. .. t a . ‘cu. a. .
5,33... mg.
q a}... 1 a .

., . ....

.i

.94“.

1;. 1. .
3.! 1

3.6.

:31... s , .
{Wuk , .
x! :angdu3ii. Lm

1 . I"\-‘.I
3.9.... Lug}?

2v .
£4 2!}- :3
5 25.13.?!
h :1 .
5d. 2 z .

9.... (iv.
339:! 6
I3.\.Lv

 

 

 

 

 

This is to certify that the
dissertation entitled

AN INVESTIGATION OF THE SIMULTANEOUS IMPACT OF
LAUNCH AND LOGISTICS STRATEGIES ON NEW PRODUCT
PERFORMANCE

presented by

GILBERT NGARIH NYAGA

has been accepted towards fulﬁllment
of the requirements for the

Ph.D. degree in Department of Marketing and
Supply Chain Management

 

 

Major Professor’svSignature

0;”?4137? //I QFCL

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

_LlBPARY
Mich... . fltate
University

 

 

 

-u-InI-n-vno-.--—--.--.-.---a--.--‘-— -I-l-I-O-‘-I--a-v-I-O-i-

--n----I-O-I—I-o-O-O-l-l-O-I-o-o-o-O-l-I-l-n.

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd-p.1

 

 

AN INVESTIGATION OF THE SIMULTANEOUS IMPACT OF LAUNCH AND
LOGISTICS STRATEGIES ON NEW PRODUCT PERFORMANCE

By

Gilbert Ngarih Nyaga

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Marketing and Supply Chain Management

2006

ABSTRACT

AN INVESTIGATION OF THE SIIVIULTANEOUS IMPACT OF LAUNCH AND
LOGISTICS STRATEGIES ON NEW PRODUCT PERFORMANCE

By

Gilbert Ngarih Nyaga

This research investigates how ﬁrms’ logistics and launch strategies impact new
product performance under different competitive, demand, product, and ﬁrms
characteristics. Speciﬁcally, the research examines new product launch in companies
that deal with durable and non-durable consumer products. Based on data from three
case studies, a simulation model is developed to examine the overall and interaction
effects of logistics strategies, launch strategies, demand characteristics, product
characteristics, and ﬁrm resources on new product performance. New product
performance is examined in terms of customer service (as measured by order and case ﬁll
rates), average system inventory, and proﬁtability.

Results from simulation experiments indicate that logistics strategy is a major
determinant of new product performance while launch strategy affects inventory and
proﬁtability but not customer service. Logistics and launch strategy have a signiﬁcant
interaction effect on performance. Moreover, demand uncertainty, competitive
environment, product characteristics, and ﬁrm resources impact new product
performance either directly or interactively.

Managerial implications are discussed and a framework identifying appropriate
strategy in different situations is provided. Suggestions for ﬁiture research are also

provided.

Copyright by
GILBERT NGARIH NYAGA
2006

Dedicated to my sister, Agnes Mairu who taught me how to write my name
and sacriﬁced her studies so that I continue with mine.

iv

ACKNOWLEDGEMENTS

I wish to acknowledge the assistance and support of several people who have
made it possible for me to complete this dissertation and my doctoral program.

I would like to thank my dissertation committee for guiding me, reading previous
drafts, and offering invaluable insights. In particular, I thank Dr. David J. Closs,
committee chair, for supporting and mentoring me throughout my doctoral studies and
expertly guiding me in dissertation research and writing; Dr. Roger J. Calantone for
encouraging me to explore new product launch and for his constructive criticisms; Dr.
Regina McNally for her thorough review of previous draft; and Dr. Alexandre M.
Rodrigues for his help in simulation development. Their patience, criticisms, and
encouragement made it possible for me to complete the dissertation.

I am grateful to faculty, staff, and doctoral colleagues in the Department of
Marketing and Supply Chain Management for their assistance and inspiration throughout
the doctoral studies. I am especially grateful to Prof. Robert Nason, Department Chair,
for providing means of ﬁnancial support throughout, to Cheryl Lundeen, member of staff,
for readily assisting me in many ofﬁce related tasks, and to M. Doug Voss, my cheerful
ofﬁcemate, colleague, and friend.

I am indebted to Dr. Allan M. Winkler, Professor of History at Miami University,
OH, for encouraging and helping me pursue graduate studies in the US. and Dr. Thomas
W. Speh, Professor of Marketing at Miami University, OH, for mentoring and

encouraging me to pursue a doctorate in the logistics and supply chain management area.

I am also indebted to my family for their love and faith in me while am so far
away from home. My parents, Nyaga Karundu and Mbuya Nyaga, for teaching me the
value of hard work; my brothers and sisters for always being there for me; my nieces and
nephews for challenging me to be their great uncle and role model — I couldn’t let them
down; and my wife, Lucy Muthoni, for being by my side in the tough journey that the
doctoral program has been. Her love, support, patience, and understanding have made

my dream come true.

vi

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. ix
LIST OF FIGURES ............................................................................................................. x
CHAPTER I: INTRODUCTION ......................................................................................... 1
1.0 Introduction ............................................................................................................. 1

1.1 Background .............................................................................................................. 1
1.1.] New Product Performance .............................................................................. 3

1.1.2 Logistics Strategy ............................................................................................ 4

1.1.3 Launch Strategy .............................................................................................. 7

1.1.4 Interaction between Launch Strategy and Logistics Strategy ......................... 9

1.1.5 Environmental Factors ................................................................................. 12

1.1.6 Product Characteristics ................................................................................ 13

1.1.7 Firm Resources ............................................................................................. 14

1.2 Research Objectives ............................................................................................... 14

1.3 Research Questions ................................................................................................ 15

1.4 Research Method ................................................................................................... 15

1.5 Justiﬁcation of Study ............................................................................................. 17

1.6 Organization of Chapters ....................................................................................... 18
CHAPTER H: LITERATURE REVIEW .......................................................................... 19
2.0 Introduction ............................................................................................................ 19

2.1 Issues Related to New Product Launch ................................................................. 19
2.2 Determinants of New Product Performance .......................................................... 20
2.2.1 Logistics Strategies ....................................................................................... 22

2.2.2 Launch Strategy ............................................................................................ 27

2.2.3 Environmental Factors ................................................................................. 29

2.2.4 Product Characteristics ................................................................................ 33

2.2.5 Firm Resources ............................................................................................. 34

2.3 Models and Techniques for Launch Analysis ........................................................ 35
2.3.1 Case Study Research ..................................................................................... 36

2.3.2 Simulation Modeling Review ........................................................................ 3 7

2.4 Conceptual Model and Research Hypotheses ........................................................ 39
2.5 Conclusion ............................................................................................................. 43
CHAPTER HI: RESEARCH METHOD ........................................................................... 45
3.0 Introduction ............................................................................................................ 45

3.1 Case Study Research .............................................................................................. 45
3.1.1 Case Study Design ........................................................................................ 46

3.1.2 Evaluating Quality of Research Design ........................................................ 48

3.1.3 Data Collection ............................................................................................. 50

3.1.4 Data Analysis ................................................................................................ 51

vii

3.1.5 Case Study] .................................................................................................. 51

3.1.6 Case Study 11 ................................................................................................. 54
3.1.7 Case Study 111 ................................................................................................ 55
3.1.8 Key Findings from Case Studies ................................................................... 5 7

3.2 Simulation Modeling ............................................................................................. 59
3.2.] Simulation Model .......................................................................................... 59
3.2.2 Model Verification and Validation ............................................................... 64
3.2.3 Experimental Factors .................................................................................... 66
3.2.4 Fixed Parameters .......................................................................................... 72
3.2.5 Performance Variables ................................................................................. 73

3. 2. 6 Number of Replications ................................................................................. 77

3. 2. 7 Data Analysis ................................................................................................ 77
CHAPTER IV: RESULTS ................................................................................................. 79
4.0 Introduction ............................................................................................................ 79
4.1 MANOVA Assumptions and Data Conformity to Assumptions ........................... 79
4.2 Overall Results ....................................................................................................... 82
4.3 Multivariate Results ............................................................................................... 83
4.4 Univariate Results .................................................................................................. 86
4.4.1 Univariate Results: Three-way Interactions ................................................. 86
4.4.2 Univariate Results: T wo-way Interactions ................................................... 94
4.4.3 Univariate Results: Main Effects ................................................................ 106

4.5 Hypotheses Testing .............................................................................................. 114
4.5.1 Hypotheses H1 a .......................................................................................... 114
4.5.2 Hypotheses H1b .......................................................................................... 114
4.5.3 Hypotheses H20 .......................................................................................... 115
4.5.4 Hypotheses H2b .......................................................................................... 115
4.5.5 Hypotheses H3 ............................................................................................ 116
4.5.6 Hypotheses H4a .......................................................................................... 116

4. 5. 7 Hypotheses H4b .......................................................................................... 116
4.5.8 Hypotheses H5a .......................................................................................... 11 7
4.5.9 Hypotheses H5b .......................................................................................... I 18
4.5.10 Hypotheses H6 .......................................................................................... 119
4.5.11 Hypotheses H7 .......................................................................................... 119
4.5 Discussion of Results ........................................................................................... 121
CHAPTER V: IMPLICATIONS AND CONCLUSION ................................................ 128
5.0 Introduction .......................................................................................................... 128
5.2 Managerial Implications ...................................................................................... 128
5.3 Research Contribution ......................................................................................... 137
5.4 Limitations and Opportunities for Future Research ............................................. 138
5.5 Conclusion ........................................................................................................... 140
APPENDICES ................................................................................................................. 141
REFERENCES ................................................................................................................ 146

viii

LIST OF TABLES

Table 3.1 - Product Launch by Month and Market Segment ............................................. 69
Table 3.2 - Effect of Competitive Reaction on New Product Demand ............................. 71
Table 3.3 - Effect of High Product Advantage on New Product Demand ......................... 72
Table 3.4 - Production Cost Rate ............................................................................ I ........... 74
Table 3.5 - Transportation Cost Rate ................................................................................. 75
Table 3.6 - Inventory Carrying Cost Rate .......................................................................... 76
Table 3.7 - Stockout Penalty Cost ...................................................................................... 76
Table 4.1 - Skewness and Kurtosis Statistics ..................................................................... 80
Table 4.2 - Summary of Signiﬁcance Test ........................................................................ 83
Table 4.3 - Multivariate Test Results ................................................................................. 84
Table 4.4 - Univariate Test for Three-way Interactions .................................................... 86
Table 4.5 - Univariate Test for Two-way Interactions ...................................................... 94
Table 4.6 - Univariate Test for Main Effects ................................................................... 107
Table 4.7 - Summary of Hypotheses Tests ...................................................................... 120
Table 4.8 - Summary of Results Matrix 1 ........................................................................ 126
Table 4.9 - Summary of Results Matrix 2 ........................................................................ 127
Table 5.1 - Best Strategy Matrix ...................................................................................... 134
Table 5.2 - Strategy Analysis Chart 1 .............................................................................. 135
Table 5 .3 - Strategy Analysis Chart 2 .............................................................................. 136

ix

LIST OF FIGURES

Figure 1.1 - Logistics Strategy x Launch Strategy Matrix ................................................. 10
Figure 2.1 - Conceptual Framework .................................................................................. 39
Figure 3.1 - Supply Chain Network ................................................................................... 60
Figure 3.2 - List of Events ................................................................................................. 62
Figure 4.1 - Frequency histograms for the Dependent Variables ...................................... 81
Figure 4.2 - Effect of LOG*LAU*DEM on Order Fill Rate ............................................. 87
Figure 4.3 - Effect of LOG*LAU*DEM on Case Fill Rate ............................................... 88
Figure 4.4 - Effect of LOG*LAU*DEM on Average System Inventory .......................... 89
Figure 4.5 - Effect of LOG*LAU*DEM on Proﬁtability .................................................. 90
Figure 4.6 - Effect of LOG*LAU*CAP on Case Fill Rate ................................................ 91
Figure 4.7 - Effect of LOG*LAU*CAP on Proﬁtability ................................................... 92
F igurc 4.8 - Effect of LOG*DEM*CAP on Average System Inventory ........................... 93
Figure 4.9 - Effect of LOG*DEM*CAP on Proﬁtability .................................................. 93
Figure 4.10 - Effect of LOG*LAU on Order Fill Rate ...................................................... 95
Figure 4.11 - Effect of LOG*LAU on Case Fill Rate ........................................................ 96
Figure 4.12 - Effect of LOG*LAU on Average System Inventory ................................... 96
Figure 4.13 - Effect of LOG*LAU on Proﬁtability ........................................................... 97
Figure 4.14 - Effect of LOG*DEM on Order Fill Rate ..................................................... 98
Figure 4.15 - Effect of LOG*DEM on Case Fill Rate ....................................................... 98
Figure 4.16 - Effect of LOG*DEM on Average System Inventory ................................... 99
Figure 4.17 - Effect of LOG*DEM on Proﬁtability ........................................................ 100

