
Chesbrough, H. W. 2003. The Era of Open Innovation. MIT Sloan Management Review, 44(3):
35.
Chew, E., & Whitbread, C. 2002. Parts Makers Getting Choosy About the Deals They Chase,
Automotive News.
Choi, T. Y., & Wu, Z. 2009. Triads in Supply Networks: Theorizing Buyer–Supplier–Supplier
Relationships. Journal of Supply Chain Management, 45(1): 8-25.
Chopra, S., & Sodhi, M. S. 2004. Managing Risk To Avoid Supply-Chain Breakdown. MIT
Sloan Management Review, 46(1): 53-62.
Coase, R. H. 1937. The Nature of the Firm. Economica, 4(16): 386-405.
Corbin, J., & Strauss, A. 2008. Basics of Qualitative Research (3rd ed.). Thousand Oaks, CA:
Sage Publications, Inc.
Council, S. C. 2008. Supply Chain Operations Reference Model: Supply Chain Council.
Croson, R., & Donohue, K. 2003. Impact Of POS Data Sharing On Supply Chain Management:
An Experimental Study. Production & Operations Management, 12(1): 1-11.
Das, A., & Narasimhan, R. 2001. Process-Technology Fit and its Implications for Manufacturing
Performance. Journal of Operations Management, 19(5): 521-540.
Das, A., Narasimhan, R., & Talluri, S. 2006. Supplier Integration--Finding an Optimal
Configuration. Journal of Operations Management, 24(5).
Dawes, J. 2008. Do Data Characteristics Change According to the Number of Scale Points Used?
International Journal of Market Research, 50(1): 61-77.
Dyer, J., & Singh, H. 1998. The Relational View: Cooperative Strategy and Sources of
Interorganizational Competitive Advantage. The Academy of Management Review,
23(4): 660-679.
Dyer, W. G., Jr., & Wilkins, A. L. 1991. Better Stories, Not Better Constructs, to Generate Better
Theory: A Rejoinder to Eisenhardt. The Academy of Management Review, 16(3): 613619.

178

Eisenhardt, K. M. 1989. Building Theories From Case Study Research. The Academy of
Management Review, 14(4): 532.
Eisenhardt, K. M. 1991. Better Stories and Better Constructs: The Case for Rigor and
Comparative Logic. The Academy of Management Review, 16(3): 620-627.
Eisenhardt, K. M., & Graebner, M. E. 2007. Theory Building from Cases: Opportunities and
Challenges. Academy of Management Journal, 50(1): 25.
Esper, T. L., Defee, C. C., & Mentzer, J. T. 2010. A Framework of Supply Chain Orientation.
The International Journal of Logistics Management, 21(2): 161-179.
Ettlie, J. E. 2007. Empirical Generalization and the Role of Culture in New Product
Development*. The Journal of Product Innovation Management, 24(2).
Ettlie, J. E., & Pavlou, P. A. 2006. Technology-Based New Product Development Partnerships*.
Decision Sciences, 37(2).
Fawcett, S. E., Jones, S. L., & Fawcett, A. M. 2012. Supply Chain Trust: The Catalyst for
Collaborative Innovation. Business Horizons, 55(2): 163-178.
Fawcett, S. E., Waller, M. A., Miller, J. W., Schwieterman, M. A., Hazen, B. T., & Overstreet,
R. E. 2014. A Trail Guide to Publishing Success: Tips on Writing Influential Conceptual,
Qualitative, and Survey Research. Journal of Business Logistics, 35(1): 1-16.
Fisher, M. L. 1997. What is the Right Supply Chain for Your Product? Harvard Business
Review, 75(2): 105-116.
Ford, T. C., Colombi, J. M., Graham, S. R., & Jacques, D. R. 2007. A Survey on Interoperability
Measurement, 12th International Command and Control Research and Technology
Symposium. Newport, RI: United States of America Department of Defense.
Ford, T. C., Colombi, J. M., Jacques, D. R., & Graham, S. R. 2009. A General Method of
Measuring Interoperability and Describing Its Impact on Operational Effectiveness. The
Journal of Defense Modeling and Simulation: Applications, Methodology, Technology,
6(1): 17-32.
Fornell, C., & Larcker, D. F. 1981a. Evaluating Structural Equation Models with Unobservable
Variables and Measurement Error. JMR, Journal of Marketing Research, 18(1): 39.
Fornell, C., & Larcker, D. F. 1981b. Structural Equation Models with Unobservable Variables
and Measurement Error: Algebra and Statistics. JMR, Journal of Marketing Research,
18(3): 382.
Friedman, T. L. 2005. The World is Flat. New York: Farrar, Straus and Giroux.
Frohlich, M. T., & Westbrook, R. 2001. Arcs of Integration: An International Study of Supply
Chain Strategies. Journal of Operations Management, 19(2): 185.

179

Geyskens, I., Steenkamp, J.-B. E. M., & Kumar, N. 2006. Make, Buy, or Ally: A Transaction
Cost Theory Meta-Analysis. Academy of Management Journal, 49(3): 519-543.
Goodale, J. C., Kuratko, D. F., Hornsby, J. S., & Covin, J. G. 2011. Operations management and
corporate entrepreneurship: The moderating effect of operations control on the
antecedents of corporate entrepreneurial activity in relation to innovation performance.
Journal of Operations Management, 29(1-2): 116-127.
Griffith, D. A., Harvey, M. G., & Lusch, R. F. 2006. Social Exchange in Supply Chain
Relationships: The Resulting Benefits of Procedural and Distributive Justice. Journal of
Operations Management, 24(2).
Gulati, R., & Sytch, M. 2007. Dependence Asymmetry and Joint Dependence in
Interorganizational Relationships: Effects of Embeddedness on a Manufacturer's
Performance in Procurement Relationships. Administrative Science Quarterly, 52(1): 3269.
Hair, J. F., Hult, G. T., Ringle, C. M., & Sarstedt, M. 2014. A Primer on Partial Least Squares
Structural Equation Modeling (PLS-SEM). Los Angeles: Sage Publications Inc.
Harmancioglu, N., Droge, C., & Calantone, R. 2009. Strategic Fit to Resources Versus NPD
Execution Proficiencies: What Are Their Roles in Determining Success? Academy of
Marketing Science. Journal, 37(3): 266.
Hart, P., & Saunders, C. 1997. Power and Trust: Critical Factors in the Adoption and Use of
Electric Data Interchange. Organization Science, 8(1): 23-42.
Heide, J. B., & John, G. 1990. Alliances in Industrial Purchasing: The Determinants of Joint
Action in Buyer-Supplier Relationships. Journal of Marketing Research, 27(1): 24-36.
Hirsch, P. M., Friedman, R., & Koza, M. P. 1990. Collaboration or Paradigm Shift?: Caveat
Emptor and the Risk of Romance with Economic Models for Strategy and Policy
Research. Organization Science, 1(1): 87-97.
Hoegl, M., & Wagner, S. M. 2005. Buyer-Supplier Collaboration in Product Development
Projects. Journal of Management, 31(4): 530-548.
Hofer, A. R., Knemeyer, A. M., & Murphy, P. R. 2012. The Roles of Procedural and Distributive
Justice in Logistics Outsourcing Relationships. Journal of Business Logistics, 33(3):
196-209.
Hofstede, G. 1984. Culture's consequences: International differences in work-related values:
sage.
Hult, G. T. M., Ketchen Jr, D. J., & Nichols Jr, E. L. 2003. Organizational Learning as a
Strategic Resource in Supply Management. Journal of Operations Management, 21(5):
541-556.

180

James, L. R., & Brett, J. M. 1984. Mediators, Moderators, and Tests for Mediation. Journal of
Applied Psychology, 69(2): 307.
Kale, P., & Singh, H. 2009. Managing Strategic Alliances: What Do We Know Now, and Where
Do We Go From Here? Academy of Management Perspectives, 23(3): 45-62.
Katz, J. 2001. Analytic Induction. In N. J. Smelser, & P. B. Baltes (Eds.), International
Encyclopedia of the Social & Behavioral Sciences, Vol. 1: 480-484. Oxford, UK:
Elsevier.
Kee-hung, L. 2009. Linking Exchange Governance with Supplier Cooperation and Commitment:
A Case of Container Terminal Operations. Journal of Business Logistics, 30(1): 243263.
Lampel, J., & Giachetti, C. 2013. International Diversification of Manufacturing Operations:
Performance Implications and Moderating Forces. Journal of Operations Management,
31(4): 213-227.
Langley, M., White, J. B., & Boudette, N. 2006. In Alliance Talks, GM to Seek Billions From
Renault, Nissan. (cover story). Wall Street Journal - Eastern Edition, 248(74): A1-A10.
LaReau, J. 2006. Kerkorian Proposal Surprises GM. Automotive News, 80(6210): 8-8.
Laursen, K., & Salter, A. 2006. Open for Innovation: The Role of Openness In Explaining
Innovation Performance Among U.K. Manufacturing Firms. Strategic Management
Journal, 27(2): 131-150.
LaVean, G. 1980. Interoperability in Defense Communications. Communications, IEEE
Transactions on, 28(9): 1445-1455.
Lee, H. L. 2004. The Triple-A Supply Chain. Harvard Business Review, 82(10): 102-112.
Lee, R. P., & Johnson, J. L. 2010. Managing Multiple Facets of Risk in New Product Alliances*.
Decision Sciences, 41(2): 271-300.
Lipparini, A., Lorenzoni, G., & Ferriani, S. 2014. From Core to Periphery and Back: A Study on
the Deliberate Shaping of Knowledge Flows in Interfirm Dyads and Networks. Strategic
Management Journal, 35(4): 578-595.
Luca, L. M. D., & Atuahene-Gima, K. 2007. Market Knowledge Dimensions and CrossFunctional Collaboration: Examining the Different Routes to Product Innovation
Performance. Journal of Marketing, 71(1): 95-112.
Luzzini, D., Caniato, F., Ronchi, S., & Spina, G. 2012. A Transaction Costs Approach to
Purchasing Portfolio Management. International Journal of Operations & Production
Management, 32(9): 1015-1042.