Figure 4.18 - Effect of LOG*CAP on Order Fill Rate .................................................... 101

Figure 4.19 - Effect of LOG*CAP on Case Fill Rate ...................................................... 101
Figure 4.20 - Effect of LOG*CAP on Average System Inventory .................................. 102
Figure 4.21 - Effect of LOG*CAP on Proﬁtability ......................................................... 102
Figure 4.22 - Effect of LAU*DEM on Order Fill Rate ................................................... 103
Figure 4.23 - Effect of LAU‘tDEM on Case Fill Rate ..................................................... 103
Figure 4.24 - Effect of LAU*DEM on Average System Inventory ................................. 104
Figure 4.25 - Effect of LAU*DEM on Proﬁtability ........................................................ 104
Figure 4.26 - Effect of LAU*CAP on Average System Inventory .................................. 105
Figure 4.27 - Effect of DEM*CAP on Average System Inventory ................................. 106
Figure 4.28 - Effect of DEM*CAP on Proﬁtability ......................................................... 106
Figure 4.29 — Effect of Logistics Strategy on Order and Case Fill Rate .......................... 108
Figure 4.30 - Effect of Logistics Strategy on Average System Inventory ....................... 108
Figure 4.31 - Effect of Logistics Strategy on Proﬁtability .............................................. 108
Figure 4.32 - Effect of Launch Strategy on Order Fill Rate ............................................ 109
Figure 4.33 - Effect of Launch Strategy on Average System Inventory ......................... 110
Figure 4.34 - Effect of Launch Strategy on Proﬁtability ................................................. 110
Figure 4.35 - Effect of Demand Characteristics on Order and Case Fill Rate ................. 111
Figure 4.36 — Effect of Demand Characteristics on Average System Inventory ............. 112
Figure 4.37 - Effect of Demand Characteristics on Proﬁtability ..................................... 112
Figure 4.38 - Effect of Production Capacity on Order and Case Fill Rate ...................... 113
Figure 4.39 - Effect of Production Capacity on Average System Inventory ................... 113
Figure 4.40 - Effect of Production Capacity on Proﬁtability ........................................... 113

xi

CHAPTER I: INTRODUCTION

1.0 Introduction

Although the importance of logistics capabilities in the attainment of ﬁrms’
competitive goals is underscored in many studies (Bowersox et al. 1989; MSUGLRT
1995; Fawcett et a1. 1997; Daugherty et a1. 1998; Stank et al. 2003; Zacharia and Mentzer
2004), the impact of these capabilities on new product performance is rarely addressed
(Bowersox et a1. 1999; Di Benedetto 1999; Hart and Tzokas 2000; Calantone et a1. 2005).
This research seeks to ﬁll the gap by examining the impact of logistics strategies on new
product performance. Speciﬁcally, the research examines new product launch process in
three companies that deal with durable and non-durable consumer products. Based on
data from these case studies, a simulation model is developed to examine the overall and
interaction effects of logistics strategies, launch strategies, demand characteristics, and

ﬁrm size on new product performance.

1.1 Background

The decisions and activities necessary to deliver a new product to its target market
have been referred to under the collective terms of launch strategy, market entry, product
launch, product introduction, or market launch (Hultink and Hart 1998). Extant literature
shows that new product launch is the most expensive, risky and time-consuming stage in
the new product development process (Guiltinan 1999; Droge et al. 2000; Calantone et al.
2005). Product delivery process greatly impacts new product performance (Cooper and
Kleinschmidt 1993) while distribution support is critical where market acceptance is

difﬁcult to achieve and diffusion is slow (Guiltinan 1999). Consequently, logistics

capabilities can improve new product launch success and minimize losses from
unsuccessful launches (Calantone et al. 2005). Companies can reduce the cost of new
product launch by prioritizing inventory deployment and emphasizing ﬂexible and
responsive logistics capabilities to mitigate risks associated with demand uncertainties
(Fisher 1997; Pagh and Cooper 1998; Bowersox et al. 1999; Di Benedetto 1999).
However, few studies have examined the simultaneous effect of launch strategies and
logistics strategies on new product performance in different competitive and demand
situations.

A few studies have developed conceptual frameworks to examine the role of
logistics and supply chain strategies in new product launch process. For example, Fisher
(1997) examines the type of supply chain strategy that is most suitable for given
products. He identiﬁes two product categories, functional and innovative products, and
two supply chain strategies, efﬁcient and responsive strategies. He posits that functional
products require efﬁcient supply chain processes while innovative products require
responsive supply chain processes. Therefore, he recommends that companies
introducing new products should match their supply chain strategies to the appropriate
product type.

Bowersox et al. (1999) posit that ﬁrms adopt two alternative strategies when
launching new product: traditional anticipatory launch strategies and lean launch
strategies. Traditional launch is demand forecasts driven and is based on anticipatory
logistics (push) while lean launch is premised on principles of postponement supported
by response-based logistics (pull) and supply chain capabilities. The authors argue that

lean launch strategy enhances new product introduction by minimizing out-of-stock

potential, allowing greater ﬂexibility in product variant, and reducing inventory exposure.
Closs et a1. (1998) empirically compared anticipatory and response-based strategies under
conditions of demand uncertainty and concluded that response-based strategies
outperform anticipatory strategies in terms of customer service delivered. However,
there are no empirical tests to validate Bowersox et al. (1999)’s conceptualizations.

Calantone et al. (2005) further explore Bowersox et al.’s (1999) arguments by
presenting an assessment of potential pitfalls of traditional anticipatory launch strategy
and ways in which the alternative lean launch mitigates those problems. Using two
company cases (Dell Computers and Benetton), the authors demonstrate the successful
application of lean launch methods supported by ﬂexible supply chain capabilities.
However, the study does not assess the impact of prevailing competitive environment and
stipulations are not empirically tested.

This research seeks to empirically test relationships and conceptualizations
stipulated in the aforementioned studies, and present a comprehensive framework to help
in quantitatively determining the simultaneous impact of ﬁrms’ logistics and launch
strategies, competitor response, product characteristics, and demand characteristics on

new product performance.

1.1.1 New Product Performance

Several measures of new product performance are examined in extant literature.
These measures have been categorized as ﬁnancial performance (i.e., proﬁts, sales, costs,
etc.), market acceptance/share (i.e., market position, sales levels, sales performance

compared to competition, etc.), and technical performance (i.e., product quality, technical

performance, etc.) (Montoya-Weiss and Calantone 1994; Green, Barclay and Ryan 1995;
Hultink and Robben 1999). The performance measures examined in this research are:
customer service, average system inventory, and proﬁtability. These measures have been
widely used in past studies (Chow et al. 1994; Closs et al. 1998; Glasserman and Wang
1998; Song 1998; Gunasekaran et al. 2001; Bowersox et al. 2002; Haughton 2002; Grifﬁs
et al. 2004).

Customer service is measured in terms of order ﬁll rate (percentage of customer
orders ﬁlled complete within a speciﬁc time frame) and case ﬁll rate (percentage of cases
that are delivered complete within a speciﬁed time period). Average system inventory is
examined in terms of system-wide inventory holdings at a given stage and time in the
logistics system (Closs et al. 1998). Proﬁtability is derived from total sales less the

product cost, inventory carrying cost, transportation cost, and stockout penalty.

1.1.2 Logistics Strategy

According to the Council of Supply Chain Management Professionals (CSCMP),
logistics management is that part of supply chain management that plans, implements,
and controls the efﬁcient, effective forward and reverse ﬂow and storage of goods,
services and related information between the point of origin and the point of consumption
in order to meet customers' requirements (www.cscmp.org). Thus, logistics strategy can
be viewed as a set of objectives, decisions and activities associated with logistics
management. In previous studies, two types of logistics strategy relating to new product
introduction are examined: anticipatory and response-based strategies (Fisher 1997;

Bowersox et al. 1999; Calantone et al. 2005).

In anticipatory strategy, sufﬁcient pre-introduction inventory is deployed in the
distribution channels so as to meet forecasted demand. Although forecasting techniques
have improved signiﬁcantly in recent years, it is almost impossible to predict with
absolute precision the expected sales and likely performance of a new product when
introduced in the market. Thus, anticipatory logistics strategy has several risks and cost
implications, which can negatively affect ﬁnancial and market results whether the launch
is a success or failure (Bowersox et al. 1999; Calantone et a1. 2005). These include high
inventory carrying costs due to overstocking, lost sales due to inventory shortages (under-
stocking), and limited product variation, which may lead to customer dissatisfaction.

In a response-based strategy, a ﬂexible logistics system capable of responding
rapidly to changes in actual demand requirements is used to support limited inventory
commitment during new product introduction. The strategy enables better management
of risks associated with demand uncertainties through ﬂexible and responsive delivery
capabilities supported by upstream supply chain coordination (Bowersox et al. 1999).
The strategy is based on postponement principles (Zinn and Bowersox 1988; C00per
1993; Pagh and Cooper 1998; Waller et al. 2000). Postponement reduces the risk of
improper inventory deployment and the need to stock inventory for all product variations,
which contributes to reduction in inventory carrying costs. Consequently, response-based
strategy mitigates cost concerns associated with anticipatory strategy. However, smaller
shipment quantities and ﬂexible responses result in shipment cost increases, which if not
well planned can signiﬁcantly increase the cost of launch.

Bowersox et al. (1999) posit that four strategic issues in new product literature

have signiﬁcant implications on logistics processes: what to launch, where to launch,

when to launch, and how to launch. These are related to Hultink et al. (1998)’s launch
strategy classiﬁcation where the ﬁrst three relate to the what, where, and when to launch
strategic decision questions while how to launch is similarly stipulated as a tactical
decision question. First, what to launch raises the question of the ability of existing
logistics processes to support the new product deployment. For example, if the product is
perishable, can the current logistics system support its delivery to customers on time and
in good condition? Second, where to launch addresses accessibility of markets via
existing logistics systems. For example, a ﬁrm might avoid a market that has poor
infrastructure since logistics costs at launch may be excessive. However, where
infrastructure is good, the new product deployment can be effectively planned and
executed at minimal logistics costs. Third, when to launch hinges on capacity of existing
logistics systems to support the anticipated scale of launch. For example, even if
production capacity can be ramped up to meet seasonal demand, the volume deployed
may be constrained by the capacity of existing logistics system. Finally, how to launch
reﬂects the ﬁrms’ launch strategy.

The foregoing discussion indicates that logistics strategy inﬂuences new product
launch planning and performance. Indeed, as Di Benedetto (1999) posits, most
successful new product launches are characterized by greater perceived involvement of
the logistics functions in marketing, sales, distribution, and sales planning. Integrating
logistics functions with other internal functions of the ﬁrm is expected to improve
coordination of products ﬂow and consequently improve the ﬁrrn’s ability to handle new
product demand uncertainties. Moreover, this coordination is important when handling

product recalls and other reverse logistics functions. Consequently, logistics strategy is

stipulated to directly impact new product performance. However, this has not been
empirically tested. Indeed, there is no empirical comparison of anticipatory and

response-based logistics strategies in new product launch literature.

1.1.3 Launch Strategy

Launch strategy is deﬁned as decisions and activities necessary to present a
product to its target market and begin to generate income from sales (Choffray and Lilien
1984; Yoon and Lilien 1985; Green et al. 1995; Hultink et al. 1998). Several studies
have attempted to determine a generic launch strategy. For example, Hultink et al. (1997;
1998) posit that managers generally make one set of launch decisions prior to beginning
of new product development (strategic decisions) and another set of decisions after
conceptualization and development of the product is complete (tactical decisions).
Strategic decisions address the what, where, when, and why questions while tactical
decisions address the how to launch question. Green et al. (1995) posit that launch
strategy consists of three managerial decisions that affect the product’s long-term
performance, viz., timing of entry, magnitude of investment, and competitive emphasis.
Hultink and Robben (1999) classify launch decisions into three categories, namely,
strategic/market decisions, timing-related decisions, and marketing mix decisions.