181

Manning, P. K. 1982. Analytic Induction. In P. K. Manning, & R. B. Smith (Eds.), A Handbook
of Social Science Methods 273-302 Cambridge, MA
: Ballinger.
McAfee, B., Glassman, M., & Honeycutt Jr, E. D. 2002. The Effects of Culture and Human
Resource Management Policies on Supply Chain Management Strategy. Journal of
Business Logistics, 23(1): 1-18.
McIvor, R. 2009. How the Transaction Cost and Resource-Based Theories of the Firm Inform
Outsourcing Evaluation. Journal of Operations Management, 27(1): 45-63.
Melnyk, S., Davis, E., Spekman, R., & Sandor, J. 2010. Outcome-Driven Supply Chains. MIT
Sloan Management Review, 51(2): 33-38.
Mentzer, J. T., DeWitt, W., Keebler, J. S., Min, S., Nix, N. W., Smith, C. D., & Zaharia, Z. G.
2001. Defining Supply Chain Management. Journal of Business Logistics, 22(2): 1.
Mills, J. F., & Camek, V. 2004. The Risks, Threats and Opportunities of Disintermediation: A
Distributor's View. International Journal of Physical Distribution & Logistics
Management, 34(9): 714-727.
Mohr, J., & Spekman, R. 1994. Characteristics of Partnership Success: Partnership Attributes,
Communication Behavior, and Conflict Resolution Techniques. Strategic Management
Journal, 15(2): 135-152.
Moldoveanu, M. C., & Baum, J. A. C. 2011. "I Think You Think I Think You're Lying": The
Interactive Epistemology of Trust in Social Networks. Management Science, 57(2): 393412.
Montoya, M. M., Massey, A. P., Hung, Y.-T. C., & Crisp, C. B. 2009. Can You Hear Me Now?
Communication in Virtual Product Development Teams*. Journal of Product
Innovation Management, 26(2): 139-155.
Mooney, C. Z., & Duval, R. D. 1993. Bootstrapping: A Nonparametric Approach To Statistical
Inference: Sage.
Mosquet, X., Russo, M., Wagner, k., Zablit, H., & Arora, A. 2014. Accelerating Innovation:
New Challenges for Automakers, BCG Perspectives, Vol. January 2014. Boston: Boston
Consulting Group.
Mouzakitis, S., Sourouni, A.-M., & Askounis, D. 2009. Effects of Enterprise Interoperability on
Integration Efforts in Supply Chains. International Journal of Electronic Commerce,
14(2): 127-155.
Nidumolu, R., Ellison, J., Whalen, J., & Billman, E. 2014. The Collaboration Imperative. (cover
story). Harvard Business Review, 92(4): 76-84.

182

Novak, S., & Eppinger, S. D. 2001. Sourcing by Design: Product Complexity and the Supply
Chain. Management Science, 47(1).
Nunnally, J. C. 1978 Psychometric Theory. New York: McGraw-Hill Book Company.
Nunnally, J. C., Bernstein, I. H., & Berge, J. M. t. 1967. Psychometric Theory: McGraw-Hill
New York.
Nyaga, G. N., Whipple, J. M., & Lynch, D. F. 2010. Examining Supply Chain Relationships: Do
Buyer and Supplier Perspectives on Collaborative Relationships Differ? Journal of
Operations Management, 28(2): 101-114.
Oliver, C. 1997. Sustainable Competitive Advantage: Combining Institutional and ResourceBased Views. Strategic Management Journal, 18(9): 697-713.
Olson, M. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups
(twentieth ed.). Cambridge Massachusetts, London England: Harvard University Press.
Otjacques, B., Hitzelberger, P., & Fernand, F. 2007. Interoperability of E-Government
Information Systems: Issues of Identification and Data Sharing. Journal of Management
Information Systems, 23(4): 29-51.
Parise, S., & Casher, A. 2003. Alliance Portfolios: Designing And Managing Your Network of
Business-Partner Relationships. Academy of Management Executive, 17(4): 25-39.
Park, S. H., & Ungson, G. R. 2001. Interfirm Rivalry and Managerial Complexity: A Conceptual
Framework of Alliance Failure. Organization Science, 12(1): 37-53.
Perols, J., Zimmermann, C., & Kortmann, S. 2013. On The Relationship Between Supplier
Integration and Time-To-Market. Journal of Operations Management, 31(3): 153-167.
Peugh, J. L., & Enders, C. K. 2004. Missing Data in Educational Research: A Review of
Reporting Practices and Suggestions for Improvement. Review of Educational Research,
74(4): 525-556.
Prahinski, C., & Benton, W. C. 2004. Supplier Evaluations: Communication Strategies to
Improve Supplier Performance. Journal of Operations Management, 22(1).
Rosen, A. M. 2006. Identification and Estimation of Firms' Marginal Cost Functions With
Incomplete Knowledge of Strategic Behavior. The Institute for Fiscal Studies
Department of Economics, University College London, CeMMAP Working Paper
CWP03/07.
Rothenberg, S., & Ettlie, J. E. 2011. Strategies to Cope with Regulatory Uncertainty in the Auto
Industry. California Management Review, 54(1): 126-144.

183

Sampson, R. C. 2007. R&D Alliances and Firm Performance: The Impact of Technological
Diversity and Alliance Organization on Innovation. Academy of Management Journal,
50(2): 364-386.
Sanders, N. R. 2008. Pattern of Information Technology Use: The Impact on Buyer–Suppler
Coordination and Performance. Journal of Operations Management, 26(3): 349-367.
Schein, E. H. 1985. Organizational Culture and Leadership. San Francisco, CA: Josse-Bass
Inc.
Schoenherr, T., & Swink, M. 2012. Revisiting the Arcs of Integration: Cross-Validations and
Extensions. Journal of Operations Management, 30(1/2): 99-115.
Siemsen, E. 2011. The Usefulness of Behavioral Laboratory Experiments in Supply Chain
Management Research. Journal of Supply Chain Management, 47(3): 17-18.
Simon, H. A. 1979. Rational Decision Making in Business Organizations. The American
economic review: 493-513.
Stuart, F. I. 1997. Supplier Alliance Success and Failure: A Longitudinal Dyadic Perspective.
International Journal of Operations & Production Management, 17(6): 539.
Uzzi, B. 1997. Social Structure and Competition in Interfirm Networks: The Paradox of
Embeddedness. Administrative Science Quarterly, 42(1): 35.
Varadarajan, P. R. 1985. A Two-factor Classification of Competitive Strategy Variables.
Strategic Management Journal, 6(4): 357-375.
Villena, V. H. R., Elena; Choi, Thomas Y. 2011. The Dark Side of Buyer–Supplier
Relationships: A Social Capital Perspective. Journal of Operations Management, 29(6):
561-576.
Wagner, S. M. 2012. Tapping Supplier Innovation. Journal of Supply Chain Management,
48(2): 37-52.
Weinstein, R. 2005. RFID: A Technical Overview and its Application to the Enterprise. IT
Professional, 7(3): 27-33.
Wernerfelt, B. 1984. A Resource-Based View of the Firm. Strategic Management Journal, 5(2).
Williamson, O. E. 1979. Transaction-Cost Economics: The Governance of Contractual
Relations. Journal of Law & Economics, 22(2): 233-261.
Williamson, O. E. 2008. Outsourcing: Transaction Cost Economics And Supply Chain
Management*. Journal of Supply Chain Management, 44(2): 5-16.
Yin, R. K. 1981. The Case Study Crisis: Some Answers. Administrative Science Quarterly,
26(1): 58-65.

184

Yin, R. K. 2009. Case Study Research: Design and Methods (4th ed.): Sage Publications.
Zaheer, A., & Venkatraman, N. 1995. Relational Governance as an Interorganizational Strategy:
An Empirical Test of the Role of Trust in Economic Exchange. Strategic Management
Journal, 16(5): 373-392.
Zhang, C., Henke, J. W., & Griffith, D. A. 2009. Do Buyer Cooperative Actions Matter Under
Relational Stress? Evidence from Japanese and U.S. assemblers in the U.S. automotive
industry. Journal of Operations Management, 27(6): 479-494.
Zhang, C., Viswanathan, S., & Henke Jr, J. W. 2011. The boundary spanning capabilities of
purchasing agents in buyer-supplier trust development. Journal of Operations
Management, 29(4): 318-328.
Zhou, H., Benton, W. C., Schilling, D. A., & Milligan, G. W. 2011. Supply Chain Integration
and the SCOR Model. Journal of Business Logistics, 32(4): 332-344.

185