These launch strategy conceptualizations can be viewed as different ends of the
same spectrum. For example, the studies concur that entry timing is an important launch
strategy component. Moreover, marketing mix decisions are integral to launch strategy.
The magnitude of investment component stipulated in Green et al. ( 1995) is not only a

strategic decision, but is also related to marketing mix decisions (i.e., promotion

expenditure, distribution coverage, etc.). Therefore, to comprehensively examine and
understand launch strategy, it is important to address both strategic and tactical decisions.
This research examines distribution aspects of launch strategy in both strategic and
tactical context. Speciﬁcally, the research focuses on scale of launch as it impacts new
product performance. Distribution decisions are viewed in terms of launch size (i.e.
where to launch). Calantone et al. (2002) differentiate between large size and small size
launch strategy. They refer to these as fat and lean, respectively. In other studies, large
size launch where new product is introduced in all markets simultaneously is referred to
as national launch while small size launch in which product is only introduced in select
markets only, then gradually introduced in the remaining markets is referred to as rollout
strategy (Bronnenberg and Mela 2004).

In fat launch, a ﬁrm introduces new product in all markets simultaneously. For
example, if a ﬁrm has delineated its customer market into ten segments, and it launches
new product in all ten segments, then this is fat or national launch. In lean or rollout
launch, a ﬁrm introduces new product in select markets only then sequentially launches
in other markets. For example, if a ﬁrm in the above case launched new product in only
three of the ten segments, then another three, two, and two sequentially, then this is lean
or rollout launch. In this research, the terms national and rollout launch are used instead
of fat and lean launch, respectively so as to distinguish launch strategy from logistics
strategies.

Launch strategy is expected to directly impact new product performance.

Consequently, these alternative launch strategies will result in different new product

performance. However, no test has been done to determine which strategy is better in

which situations.

1.1.4 Interaction between Logistics Strategy and Launch Strategy

Although past studies suggest that logistics strategy and launch strategy directly
impact new product performance, there are no studies examining the interaction effect
between these strategies. According to Bowersox et al. (1999) and Calantone et al.
(2005), rollout launch is associated with response-based logistics system while national
launch is associated by anticipatory logistics. However these relationships have not been
empirically tested. Moreover, the conceptualization of launch strategy adopted in this
research differs from that posited by Bowersox et al. (1999) and Calantone et al. (2005).

The view adopted in this research clearly distinguishes launch strategy
dimensions from logistics strategy dimensions. For rollout launch, ﬁrms can either
produce sufﬁcient new product inventory to deploy to selected market segments based on
demand forecast or adopt a market-responsive approach in which production and
inventory deployment is done only after customer orders are received. Similarly, in
national launch, ﬁrms can deploy inventory to market based on forecasted demand or by
responding to customer orders. Thus, examining effectiveness of the four possible
combinations of logistics strategy and launch strategy may yield important managerial
and theoretical insights.

A 2x2 matrix is used to identify the possible outcomes of interaction between

launch strategy and logistics strategy, as shown below.

Figure 1.1: Logistics Strategy x Launch Strategy Matrix

 

 

Logistics strategy
Anticipatory Response-based
Rollout Opportunity loss? Okay?
Launch
Strategy Insufﬁcient
National Inefﬁciency? c a p a c ity"

 

 

 

 

In the top left box, ﬁrms adopt rollout launch supported by anticipatory logistics.
For example, a ﬁrm launches new product in three out of ten market segments by
producing sufﬁcient inventory to meet forecasted demand in those segments, and then
deploying the inventory into the distribution system in anticipation of forecasted sales.
This approach has two major risks. First, the ﬁrm may incur high inventory carrying cost
since inventory is deployed into the distribution system before commencement of sales.
If product fails, the ﬁrm will incur inventory carrying cost due to overstock, obsolescence
cost, and write-off costs as well as opportunity cost for tying capital in deployed
inventory. Secondly, if product is successful, the ﬁrm will have lost an opportunity to
increase sales, market share, and customer loyalty. If competitors respond quickly to the
new introduction, opportunity losses may have signiﬁcant long term implications.
Consequently, it is expected that rollout launch supported by anticipatory logistics has a
high risk of opportunity loss.

In the bottom left box, ﬁrms adopt national launch supported by anticipatory
logistics system. For example, a ﬁrm launches new product in all market segments by
deploying sufﬁcient inventory into the distribution system based on demand forecasts.
This approach has several cost implications. First, under uncertain demand environment,

the ﬁrm may overstock in some markets, or worse, in all markets thereby incurring high

10

inventory carrying costs and associated opportunity costs. Second, if product succeeds in
some market segments and fails in others, the ﬁrm may have to transship inventory from
poorly performing market segments to those segments with high demand and inventory
shortage. Finally, if new product is a failure, the company will incur high obsolescence
and write-off cost. Since launch is done in multiple market segments, the cumulative
inventory carrying costs may be extremely high, making this approach highly inefﬁcient.

In the top right box, ﬁrms adopt rollout launch supported by response-based
logistics. For example, a ﬁrm launches new product in three out of ten market segments
by deploying new product into the distribution system only upon receiving customer
orders. Since market segments in which new product is introduced are few, it is expected
that the ﬁrm will manage a response-based logistics system more effectively. In other
words, it is easier and more efﬁcient to manage a market-responsive logistics system
when serving a small market than when serving a large market. Therefore, it is expected
that supporting rollout launch with a response—based logistics is most appropriate.

In the bottom right box, ﬁrms adopt national launch supported by response-based
logistics. For example, a ﬁrm launches new product in all market segments by deploying
inventory into the distribution system only upon receiving orders from the customers.
The challenge with this approach is that managing a response-based logistics system for
all its customers may be extremely expensive and inefﬁcient. Furthermore, the ﬁrm may
not have sufﬁcient capacity (i.e., ramping up manufacturing or distribution capacity) to
respond appropriately to increased customer orders from many segments. Therefore, it is
expected that national launch supported by response-based logistics will have insufﬁcient

capacity problems.

11

This research makes an important contribution by testing proposed interaction
outcomes, and identifying the combination of logistics and launch strategies that impacts
new product performance most signiﬁcantly. Moreover, the research tests how these

interactions affect performance under different competitive and demand conditions.

1.1.5 Environmental Factors

Several environmental factors are stipulated to impact new product performance.
This study examines two factors: competitive response and demand uncertainty.
Competitive response is a set of decisions made by a ﬁrm in response to an observed new
product launch (Hultink and Langerak 2002), and have the effect of reducing demand for
the new product. They are inﬂuenced by several drivers. First, characteristics of new
product launch strategy such as product newness (innovativeness) triggers retaliatory
behavior from competitors especially in highly competitive markets (Kuester et al. 1999;
Debruyne et a1. 2002). Second, industry factors such as market structure inﬂuence the
intensity and speed of response (Bowman and Gatignon 1995; Debruyne et al. 2002). For
example, competitor response is more rapid and aggressive in markets exhibiting high
grth rates (Yoon and Lilien 1985). Third, market signaling, which refers to a set of
signals generated by the scope, intensity, and competitiveness of a new product launch is
stipulated to inﬂuence competitive response (Heil and Walters 1993). Finally,
competitive response tends to take longer when switching costs are high, market grth
is slow, time required for developing new products is long, and when reacting ﬁrm’s
market share is high (Bowman and Gatignon 1995). Although past studies demonstrate

that competitive response inﬂuences new product performance, no study has been done to

12

investigate its moderating effect on the impact of launch strategy and logistics strategy on
new product performance.

Demand for new products may be highly unpredictable. As a result, responsive
logistics processes are desirable since they enable better management of uncertainty. For
example, ﬁrms introducing new to market (ﬁrst time) products tend to rely on forecasts
based on potential customer surveys and focus groups. These forecasts are susceptible to
high margin of error, which can have signiﬁcant impact on launch cost and performance.
Thus, when making launch decisions, it is expected that ﬁrms will take cognizance of

prevailing demand uncertainty.

1.1.6 Product Characteristics

Extant literature shows that the main product characteristic inﬂuencing new
product launch planning is product advantage. Product advantage refers to the bundle of
tangible and intangible beneﬁts that a product offers to its customers that are unique and
superior to competitive brands (Cooper and Kleinschmidt 1987; Montoya-Weiss and
Calantone 1994; Li and Calantone 1998). Cooper and Kleinschmidt (1987) found that
product advantage was the number one factor inﬂuencing new product commercial
success. Other studies suggest that product advantage is associated with incumbent’s
competitive response and market acceptance, which may affect new product performance
Olap and Souder 1994; Hultink and Robbin 1999; Kuester et al. 1999; Debruyne et al.
2002). Consequently, it is stipulated that product advantage has a direct impact on new

product performance.

13

1.1.7 Firm Resources

Firm resources enable ﬁrms to implement strategies that provide competitive
advantage in the market place. Broadly perceived, they refer to the tangible and
intangible assets that ﬁrms use to develop and implement their strategies (Ray et al.
2004). According to the resource-based view of the ﬁrm, ﬁrm performance differences
arise because ﬁrms possess unique resources and capabilities that are rare, valuable,
inimitable, and non-substitutable, which enable them to appropriate above normal returns
(Wernerfelt 1984; Bamey 1991; Peteraf 1993). Therefore, ﬁrms leverage their resources
and capabilities to realize superior proﬁts and sustainable competitive advantage (Barney
1991; Peteraf 1993; Day 1994). In this research, it is stipulated that ﬁrms’ resource
endowment impacts new product performance. For example, a ﬁrrn’s transportation and
warehouse capacity may dictate how much inventory (and where) is deployed in the

market place, which directly affects new product performance.

1.2 Research Objectives

The overall objective of this research is to understand how logistics strategies
affect new product introduction, the interaction between logistics and launch strategies,
and the moderating effect of demand factors on new product performance. The effect
may vary with type of product and industry. Thus, the study examines companies dealing
with durable and non-durable consumer goods. Based on the case studies, a simulation
model is developed to enable realization of three main research objectives:

1. To examine how logistics and launch strategies affect new product performance.

14

2. To examine the relationship between logistics strategy and launch strategy during
new product introduction, and how this interaction affects performance.
3. To identify optimal combination of logistics and launch strategies that result in most

successﬁil new product launch under different competitive and demand situations.

1.3 Research Questions

To realize these objectives, the study addresses the following main research questions:

1. What effects do logistics and launch strategies have on new product performance?

2. What is the interaction effect between logistics and launch strategies on new product
performance?

3. How do competitive and demand situations affect the impact of logistics and launch
strategies on new product performance?

4. What effect does product characteristic have on new product performance?

5. What effect do ﬁrm resources have on new product performance?

1.4 Research Method

The research is done in two phases. First, case studies of companies that regularly
launch new products (the representation ﬁrms market durable consumer products and
non-durable consumer products) are done. Each case study identiﬁes the ﬁrrn’s new
product launch process including logistics strategies employed. Past studies have used
case study research to examine dynamics present within single settings (Eisenhardt 1989;
Miles and Huberman 1994; Yin 2003). According to Yin (2003:13), a case study is “an

empirical inquiry that investigates a contemporary phenomenon within its real-context,

15

especially when the boundaries between phenomenon and context are not clearly
evident.” The aim is to investigate contextual conditions that might be pertinent to the
phenomena under study.

In the second phase, a simulation design modeled after data ﬁom case studies is
developed. Simulation modeling technique is widely applied to many problems and
systems that are difﬁcult to feasibly construct or experiment upon in ‘real world’ or that
are too complex to model analytically (Rees et al. 2002). It is a tool for planning and
analysis that models real-world settings and estimates outcomes given a set of situational
characteristics (Closs et al. 1998). Unlike most analytic models, simulation models can
accommodate dynamic interrelationships and assumptions are less restrictive (Marucheck
and McClelland 1992). This means that simulation analysis enables a more detailed
assessment of interrelationships between logistics and launch strategies under different
competitive and demand situations. Input and output variables, as well as uncontrollable
factors for the simulation model were determined from the case studies. The resulting
simulation model was subjected to rigorous veriﬁcation and validation process as
stipulated by Sargent (2000).

The research is done in ﬁve stages. First, case study interviews are conducted to
collect data from selected ﬁrms. Second, new product launch process ﬂows of the
studied companies are mapped. Third, a system dynamic simulation model is developed,
based on the mapped process ﬂows and ideas derived from extant literature. Fourth,
simulation runs are conducted to generate data, which is used to test study hypotheses.
Finally, analysis and discussion of research ﬁndings is done. MANOVA and ANOVA

are used to analysis data.

16

1.5 Justification of Study

New products play a signiﬁcant role in a ﬁnn’s growth and competitiveness.
Studies show that recently introduced products account for a signiﬁcant portion of ﬁnn’s
proﬁts (Debruyne et al. 2002). Indeed, new product performance is positively related to
overall organizational performance (Montoya-Weiss and Calantone 1994; Grifﬁn and
Page 1996; Hultink et al. 1998). For example, Grifﬁn (1997) posits that new product
performance accounts for one-fourth of variability in organizational performance while
Terwiesch et al. (1998) report that new product performance explains, depending upon
the market context, about 30 percent of organizational proﬁtability variance. Product
launch is the most demanding new product development phase in terms of time, ﬁnances,
and managerial resources (Bowersox et al. 1999; Debruyne et a1. 2002). Cooper and
Kleinschmidt (1988) posit that 54 percent of all new product expenditures are spent on
market launch compared to 39 percent spent on product development and 7 percent on
pre-development activities. In spite of the demonstrated importance of the launch stage,
there is limited research on the issue especially with respect to logistics strategy effects.

Thus, this study makes four important contributions. First, it investigates the role
of logistics strategies in enabling successful new product launch. This is crucial
considering the growing importance of logistics capabilities as sources of sustainable
competitive advantage in contemporary business operations (Olavarietta and Ellinger
1994; MSUGLRT 1994; Fawcett et a1. 1997; Lynch et al. 2000). Second, it considers
interaction effect between logistics and launch strategies in inﬂuencing new product
performance. Third, it considers environmental factors that moderate the impact of

logistics and launch strategies on new product performance. Finally, the research

17

employs simulation modeling, which allows researchers to understand the impact of
multiple decision factors and identify appropriate alternatives. Thus, it demonstrates the

use of simulation analysis in new product launch research.

1.6 Organization of Chapters

Chapter 1 provided an overview of dissertation focus including research
motivation, objectives, questions, and justiﬁcation. Chapter 2 presents a review of
relevant literature leading to conceptual model and research hypotheses. Chapter 3
discusses research design, including research approaches used in the study — case study
and simulation analysis. Chapter 4 presents simulation data analysis and results as well
as research ﬁndings. Chapter 5 discusses implications of research ﬁndings, research

contribution, limitations, and suggestions for future research.

18

CHAPTER II: LITERATURE REVIEW
2.0 Introduction
This chapter presents a review of literature on factors associated with new product
performance and highlights major research ﬁndings and gaps. The chapter is organized
into four sections: (a) review of issues related to new product launch, (b) determinants of
new product performance, (0) discussion of models and techniques for launch analysis
used in the study, and (d) conceptual model and research hypotheses. The chapter

concludes with a summary of key literature gaps, which the study addresses.

2.1 Issues Related to New Product Launch

Past studies indicate that new product launch represents the largest investment,
most expensive, risky, and time consuming phase in the entire new product development
process (Urban and Hauser 1993; Guiltinan 1999; Droge et al. 2000; Calantone et al.
2005). As Guiltinam ( 1999) suggests, this is attributed to a combination of production
and marketing expenditures incurred once a decision to launch is approved. Most new
product studies examine either the marketing component (market entry, promotions, etc)
or development process (proﬁciency of technical skills, new product development teams,
etc) (Montoya-Weiss and Calantone 1994; Henard and Szymanski 2001). The role of
logistics functions in the launch process has received limited attention (Bowersox et al.
1999; Di Benedetto 1999). Di Benedetto (1999) posits that greater involvement of
logistics functions in marketing, sales, distribution, inventory, and service planning
during launch can greatly enhance new product performance. Indeed, logistics

capabilities are highly intertwined with launch processes such as entry timing,

19

distribution planning, promotion planning, etc. Therefore, this research seeks to advance
new product launch knowledge by examining the simultaneous impact of launch and
logistics strategies on new product performance.

New product performance is the degree of market success attained by a product at
market maturity or the point at which product market boundaries change (Green et al.
1995). Most common measures of performance include proﬁtability, market share, unit
sales, customer loyalty, etc. Three performance measures that are easier to model are
examined in this research: customer service, system inventory, and proﬁtability. These
measures have been successfully applied in past studies (Chow et a1. 1994; Glasserman

and Wang 1998; Closs et al. 1998; Song 1998; Bowersox, et al. 2002).

2.2 Determinants of New Product Performance

Extant literature shows that several factors impact new product performance
(Cooper and Kleinschmidt 1987; Montoya-Weiss and Calantone 1994; Yap and Souder
1994; Henard and Szymanski 2001). Cooper and Kleinschmidt (1987) identiﬁed the key
new product success factors as product advantage (i.e., innovativeness), proﬁciency of
predevelopment activities (i.e., initial screening, marketing research, business analysis,
etc.), and protocol (i.e., clear deﬁnition of target market, product speciﬁcations, etc.).
Other factors included proﬁciency of technological and market-related activities,
technological synergy, marketing synergy, and market potential.

Yap and Souder (1994) identify seven factors that inﬂuence new product
development, and commercial success or failure. These are project synergies (i.e.,

marketing, engineering and production skill synergies), skills level (i.e., project

20

manager’s management and marketing skills, manufacturing skills, etc.), technology
sources (i.e., licensing, key personnel, etc.), product characteristics (i.e., product features,
compatibility, etc.), market characteristics (i.e., market growth, number of competitors,
etc.), entry strategies (i.e., pricing, distribution, positioning, etc.), and organizational
characteristics (i.e., interdepartmental communication, top management involvement,
etc.). These factors are similar to those suggested by Cooper and Kleinschmidt (1987).
For example, both studies suggest that market characteristics, technology aspects, product
characteristics, synergies, and launch proﬁciency determine new product success.

Subsequent meta-analysis studies by Montoya-Weiss and Calantone (1994) and
Henard and Szymanski (2001) build on these classiﬁcations to develop a more reﬁned
taxonomy of determinants of new product performance. Montoya-Weiss and Calantone
(1994) conducted a meta-analysis of antecedents of new product performance in which
they classify the determinants discussed in literature into four categories. These
categories are strategic factors (i.e., product advantage, technology and market synergies,
company resources, etc.), development process factors (i.e., protocol, speed to market,
technical proﬁciency, top management support, etc.), market environment factors (i.e.,
market competitiveness, market potential, etc.), and organizational factors (i.e., degree of
coordination and cooperation between and within ﬁrms, ﬁrm structure, organizational
climate, etc.). These factors are similar to those identiﬁed by Cooper and Kleinschmidt
(1987) and Yap and Souder (1994). However, the classiﬁcation is more succinct.

Henard and Szymanski (2001) also classify the determinants into four categories,
namely, product characteristics (i.e., product advantage, technological sophistication,

price, etc.), ﬁrm strategy characteristics (i.e., timing of entry, magnitude of resources

21

deployed, marketing and technological synergies, etc.), ﬁrm process characteristics (i.e.,
interfunctional coordination, proﬁciency, etc.), and market characteristics (i.e., market
potential, competitive intensity, demand characteristics, etc.). This classiﬁcation is a
reﬁnement of Montoya-Weiss and Calantone (1994). The only difference is that
predictors are placed in different categories. For example, Montoya-Weiss and
Calantone (1994) classify product advantages under strategic factors while Henard and
Szymanski (2001) classify them under products characteristics.

Most studies focus on main effects while ignoring interrelationships among
determinant variables (Henard and Szymanski 2001). Interactions between determinant
variables are expected to yield different performance outcomes. For example, new
product development processes directly affect product characteristics while market
characteristics such as competitive intensity may strongly inﬂuence a ﬁrm’s strategy and
launch budget. Consequently, it is important to examine interaction effects in detail. The
study addresses this gap by examining interaction effect of logistics and launch strategies,

product characteristics, demand characteristics, and ﬁrm size.

2.2.] Logistics Strategy

Logistics processes are inherently intertwined with new product introduction.
Bowersox et al. (1999) posit that four strategic issues in new product literature have
signiﬁcant implications on logistics processes: the what, where, when and how to launch.
These are related to Hultink et al.’s (1997; 1998) strategic—tactical launch strategy
decision questions. However, in spite of demonstrated importance of logistics functions

during new product launch, the role of logistics is substantially under-researched

22

(Bowersox et al. 1999, Di Benedetto 1999, Guiltinan 1999, Calantone et al. 2005). Thus,
more research is required to establish the extent to which logistics strategies impact new
product performance.

Three studies have addressed issues related to logistics strategy’s role during new
product introduction. The ﬁrst study, Fisher (1997), examines the most suitable match
between supply chain strategy and type of products. The author posits that developing an
effective supply chain strategy requires managers to take cognizance of the nature of
demand for the product being launched. For example, managers should consider such
factors as product life cycle, demand predictability, product variety, and market standards
for lead times and service (p.106). Based on the demand pattern, he classiﬁes products as
functional and innovative product. Functional products include staples that people buy in
a wide range of retail outlets such as grocery stores and gas stations. The demand for
these products tends to be stable and predictable, their life cycles tend to be long, and as a
result, they tend to be characterized by stiff competition, which leads to low proﬁt
margins. Innovative products on the other hand tend to have unpredictable demand and
shorter life cycles, and generally have higher proﬁt margins.

These product characteristics engender different supply chain strategies. For
example, the fairly predictable demand for functional products and low proﬁt margins
means that companies tend to focus more on proﬁt maximizing supply chain activities
with least costs. However, innovative products have high proﬁt margins and uncertain
demand, which increases the cost of lost sales and over-supplies (obsolescence cost).
Efﬁcient supply chain processes aim at supplying predictable demand efﬁciently at the

lowest possible cost while market-responsive supply chain process aim at responding

23

quickly to unpredictable demand in order to minimize stock-outs, forced markdowns, and
obsolete inventory. Therefore, functional products require efﬁcient supply chain
processes while innovative products require market-responsive processes. Companies
introducing new products should match their supply chain strategies to appropriate
product types since different products have unique delivery requirements, demand
characteristics, and proﬁtability potential. Fisher’s (1997) arguments underscore the need
to examine the impact of logistics strategies on new product performance. Moreover,
environmental conditions under which a speciﬁc match results in maximal performance
have not been investigated.

The second study, Bowersox et al. (1999), builds on Fisher’s (1997) ideas and
argues that ﬁrms launching new products adopt two alternative strategies: traditional
anticipatory launch strategies and lean launch strategies. Anticipatory strategy entails
sufﬁcient pre-introduction inventory deployment in the distribution channels so as to
meet forecasted demand. It is based on anticipatory logistics (push), in which inventory
to support projected sales are deployed into the distribution channel in anticipation of
expected sales. Lean launch strategy is based on postponement principles supported by
responsive logistics (pull) and supply chain capabilities. In this approach, only limited
inventory is deployed in the distribution channel during introduction rollout since ﬂexible
logistics system can support rapid response in case of early product success. The authors
conclude that lean strategy enhances new product performance by minimizing out-of-
stock potential, allowing greater ﬂexibility in product variant selection, and reducing
inventory exposure. However, although supported in literature, these arguments have not

been empirically tested. Moreover, the study focuses on logistics and supply chain

24

strategies during new product launch but gives limited attention to environmental factors
that may have profound impact on both logistics strategy and launch proﬁciency.

The third study, Calantone et al. (2005), builds on Bowersox et al.’s (1999)
arguments. The authors examine potential pitfalls of traditional anticipatory launch
strategy and how the alternative lean launch mitigates those pitfalls. Like Fisher (1997),
the authors posit that optimal performance can be realized by matching launch strategy
(or supply chain strategy) to market demand (which is related to product characteristics).
The authors cite two company cases (Dell Computers and Benetton) to make their
arguments and recommend greater inclusion of distribution and logistics functions in
launch planning. While the propositions make intuitive sense and are supported in
literature, no empirical or in-depth qualitative analysis has been done. This research will
empirically test these propositions.

Bowersox et a1. (1999) and Calantone et al. (2005) highlight three major risks
associated with anticipatory launch strategy. First, since this approach is premised on
demand forecasting, it is susceptible to uncertainties that accompany forecasts. As a
result, there is a risk of lost sales in case of under-stocking or high carrying cost in case
of overstocking. Secondly, pre-introduction inventory commitment makes it difﬁcult to
redeploy inventory from markets recording low success or outright failure to those that
are highly successful. In addition, it may not be possible to ramp up production to meet
demand requirements arising in some markets and even if this were possible, the
distribution system may not be able to cope with requirements in different market
segments. Consequently, there is likely to be inventory shortages in some markets and

excess inventory in other segments. Thirdly, due to demand uncertainty and attendant

25

challenges of accommodating forecast variance, anticipatory launch may be restricted to
one product variation. This is likely to negatively affect performance since customers are
increasingly demanding greater variety in product offering or product features.
Response-based strategy is shown to be a better strategy than anticipatory
approach because it enables better management of risks associated with demand
uncertainties (Bowersox et al. 1999). The strategy is premised on principles of
postponement (Alderson 1950; Bucklin 1965; Zinn and Bowersox 1988; Cooper 1993;
Pagh and Cooper 1998; Waller et al. 2000). Postponement reduces risk and uncertainty
costs tied to differentiation that occurs during manufacturing and logistics operations
(Pagh and Cooper 1998). Moreover, postponement allows a company to be ﬂexible in
developing different versions of a product as needed to meet changing customer needs
and to differentiate a product (Waller et al. 2000). In new product launch, postponement
enables reduction in risks associated with improper inventory deployment and the need to
stock inventory for all product variations. Consequently, the strategy mitigates cost
concerns associated with anticipatory strategy and enables a company to offer diverse
product variants (Calantone et al. 1999). However, response-based strategy requires
efﬁcient coordination and extensive information sharing between departments and ﬁrms
in the supply chain. In addition, shipment cost may increase substantially due to smaller
shipment quantities and ﬂexible responses, which if not well planned, may erode proﬁts.
The three studies underscore the importance of logistics in new product
introduction and make plausible propositions for further research. Speciﬁcally, the issue
of matching logistics and launch strategies under different market environments needs to

be tested empirically. Only one study, Closs et al. (1998), empirically compares

26

anticipatory and response-based supply chain strategies. The authors show that demand
uncertainty affects supply chain performance and that response-based strategy
outperforms anticipatory strategy in terms of customer service under conditions of
demand uncertainty. This means that anticipatory and response-based strategies result in
different performance outcomes under different demand situations, which supports
arguments by Bowersox et al. (1999) and Calantone et a1. (2005). Therefore, it is
important to examine not only the main effects of logistics strategy on new product
performance but also effects of interaction with other factors and moderating effect of

demand characteristics.

2.2.2 Launch Strategy

Broadly perceived, launch strategy refers to the set of decisions and activities
necessary to present a product to its target market and begin to generate income from
sales (Choffray and Lilien 1984; Green and Ryans 1990; Green et al. 1995; Hultink et al.
1998; Hultink and Robben 1999; Kotler 2003). Most studies tend to focus on a few
elements that collectively constitute launch strategy. Consequently, there is no consensus
on how to operationalize launch strategy concept (Hultink and Robben 1999). For
example, Green and Ryans (1990) examine four launch strategy variables, namely, timing
of entry, magnitude of marketing investment, magnitude of research and development
(R&D) investment, and product’s competitive advantage. Green et al. ( 1995)
operationalize launch strategy in terms of three decision components, namely, timing of

entry, magnitude of investment (i.e., R&D and promotion expenditures, etc.), and

27

competitive positioning (i.e., price, product quality, etc.), which affect long-term
performance of the product in the marketplace.

Hultink et al. (1998) distinguish between strategic and tactical launch decisions.
Based on a survey of over 600 managers of manufacturing ﬁrms in Netherlands, UK, and
USA, the authors posit that strategic launch decisions encompass elements of project’s
product strategy, market strategy, and ﬁrm strategy. These decisions, which address the
what, when, where, and why to launch questions, set the strategic context and competitive
scope into which the new product is launched (Biggadike 1979; Cooper 1993; Henard
and Szymanski 2001). Strategic decisions are more important, involve substantial
resource commitment and are difﬁcult to alter once decisions have been made (Debruyne
et al. 2002). Tactical decisions are the marketing mix decisions such as pricing,
branding, advertising and promotion, and distribution (these capture the how to launch
question). The authors conclude that managers tend to use a set of options across
different contexts, and that different strategies lead to different performance outcomes.

Hultink and Robben ( 1999) examine three launch decisions, namely,
strategic/market decisions, timing-related decisions, and marketing mix decisions. They
conclude that launch decisions that affected market acceptance of new consumer products
were relative product innovativeness, entry timing, and competitor reactions. They also
note that some launch decisions are more important in attaining new product success for
consumer products than for industrial products and vice versa, implying that launch
strategies will have different impact on new product performance depending on the

nature of product.

28

Calantone et al. (2005) conceptualize launch strategy in terms of launch size.
Depending on the size of market served, launch strategy is viewed as either fat (large
launch size) or lean (small launch size). According to this View, ﬁrms can either
distribute new product in select market segments only (lean launch) or in all market
segments (fat launch). The main premise is that ﬁrms will seek to size their launch
according to their resource capabilities and market needs. Thus, lean-fat launch
dichotomy can be viewed as a continuum where in one extreme, new product is launched
in one market segment while in the other, the product is launched in all market segments.

The type of launch strategy adopted is stipulated to directly impact performance.
Furthermore, since the scale of launch is directly related to logistics activities (i.e.
distribution centers and network, truck ﬂeet, etc.), it is argued that there is an interaction

between launch and logistics strategy, which will directly impact performance.

2.2.3 Environmental Factors

In this research, two environmental factors are examined: competitive response
and demand variability. Porter (1980) posits that competitors are motivated to react if a
competitor or product entry is viewed as threatening. Competitive response is a set of
decisions made by a ﬁrm in response to an observed new product launch (Hultink and
Langerak 2002). They are inﬂuenced by factors such as product innovativeness
(Gatignon and Xuereb 1997; Kuester et al. 1999; Debruyne et al. 2002), market structure
and organization position (Bowman and Gatignon 1995), market growth rates (Y con and
Lilien 1985), market signaling (Heil and Walters 1993), time required to develop a new

product and switching costs (Bowman and Gatignon 1995).

29

First, product characteristics such as product newness (innovativeness) triggers
retaliatory behavior from competitors especially in highly competitive markets
(Montoya-Weiss and Calantone 1994; Kuester et a1. 1999; Debruyne et al. 2002).
Debruyne et al. (2002) present a comprehensive review of literature on the impact of
product newness on competitor response and concludes that competitor response is
greater for radically new products than incrementally new products. Radically new
products are technologically superior and represent a discontinuity within an industry
while incremental innovations are logical extensions to existing knowledge and
established designs (Ali 1994; Gatignon and Xuereb 1997; Debruyne et al. 2002).
Radically new products tend to attract greater attention during launch and may have
signiﬁcant consequences in an industry as a whole. Since they are relatively new and are
not comparable to existing products, they tend to create uncertainty concerning the
success and consequences of innovation and uncertainty concerning the target market
(Debruyne et al. 2002). Robinson (1988) empirically showed that innovative products
are associated with greater competitor response since they threaten the incumbents.

Second, market growth rate affects reactive behavior of competing ﬁrms
(Bowman and Gatignon 1995; Kuester et a1 1999; Debruyne et al. 2002). Bowman and
Gatignon (1995) showed that competitor response is faster and more intense in fast
grth markets because in these markets, ﬁrms seek to establish a larger market share as
early as possible or aggressively defend their already acquired share. Moreover, as
Kuester et al. (1999) posit, high growth markets signify potential proﬁtability thereby
presenting the industry buyers with an incentive to invest and defend their market

position more vigorously. Robinson (1988) found that competitor response tends to be

30

more rather than less frequent in high growth market. The author argues that this may be
due to entrants who point out overlooked opportunities (i.e., the new product creates a
new market or causes an expansion of an existing market), incumbent sales that fall
below expected sales (i.e., as incumbents sales decline, they respond by attacking the new
product), and a longer time horizon which helps incumbents to justify the cost of reacting
(this is related to product life cycle). Hultink and Lingerak (2002) found that market
growth signiﬁcantly moderates the effect on the relationship between the entrants’ launch
decisions and the market signals that are perceived by the incumbents. Other studies
examining the impact of market growth on competitor response include Porter (1980),
Day and Wensley (1988), and Shankar (1999).

Third, the degree of market concentration is stipulated to inﬂuence competitor
response. For example, Porter (1980) posits that increased market concentration is
associated with increased competitor response since incumbents suffer market share loss
and reduced proﬁtability with entry of new ﬁrms. In a study of effects of market
characteristics and strategy on performance of 100 original and reformulated new
industrial products, Yoon and Lilien (1985) showed that the number of competitors in the
market place signiﬁcantly affects new product performance. This argument should hold
for new product entry relative to existing ones. Kuester et al. (1999) found that greater
market concentration is associated with less retaliatory from incumbents, and that larger
incumbents are less inclined to retaliate strongly. This ﬁnding contrasts Porter’s (1980)
arguments, and is consistent with other studies including Gatignon et al. (1990) and

Bowman and Gatignon (1995), who argue that ﬁrms in highly concentrated markets

31

focus more on proﬁt maximization and less on market share hence their slow retaliation.
Thus, there is no consensus on the effect of market concentration on competitor response.

Fourth, several studies have examined the inﬂuence of market signaling on
competitor response. According to Hultink and Langerak (2002), the scope, intensity,
and competitiveness of a new product launch generate a set of market signals that explain
competitive reactions from incumbents. In other words, attributes of new product launch
trigger different interpretations by an incumbent regarding motives and intentions behind
the new product entry (Heil and Walters 1993). These interpretations form the basis of
incumbents’ response strategy and actions. Heil and Walters (1993) examined three
market signals — hostility, commitment, and consequences — and concluded that hostility
and consequence signals signiﬁcantly trigger stronger competitive reactions. They deﬁne
signaling hostility as extent to which approaches used by a ﬁrm to introduce new product
is perceived as hostile while consequence signal refers to incumbents’ perception of the
impact of a new product entry on their proﬁtability. Kuester et al. (1999) combined
hostility and consequence signals into a new construct called ‘threat’ and found that
threat is a signiﬁcant predictor of reaction speed. They conclude that when competitors
are threatened based on a given product mix, they tend to retaliate with the same
instrument. Other studies such as Smith et al. (1989) show that actions perceived to be
threatening are associated with shorter response times. Thus, market signals by new
product entry will tend to invite response from incumbents.

Finally, competitive response tends to take longer when switching costs are high,
time required to develop new products is long, and when reacting ﬁrm’s market share is

high (Bowman and Gatignon 1995). Consistent with the ﬁndings in previous studies, it is

32

expected that competitor response will directly impact new product performance and also
moderate the impact of launch strategy and logistics strategy on performance.

The second environmental factor examined is demand variability. According to
Fisher (1997), product demand pattern affects the type of supply chain strategy adopted.
Demand for new products is highly unpredictable and can vary signiﬁcantly from the
average level. For example, ﬁrms introducing new to market (ﬁrst time) products tend to
rely on forecasts based on potential customer surveys and focus groups. These forecasts
are susceptible to high margin of error, which makes it very difﬁcult to plan before hand.
In cases where a product is an innovation of existing brand, the margin of error in the
forecast is likely to be lower since the demand for incremental product can be
benchmarked on the existing product demand. However, although demand uncertainty is
examined as a moderator factor in new product literature (Gatignon and Xuereb 1997),
there are no studies investigating its inﬂuence on the launch and logistics strategy-new
product performance linkage. Demand variability affects ﬁrms’ ability to fulﬁll customer
orders, which in turn impacts new product performance. Thus, it is a major consideration

when making launch decisions.

2.2.4 Product Characteristics

According to Henard and Szymanski (2001), product characteristics encompass
elements pertaining to the new product such as price, innovativeness, and manager’s
perceptions of how well the product meets customer needs. Past studies show that new
product attributes such as quality, reliability, newness, and uniqueness are important

indicators of a ﬁrm’s ability to meet customer needs (Li and Calantone 1998). These

33

attributes enable a new product to provide both tangible and intangible beneﬁts to its
customers that are superior to competitive brands (Li and Calantone 1998; Lee and
O’Connor 2003). To the extent that product attributes (i.e., innovativeness, reliability,
unique features, perceived value, etc.) offer advantages over competitor brands, then they
constitute product advantage (Lee and O’Connor 2003).

Extant literature shows that product advantage impacts new product performance
(Cooper and Kleinschimdt 1987; Li and Calantone 1998). For example, new product’s
innovativeness inﬂuences incumbent’s competitive response (Kuester et al. 1999;
Debruyne et al. 2002). Hultink and Robbin (1999) found that product innovativeness and
penetration pricing among other factors positively impacted market acceptance of a new
product. Yap and Souder (1994) posit that under low market uncertainty, new product’s
superior performance and unique features conveyed impression to the user that the
product will be premium priced, expensive to use, or failure prone due to perceived
complexity. In a nutshell, these studies underscore the inﬂuence of product advantage on
new product performance. However, the inﬂuence of product advantage on logistics and

launch strategies has not been examined.

2.2.5 Firm Resources

The resource-based view (RBV) of the ﬁrm holds that ﬁrm strategy is a product
of factors endogenous to the ﬁrm. It advances the argument that ﬁrms possess different
resources and capabilities, which enable them to generate proﬁt and build sustainable
competitive advantage (Barney 1991). To generate above normal rents, ﬁrms must

possess unique resources and capabilities that are rare, valuable, inimitable, and non-

34

substitutable (Wemefelt 1984, Barney 1991, Peteraf 1993). These resources must enable
the ﬁrm to conceive of or implement strategies that improve performance, exploit market
opportunities, or neutralize impending threats (Barney 1991). Firm resources include
tangible assets (i.e., truck ﬂeet, distribution centers, etc.) and intangible assets (i.e., brand
names, trade marks, etc.) (Olavarietta and Ellinger 1997; Fahy 2001). Capabilities are
complex bundles of skills and knowledge exercised through organizational processes,
which enable ﬁrms to coordinate their activities and make use of their resources (Day
1994). Although possessing key resources is critical, it is the manner in which resources
are deployed that enables realization of above normal returns (Penrose 1959).

This implies that when planning new product launch, ﬁrms take cognizance of
their resource endowment and adopt the strategy that enables optimal resource allocation
and rent appropriation. Generally, large size ﬁrms tend to have greater resources that
translate to greater market power, which they can use to gain competitive advantage
(Gatignon and Xuereb 1997). However, ﬁrms must also have the capability to effectively
deploy these resources to improve new product performance. For example, a ﬁrm with
small launch budget, limited distribution centers, and few trucks may opt to launch its
new product in a few market segments whereas a ﬁrm with large launch budget and
widespread distribution network may launch new product in multiple market segments
simultaneously. Thus, it is expected that a ﬁrm’s resources and capabilities substantially

determine its launch and logistics strategies.

2.3 Models and Techniques for Research Analysis

In this research, case study research approach and simulation modeling are used.

35

2.3.1 Case Study Research

Case study research approach has wide application in many ﬁelds including
business and social sciences (Gummesson 2000; Yin 2003). Yin (2003) posits that three
conditions determine the type of research approach (i.e., case study, survey, archival
analysis, etc.), viz., (i) the type of research question posed, (ii) the extent of control an
investigator has over actual behavioral events, and (iii) the degree of focus on
contemporary as opposed to historical events. In this research, case studies are done of
two companies that deal with durable and non-durable consumer goods. When
developing a simulation design, it is desirable to have knowledge of the real system
processes so that the simulation mimics the real process. Thus, the use of case study
research is an important initial step in carrying out a simulation research. Since the main
purpose of case studies is to provide sound bases for simulation design, it is important
that the case study selection reﬂect the issue under investigation. The researcher
determines whether a single case or multiple cases are to be used (Yin 2003). A single
case is suitable when the case represents a critical case to test a well-formulated theory or
to reveal a previously inaccessible phenomenon while multiple cases are desirable for
generalizability (Ellram 1996).

Case study approach allows use of different information sources (i.e., interviews,
expert opinion, etc.), and is preferred when direct observation of events being studied and
interviews of persons involved can be done (Yin 2003). According to Gummesson
(2000), case studies can be concerned with different functions, relationships, events or
processes and may involve study of historical, present, future conditions or mixture of

them all. Ellram (1996) provides a detailed analysis of case study application in logistics

36

research. The author states that case study approach can signiﬁcantly advance logistics
research. Consequently, case study approach is appropriate in this research because it

provides basis for simulation design and is used in past logistics research.

2.3.2 Simulation Modeling Review

Simulation modeling technique is widely applied to many problems and systems
that are difﬁcult to feasibly construct or experiment upon in ‘real world’ or that are too
complex to model analytically (Rees et al. 2002). Computer simulation allows different
factors to be modeled into the system. Simulation consists of systems entities, input
variables, performance variables, and functional relationships (Maria 1997). The major
steps in a simulation modeling are problem identiﬁcation, collecting real system
information and data, formulate and develop a model, design and analyze simulation
experiment, validate and verify model, perform simulation runs, and analyze results (Law
and Kelton 2000; Law et al. 1994; Maria 1997). The objective of simulation in this
research is to model the dynamic behavior of launch processes and to evaluate the
consequences of alternative conﬁgurations to new product performance.

According to Bowersox and Closs (1989), the model categories used in logistics
are: analytic, heuristic, and simulation. Analytic models use mathematical methods to
identify the optimal solution to the problem under analysis. Heuristic models use “rule of
thumb” procedures developed from a basic knowledge of the problem and are capable of
replicating complex problems. Both analytic and heuristic models are deterministic,
implying that they do not contain random variables. Simulation models are distinct

because of their capability to include stochastic situations. Since logistics planning is

37

characterized by uncertainties and resulting variance, simulation models are best suited
for the analysis, which has led to their extensive use. As Marucheck and McClelland
( 1992) point out, simulation models can accommodate dynamic interrelationships and
their assumptions are less restrictive. Consequently, simulation modeling is suitable in
this research because it enables a more detailed assessment of interrelationships between
launch and logistics strategies in different competitive and market environments.

Simulation approach has been used in many logistics studies. For example,
Rodrigues (2004) reviewed several journals for the period 1980 to 2003 to identify
articles that primarily use simulation analysis. The author found out that for the eight
journals reviewed, there were a total of 115 articles that employed simulation analysis.
The journals reviewed were the European Journal of Operational Research, Journal of
Business Logistics, Decision Sciences, Journal of Operations Research, International
Journal of Physical Distribution and Logistics Management, International Journal of
Operations and Production Management, Management Science, and International
Journal of Logistics Management. In all these studies, there is limited focus on new
product launch analysis.

A review of Journal of Business Logistics articles for the period 1987 — April
2005 shows that a total of 22 articles were based on simulation analysis. None of these
articles explicitly examined new product launch issues or logistics aspects relating to new
product introductions. International Journal of Logistics Management had two
simulation based articles for the period 1998 to April — 2005. One of the articles, Closs
et al. (1998) uses simulation analysis to examine supply chain performance under varying

information exchange and demand uncertainty conditions. International Journal of

38

Physical Distribution and Logistics Management had 27 simulation based articles for the
period 1979 to April 2005. None of these articles examined product introduction related
issues, which suggests that application of simulation approach in new product launch
research is limited. Thus, this research makes an important contribution to the growing

application of simulation modeling in logistics and new product research.

2.4 Conceptual Model and Research Hypotheses
Figure 2.1 shows the conceptual framework used in this research.

Figure 2.1: Conceptual Framework

 

 

“c-0-u-u-n—c-¢-e-o---u-o-o-o-u-

Logistics strategy 3 Environmental factors
- Anticipatory i - Demand Uncertainty
- Response-based : - Competitive Response

 

 

 

 

 

 

 

 

 

 

 

 

H19“) . .......................... .

Launch Strategy 5
- National Launch 5 NP Performance
- Rollout Launch : - Order Fill Rate

: ‘ - Case Fill Rate

I - System Inventory
Logistics x Launch Strategy ' - Proﬁtability
- Anticipatory x National

 

 

- Anticipatory x Rollout
- Response-based x National
- Response-based x Rollout

    

 

 

 

H7(+)

 

 

Product Characteristics Firm Resources
- Product Advantage - Production Capacity

 

 

 

 

 

 

 

 

 

According to the model, six determinants of new product performance are
examined. These factors are based on extant literature as discussed in preceding sections.

Their relationship to performance is stipulated in seven study hypotheses. Besides main

39

effects, interaction effects are also examined. Of special interest in this research is
interaction between logistics and launch strategy.

Past studies suggest that logistics systems and capabilities impact new product
success (Bowersox et al. 1999; Di Benedetto 1999; Guiltinan 1999; Calantone et al.
2005). Logistic strategy is conceptualized as either anticipatory or response-based. As
the preceding literature review suggests, each of these strategies result in different new
product performance outcomes. For example, anticipatory strategy involves greater
inventory deployment in the distribution system therefore it is associated with higher
average system inventory, higher inventory carrying costs, and opportunity costs.
However, with higher inventory availability, there will be fewer stockout and higher ﬁll
rates. Response-based strategy reduces average system inventory and attendant inventory
carrying costs, which may increase proﬁtability compared to an anticipatory strategy.
However, order processing and delivery costs are higher due to frequent orders and
shipments with limited economies of scale. Consequently, it is hypothesized that:

H 1 a: Logistics strategy directly and positively impacts new product performance.

H1 b: Response-based logistics strategy has a greater positive impact on new
product performance than anticipatory logistics strategy.

Launch strategy is stipulated to directly impact new product performance. As
Green and Ryan (1990) posit, entry (launch) strategy for a new product provides the
platform from which sustainable competitive advantage is gained. Past studies have
shown that launch strategy type may explain new product performance differentials
(Montoya-Weiss and Calantone 1994; Green et a1. 1995; Hultink and Robben 1999;
Henard and Szymanski 2001). Launch strategy is conceptualized as either national or

rollout. The preceding literature review suggests that a national launch strategy requires

40

greater resource commitment, higher inventory, and higher risks a than rollout launch
strategy (Calantone et al. 2005). Thus, national launch strategy is associated with greater
average system inventory and higher inventory carrying cost as well as greater
management complexity given greater market coverage. Rollout launch strategy on the
other hand is associated with less average system inventory, lower inventory cost, greater
ﬂexibility, and less management challenges given less market coverage. Consequently, it
can be argued that rollout launch strategy results in better performance than a national
launch strategy. Thus, it is hypothesized that:
H2a: Launch strategy directly and positively impacts new product performance.

H2b: Rollout launch strategy has a greater positive impact on new product
performance than national launch strategy.

The interaction between different logistics and launch strategies is stipulated to
yield different performance outcomes. For example, national launch supported by an
anticipatory logistics strategy is associated with the greatest average system inventory,
inventory carrying costs, and risks of write-off in case product fails but has least
opportunity loss if product is successful. A rollout launch supported by response-based
logistics strategy has the lowest average system, inventory carrying costs, and risks in
case product fails but has the highest opportunity loss in case product is highly
successful. Results of other strategy interactions fall between the above extremes.
Therefore, each of the four combinations of launch and logistics strategies directly and
differentially affects new product performance. Therefore, it is hypothesized that:

H3: The interaction between logistics strategy and launch strategy has a direct
and positive impact on new product performance.

Demand variability and competitive response impact new product performance

(Bowman and Gatignon 1995; Kuester et al. 1999). Demand variability is expected to

41

inﬂuence distribution plans and pose signiﬁcant challenge to inventory management,
which may profoundly impact performance. At the same time, intense competitive
response to a new product introduction signiﬁcantly erodes proﬁts and market advantage
that a ﬁrm gains from its new product sales. However, when competitive response is less
intense, a ﬁrm is able to build competitive advantage that translates into increased sales,
market share, proﬁts, and customer loyalty. Thus, it is stipulated that demand variability
and competitive response directly impact performance and moderate the impact of
logistics and launch strategies on performance. Thus, it is hypothesized that:
H4a: High demand variability negatively impacts new product performance.
H4b: High competitive response negatively impacts new product performance.
H5a: The greater the demand variability for a new product, the less the impact of?
(i) Logistics strategy on new product performance.
(ii) Launch strategy on new product performance.
(iii) Logistics and launch strategy interaction on new product performance.
H5b: The greater the competitor response to a new product introduction, the less
the impact of'
(i) Logistics strategy on new product performance.
(ii) Launch strategy on new product performance.
(iii) Logistics and launch strategy interaction on new product performance.
Extant literature shows that product advantage is a major determinant of new
product performance (Cooper and Kleinschimdt 1987; Yap and Souder 1994; Hultink and
Robbin 1999; Kuester et al. 1999; Debruyne et al. 2002; Li and Calantone 1998). As the
preceding literature review indicates, product advantage is based on product attributes
such as quality, unique features, image, etc. These advantages provide the bases for
improved performance through greater customer acceptance and subsequent sales.

However, they also tend to trigger competitive retaliation. Therefore, product advantage

results in greater demand, which may affect a ﬁrm’s ability to fulﬁll customer orders.

42

Moreover, increased demand leads to more sales and proﬁts. In high product advantage
situations, ﬁrms may maintain higher inventory in distribution system in anticipation of
increased sales, which affects inventory carrying cost. Consequently, product advantage
is expected to impact new product performance. Therefore, it is hypothesized that:

H6: New product advantage directly and positively impacts new product
performance.

Several studies within the resource-based view of the ﬁrm framework show that
ﬁrm resources inﬂuence ﬁrm strategy formulation, implementation, and performance
(Barney 1991, Peteraf 2001; Amit and Schoemaker 1993). When launching a new
product, it is expected that ﬁrms’ resource endowment will dictate the type of launch
strategy adopted. This is because resources may enhance or constrain the extent to which
ﬁrms execute a speciﬁc strategy. For example, where resources such as ﬁnancial budget,
distribution facilities, and production capacity are limited, a ﬁrm may ﬁnd it difﬁcult to
implement a national launch. Conversely, where similar resources are in abundance, the
ﬁrm can execute a national launch more efﬁciently and effectively. Large ﬁrms will tend
to have more resources than smaller ﬁrms, which mean they can implement strategies
that smaller ﬁrms cannot. Consequently, a ﬁnn’s production capacity can be used as a
surrogate of ﬁrm resources. Therefore, it is hypothesized that:

H7: Firms' production capacity directly and positively impacts new product

performance.
2.5 Conclusion

This Chapter highlights several research gaps in new product launch literature and

presents a conceptual framework for this research. First, literature review shows that the

impact of logistics strategy during new product introductions is not adequately addressed.

43

Most studies addressing the issue have tended to be conceptual in nature and their
propositions are not empirically tested. Secondly, past studies have tended to focus on
main effects of individual performance determinants while ignoring interaction effects.
Thirdly, moderating effects of competitive response and demand uncertainty on the
impact of launch and logistics strategies on new product performance has not been
addressed. Fourthly, the impact of product advantage and ﬁrm resources on new product
introduction strategies and performance has not been adequately addressed. Finally,
simulation modeling approach is increasingly being used in logistics research. However,
a search in major logistics journals indicates that it is rarely applied in new product
literature. Simulation offers an effective way to analyze and compare interaction effects
between different determinants of new product performance. Consequently, this study
seeks to ﬁll these gaps in literature by using case study and simulation modeling to
examine the simultaneous impact of launch and logistics strategies, environmental

factors, product characteristics, and ﬁrm resources on new product performance.

44

CHAPTER III: RESEARCH METHOD
3.0 Introduction
This research employs two methods: case study approach and simulation

modeling. This Chapter discusses how these approaches were applied.

3.1 Case Study Research

Data from case studies of select companies is used to build a simulation model of
new product launch process. Two companies that deal with durable consumer products
and one that deals with non-durable consumer products are studied with a view to
identifying how they conduct their new product launch. Findings from these case studies
provide the basis for development of simulated launch process and model parameters.

Case study approach is preferred for several reasons. First, it is necessary to
understand how launch process is conducted, why it is conducted in the observed manner,
and the implications of the launch process design on the overall performance. Yin (2003)
notes that case study approach is appropriate for research problems where the emphasis is
on “how” and ‘ﬁrvhy” questions. Given limited empirical studies examining the role of
logistics in new product launch processes, there is lack of essential data on the how and
why of launch process. Consequently, case studies were necessary in this research.

Secondly, most studies advocate use of case study approach to study
organizational processes that do not lend themselves easily to quantitative measures
because it enables in depth assessment of process dynamics not possible via other
research methods (Eisenhardt 1989; Marginson 2002; Yin 2003). New product launch

process brings together supply chain partners such as manufacturers, suppliers,

45

distributors, retailers, and transporters. To understand the interactions between these
players during new product introduction, it was necessary to identify and study speciﬁc
companies that have a long history of successful new product launches. Case studies
enabled better understanding of the launch process dynamics, which in turn enabled the
development of a realistic, valid, and reliable model.

Finally, case studies enable a better appreciation of dynamics at play during new
product launch, which aids in interpretation of results and formulation of theoretical and
managerial recommendations. Furthermore, this being a ﬁrst study, case studies provide
insights that are basis for future research propositions.

Case studies proceed in three main steps: case study design, data collection, and

data analysis (Eisenhardt 1989; Miles and Huberman 1994; Ellram 1996; Yin 2003).

3.1.1 Case Study Design

Case study design involves development of the questions to be asked and
identiﬁcation of criteria for judging quality of research design. In this study, research
questions were open-ended. These types of questions enable respondents to give a
detailed narrative of the events and sequences that are then reconstructed to develop a
typical launch process ﬂow. Narratives are analytic constructs used to unify groups of
events into single stories (Abbot 1990; Grifﬁn 1993; Riessman 1993; Stevenson and
Greenberg 1998). They embody sequence and time, and are therefore ideally suited to
the development of process theories and explanations (Pentland 1999). Participants not
only make sense of their world in narrative terms but they proactively plan and enact

narratives that are consistent with their expectations and values. Consequently, narratives

46

can be invaluable sources of insights about organizational processes since explanations
that draw upon narrative data tend to be close to the phenomena they purport to explain.
As Pentland (1999) states, narratives are like ruts in the road that people follow and
thereby re-create.

Abbott (1990) suggests three categories of questions that one can address in a
narrative context: the existence and classiﬁcation of sequential patterns, the antecedents
of, and the consequences of these patterns. Riessman (1993255) suggests use of ﬁve to
seven broad questions about the topic of inquiry (i.e., “tell me what happened”) with
supplementary questions to probe further (i.e., “can you tell me more about that”). In this
research, interview questions seek general and speciﬁc information new product launch
information in each company (Refer to Appendix A for Sample Questions).

The ﬁrst set of questions examines company’s new product launch process in
general. For example, the respondents were asked to describe how their company plans
and executes new product introduction. This is important to the researcher because it
provided a general overview of launch process in each company and the context in which
speciﬁc product launches are conducted. The second set of questions addresses speciﬁc
product launch activities. Interview respondents are asked to discuss a recent product
launch in their company with respect to product characteristics, launch process, logistics
processes during the launch, prevailing market characteristics, and performance outcome.
Responses to these questions provided the bases for the design of a simulation model that
replicates a real product launch process as closely as possible. Pattern matching approach
(Eisenhardt 1995; Yin 2003) is used to weave together the processes and events described

in interviews into a coherent launch process ﬂow.

47

3.1.2 Evaluating Quality of Research Design

There are four major criteria of ensuring reliability and validity in case study
research: construct validity, internal validity, external validity, and reliability (Eisenhardt
1989; Miles and Huberman 1994; Yin 2003; Hefferrnan 2004).

Construct validity establishes correct operational measures for concepts being
studied. According to Yin (2003), construct validity can be achieved in three steps: use
multiple sources of evidence, establish chain of evidence, and have key informants
review draft case study report. To ensure construct validity, at least three executives
involved in new product development and/or launch in each company are interviewed.
The respondents provide detailed information about their company’s launch process and
the different activities that take place from the time a ﬁnished product is ready for market
to the time it reaches consumers. These interviews validated assumptions made earlier
based on information available in extant literature such as transportation lead times,
inventory policies, and supply chain network.

The interviews enabled the researcher to develop a launch process ﬂow showing
important linkages and events in the supply chain network. The mapped process ﬂow is
considered representative of a typical new product launch in the consumer goods industry
as discussed in the interviews. As a ﬁnal step to achieve construct validity, interview
respondents were asked to review the mapped process ﬂow and major assumptions in the
model so as to ensure that the process ﬂow is a true representation of the launch process
in their companies and industry. The respondents concurred that the mapped process is

fairly representative of a typical launch process. Consequently, construct validity is met.

48

Internal validity requires the researcher to establish causal relationships, whereby
certain conditions are shown to lead to other conditions. Researchers can do pattern
matching, explanation building, address rival explanations, and/or use logic models (Yin
2003). In this research, pattern matching is done. Launch activities and processes
discussed in the interviews and those documented in extant literature are matched to
develop reasonable launch process linkages and ﬂows. During interviews, respondents
explained the rationale behind most launch process linkages. Therefore, mapped process
ﬂows indicate the sequence of events during a typical product launch. To the extent that
the mapped launch process ﬂows and linkages represented correct causal relationships,
internal validity is met.

External validity establishes the domain to which a study’s ﬁndings can be
generalized (Yin 2003). As Yin (2003) posits, case studies rely on analytical
generalizations unlike survey research, which rely on statistical generalizations. In this
research, use of three case studies increases replication logic thereby increasing analytical
generalizations of the ﬁndings. The mapped launch process incorporates information
from all case studies and case studies ﬁndings are compared with relevant information in
extant literature. The comparison shows that mapped process ﬂows are representative of
ﬂows in consumer goods industry. Therefore, external validity is achieved.

Finally, reliability requires the researcher to demonstrate that operations of a
study such as data collection procedures can be repeated with same results (Eisenhardt
1989; Yin 2003; Hefferman 2004). In this research, three strategies are used to ensure
reliability: case study protocol, case study database, and review of interview transcripts to

ensure accuracy. A case study protocol contains survey instruments, procedures and

49

general rules to be followed when using the instrument (Yin 2003). According to Yin
(2003: 69), a typical protocol should have an overview of case study project (a statement
about the project, its objectives, people involved, issues being investigated, etc.), ﬁeld
procedures (sources of information), case study questions (speciﬁc questions to guide the
data collection), and a guide of case study report (outline, format of the narratives, use
and presentation of other documentation). In this research, a case study protocol was
developed and followed to the letter during interviews. Before interview appointments,
respondents were provided with survey questions and a one-page overview of research
objectives, research design, interview procedure, and expected contribution (See
APPENDD( A, B, C, and D).

During interviews, key issues discussed are written down. The hand-written
interview notes are then transcribed into a word document and stored for future reference.
The transcribed notes are double-checked to ensure that all information is noted and

accurately recorded. Therefore, reliability is achieved.

3.1.3 Data Collection

Data is collected through face-to-face and telephone interviews with managers in
three companies. The interview respondents are provided with interview questions
before hand. The interviews take about one and half hours. In each case study, interview
respondents are managers involved in logistics and new product processes. For example,
respondents in one case study include logistics manager and program manager for supply
chain in new product area while in another case study, respondents include heads of

operations, new brands and marketing in new product area. These respondents are highly

50

knowledgeable in new product introduction processes. They share insightful information

not only about their companies but also about industry trends.

3.1.4 Data Analysis

Case study data analysis involves transcription of interview notes and generation
of an analytical diagram of causal relationships and events as well as interpretation of
sequences of action that are generalizable across multiple new products launch scenarios.
The analytical diagram represents the researcher’s interpretation of relationships and
sequence of events constituting new product launch process. A within-case and cross-
case analysis is done to determine reliability and validity of the results. The case studies

ﬁndings are presented below.

3.1.5 CASE STUDY I

This case study involved a company that deals with non-durable consumer goods.
The company is functionally divided into departments such as marketing, operations,
ﬁnance, etc. New product development, testing, and commercialization are handled by
the marketing department. In this company, new product launch process starts with
research (qualitative and quantitative) to determine potential market, requirements to get
new product to market in a timely manner, asset requirements for rollout, and production
capacity requirements. Engineering division assesses production requirements and
determines which manufacturing facilities will be used to make new product and when to

produce. Sales division assesses distribution base. Sales teams make new product

51

(sample) presentations to select distributors and retailers. Thereafter, a distribution kick-
off team makes presentations throughout the country.

The ﬁrst production run is based on forecasts but subsequent productions are
based on distributor orders (with some cushion inventory added to these orders). This
company uses a three-tier distribution system in which the company only supplies to
distributors, who then supply the retailers. Generally, orders are shipped weekly or in
one to four weeks. The company has limited storage capacity at its manufacturing plants.
If extra capacity is needed, the company uses its distributors’ capacity. There are more
than 500 distributors in the US. The company carries 3-4 weeks worth of inventory.

Recently, the company launched a product that was quite new to the market place.
Although market for the product category is at mature stage, the company introduced
features that were new to the market. The target market was women consumers between
ages 21 to 45 years. The product was tested in three market clusters selected based on
market issues such as target customer proﬁles and distribution requirements. Test
marketing is done to minimize risks of launch, assess pricing, and to understand market
characteristics and consumption behavior. Product performance in these markets would
determine the extent of national introduction. Production was done in the Southeast
(USA) while one test market was in the Northwest. Thus, transporting the product to this
test market was expensive and challenging. For this company, the average shipping
distance to distribution facilities is 315 miles. The average lead time is 2 days.
According to respondents, high cost of shipping could be minimized if new product

production was done in multiple facilities throughout the country.

52

Generally, competition is stiff at both national (with national brands) and local
(with local brands) level. Competition is based on price, packaging, product innovation,
brand equity, etc. The industry is heavily regulated. Consequently, price competition is
less intense. Companies compete on brand marketing and promotions. The top three
brands account for 50% of the US. market. Thus, the industry is characterized by high
concentration. According to interview respondents, market trends for major brands in the
industry do not change abruptly. Therefore, customer demand is relatively predictable
and forecasts are fairly dependable. How, this may not be the case for new products.

The company has multiple measures to determine new product success/failure.
These include product sales (did the sales meet expectations), brand awareness, customer
loyalty to product/brand, feedback from distributors, cost of launch, distribution cost,
distribution effectiveness and velocity, etc. The aforementioned product was poorly
received in the market place. After three to six months of test marketing, it was clear that
the product was bound to fail as sales and distributor orders declined. Most raw materials
used for the product were new and had been ordered in advance based on sales forecasts.
As a result, the company had excess raw material and ﬁnished goods inventory.

In another case, the company successﬁrlly launched a new product nationally
(nation-wide launch). This product was produced simultaneously in six facilities located
in different parts of the country. The new product sales and market share increased
substantially within six to eight weeks. According to respondents, nation-wide launch is
a risky undertaking. Thus, it is crucial for the company to have sufﬁcient ﬁnancial

resources before undertaking a nation-wide launch.

53

3.1.6 CASE STUDY II

This case study involved a company that manufactures durable consumer goods
(ofﬁce equipment/furniture). The company operates a make-to-order production model
where products are made only upon receipt of customers’ orders. New product launch
process starts with determination of how and where the product will be manufactured.
The company can make new product in three ways: product can be made entirely in the
ﬁrrn’s facilities (i.e., from raw material to ﬁnished goods), product parts can be
purchased from suppliers then assembled in the ﬁrms, or production of new product
could be outsourced to a contract manufacturer. Next, a product placement analysis and
pilot testing is done. This is primarily done in exhibitions at NeoCon, the World’s Trade
Fair for Interior Design and Facilities Management. Also, product demonstrations are
done at dealers’ facilities. Once the decision to launch a new product is made, initial
production is done based on ﬁrst order entries from dealers.

In the recent past, the company introduced a product that had many innovative
features and was marketed as new, innovative, comfortable, and easy to use. It was very
successful. Product demand grew rapidly outstripping production capacity, which led to
delays in order fulﬁllment. Therefore, the company had to expand its production
capabilities. Initial four-week’s production was for dealers to display and demonstrate to
potential customers. Most components and sub-assemblies for the new product were
bought from suppliers. The product was launched nationally including in Canada and
Central America. Subsequent launches were done in Europe and Asia. According to

interview respondents, market placement analysis necessitated the nation-wide launch.

54

The company handles 10 percent of transportation while the rest is outsourced to
third party providers and distributors. Most shipments are less-than-truckload. Typical
lead time from manufacturing to ready to ship is 1 day and actual delivery takes 2 days.
The distribution network comprises manufacturer, distributors/dealers, and retailers. The
company ships product to distributors/dealers and then dealers supply to retailers.

Competition is generally intense but demand is fairly predictable. Competitors
generally react to new product introduction by offering similar products. New product
success is measured in terms of meeting new product revenue targets, proﬁtability, and
customer satisfaction among others. The aforementioned new product launch was

successﬁrl based on high sales, proﬁts, and customer satisfaction ratings.

3.1.7 CASE STUDY III

This case study involved a large company that deals with durable consumer
products. The company has operations or sells products in 30 countries. The company is
organized into teams around product categories. New product launch is planned and
executed by these product category teams. There are functional units in each area such as
marketing, industrial design, manufacturing, purchasing, supply chain, etc. In each
product category, there is a process-matrix of activities to be performed by each
ﬁrnctional area before and during launch. For example, there are activities and decisions
dealing with new product development, product strategy, concept selection, product
engineering, production decisions, etc. The marketing team communicates customer
requirements and develops retail labels. Production of the new product is forecast driven.

Demand forecast are for 12-months with a monthly review.

55

In the recent past, the company launched a new product that was highly successﬁrl
and another that did not do as well. The successful product was highly innovative and
required new product design and tooling. Therefore, the company had to displace old
products at its manufacturing facility to accommodate new product needs. Initial sales
target was 150,000 units. Within six months of launch, sales had doubled, thereby
outstripping production capacity. By the end of ﬁrst year since launch, sales exceeded
400,000 units. According to respondents, planning for the new product was conservative,
a practice that is common when launching new products. Production of the unsuccessful
product was done in Asia. Due to low sales, the company had to cut down production.
In addition, the company bought back stock from its distributors so as to create room for
successful products. Usually, retailers own the product stock in their channel.

The company typically launches new product nationally. However, there are
cases where new product is launched by region, with delays of about 6-8 weeks in
accordance with production capacity and retail channel requirements. Generally, the
company plans with customers (distributors and retailers) on the quantity and type of
product to ship where and when. For example, a major retailer with 1,800 stores (such
as Home Depot or Lowe’s) may not expect the manufacturing ﬁrm to instantly ship new
product to all stores. Instead, it may request the manufacturer to have new product
shipped to speciﬁc stores in a sequential manner, which helps rationalize production.

The company has manufacturing facilities across North America. Each facility
ships to the company’s only distribution centers in Wisconsin and to major customers
(national wholesalers such as Ferguson, regional wholesalers, national retailers such as

Home Depot, and regional retailers). However, the company is in the process of

56

changing this model to one where they have regional distribution centers focused on
fewer markets so as to reduce total haul length and duration. The company meets the
cost of freight to distributors and major customers. The average haul length is 1100
miles. The average order cycle is 14 days. However, there are contractual obligations
with some customers to deliver the product in shorter leads times. For example, Home
Depot requires deliveries to be done in 7-10 days while Lowe’s requires deliveries in 7
days. Total inventory in the system (inventory at the production, storage, and transit
systems) ranges from 20 to 40 days.

Competition in the industry is generally stiff. The company is facing competition
from local competitors as well as foreign brands especially from Asia. Usually, it takes
about a year for competitors to create or copy newly launched product designs. Although
demand is fairly stable, it is difﬁcult to plan for or understand demand for a new product.
To measure new product launch success, the company uses indicators such as sales,
proﬁts, and a Vitality Index of 30 percent. This index measures the target to net sales for

a new product.

3.1.8 Key Findings from Case Studies

I Production Capacity —- Inﬂuences timing and scale of new product introduction.
Companies may scale down the number of markets where new product is launched if
capacity limitations so demand.

I Logistics Strategy and Inventory Policies — Companies launching new products often
adopt either “push” or “pull” logistics strategy. For Push strategy, ﬁrms manufacture

new products based on demand forecasts and then ‘push’ the ﬁnished products to

57

distributors and retailers to await sales. In effect, ﬁrms deploy inventory to
distributors and retailers while maintaining minimal stock at the manufacturing
facility. For Pull strategy, ﬁrms withhold shipping new products to distributors and
retailers until they receive orders so that inventory is held at the companies’ plant
warehouses or distribution centers. At the same time, ﬁrms can adopt either “lean” or
“fat” inventory policy. In a Lean policy, ﬁrms maintain inventory required to cover
one order cycle time while in a Fat policy, ﬁrms maintain higher inventory (i.e., buffer
inventory) to cover more order cycle times. Fat policy results in greater average
system inventory than a Lean policy. The aforementioned Push strategy and Fat policy
are characterized by higher inventory in the distribution system and more inventory at
the retail level, which reﬂects anticipatory strategy. On the other hand, Pull strategy
and Lean policy are characterized by lower inventory in the distribution system and
less inventory at the retail level, which reﬂects a response-based strategy. Therefore,
logistics strategy can be conceptualized as two alternative strategies: Push vs. Pull,
and Fat vs. Lean. Transportation lead times from manufacturers to distributors and
major retailers range from 2 to 5 days. The main mode of transport is road.

Launch Strategy — when launching new product, companies adopt either a national
strategy (simultaneous nation-wide launch) or a rollout strategy (launch in some
markets then to others sequentially until all markets are covered over time).

Demand Variability — New product demand is highly unpredictable. Generally, new
product demand varies signiﬁcantly from the daily average. This makes planning for

and executing new product launch strategies very challenging.

58

' Competitive Response — Competitors respond to new product launch in many ways,
which affects new product’s sales and market share. Aggressive competitor response
lowers demand for a new product, and if not countered, could lead to product failure.

. Product Advantage — Demand for products that are perceived to have many advantages
by the customers tend to grow rapidly, in some cases by up to four times during the

year the product is launched (for consumer goods).

3.2 Simulation Modeling

Supply chain analysis presents a challenge because of system complexity and the
large number of variables involved. Thus, analogous models are built to examine
speciﬁc system variables in a controllable environment. Computer simulation modeling
presents a practical basis for representing complex systems and helps to realistically
analyze performance trade-offs (Venkateswaran 2004). Computer simulation refers to
methods of studying a wide variety of models of real world systems by numerical
evaluation using software designed to imitate the system’s operations or characteristics
(Kelton et al. 2003). This research uses ARENA, a dynamic simulation software tool that

combines SIMAN simulation language with a graphics component (Kelton et al. 2003).

3.2.] Simulation Model
The conceptual supply chain network studied is dynamic multi-echelon,
comprising production functions, distribution functions, customer (market) segments, and

information ﬂows. Figure 3.1 shows the simulated supply chain network.

59

Figure 3.1: Supply Chain Network

 

 

 

,_p Distributor 1

 

 

 

 

Manufacturer

 

 

 

 

,_---_________L_--___--_,

 

 

_; Distributor 2

 

 

 

—’ Products
<— - - - Orders

 

 

 

 

In this network, Manufacturer serves Distributor 1 and Distributor 2. Distributor 1
serves Retailers 1, 2, 3, 4 and 5 exclusively while Distributor 2 serves Retailers 6, 7, 8, 9
and 10 exclusively. Retailers exclusively serve their respective customer segments, i.e.,
Retailer 1 serves Customer Segment 1 only. Transshipments between distributors or
between retailers are not allowed. Demand in each retail location is independent and may
vary at each Customer Segment throughout the simulation period.

According to the model, customers place orders (buy) at the retail facility.
Retailers acquire inventory supplies from distribution centers, which in turn replenish
with supplies from manufacturer. Customer demand and transit lead times follow
stochastic distributions. The average daily customer demand was 100 units. Demand

was generated using Triangular Distribution in excel. For example, low demand

60

 

variability was generated with TRIA(75,100,125) while high demand variability was
generated with TRIA (0,100,200). In ARENA, TRIA(75,100,125) results in a daily
demand with mode of 100, minimum daily demand of 75, and a maximum daily demand
of 125. This represents a truncated Normal Distribution. Stochastic lead times were
imposed to represent order processing and transportation delays. The transit time from
manufacturer to distribution centers is TRIA(2,3,4), which represents a transit time mode
of 3 days, minimum of 2 days, and a maximum of 4 days. Transit time from distribution
centers to the retailers was TRIA (1,2,3), which represents a mode of 2 days, minimum of
1 day, and a maximum of 3 days. Target and maximum levels of inventory in each
location are deﬁned based on logistics strategy tested. Safety stock levels are based on
forecasted average daily demand and target levels. On a typical simulated day, a
sequence of events take place as summarized in Figure 3.2 below.

At retail level, retailers receive replenishment shipments that were scheduled to
arrive (in-transit inventory) at the beginning of the day. These shipments may include
backorders from past unfulﬁlled demand. If backorders are allowed, then all demand is
eventually met. However, in this simulation, backorders were not allowed. This is
because backorders would add complexity to inventory deployment and customer service
thereby making it difﬁcult to identify speciﬁc strategy implication on performance. Any
unfulﬁlled orders were disposed and recorded as stock outs. Retailers update their
inventory positions after receiving replenishments shipments. After that, customer orders
are received at retail locations. If there is sufﬁcient inventory, orders are ﬁlled and

shipped. If there is no sufﬁcient inventory, partial orders are shipped. The order that

61

remains unfulﬁlled is considered a backorder. Since backorders are not allowed in this
simulation, stock cuts are calculated whenever orders are not ﬁlled.

Figure 3.2: List of Events

 

 

 

 

Retail: Receive Replenishment Retail: Receive Retail: Ship
Shipments & Backorders Customer Orders Customer Orders 7|

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Retail: Create Distributor: Receive
backorders replenishment orders
& backorders

 

 

 

 

 

 

 

 

 

 
  

 

 

   

Distributor: Ship Change Yes Distributor: Create

 
 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

Replenishment Orders replenishment Replenishment
& Backorders . Ian? Orders
No
,- ' - ' r - r r ' - - -. Manufacturer: Record Statistics
I Forecasts I _______ Production Planning

 

 

 

 

 

 

 

 

 

Retailers then evaluate replenishment requirements based on an order-up-to
policy. According to order-up-to policy, whenever inventory levels drop below target
level, orders are placed to bring the combination of inventory and on-order up to a target
level. Target level and reorder point are deﬁned in days of demand at each location
depending on inventory management policy and logistics strategy in place. Thus, they
are dynamically calculated based on demand forecast information. During each review
period, average daily forecasts are recalculated and then multiplied by the deﬁned target
level (in days) to obtain the quantity to be maintained at the facility. Inventory is
reviewed daily. If inventory is below maximum target level, a replenishment order is

created. The quantity ordered is the amount necessary to meet maximum inventory

62

target. Thus, after reviewing replenishment requirements, retailers create replenishment
orders to distributors.

At the distribution center, replenishment orders from retailers are received,
processed, and shipped in the same manner as customer order fulﬁllment process at the
retail location. Unfulﬁlled replenishment orders are disposed and recorded as stock outs.
Distributor inventory position (including transit-inventory) and replenishment
requirements are then evaluated in a similar process to that at retail level. However,
target levels vary ﬁ'om those at the retail center depending on inventory and logistics
policies. Distributors then create replenishment orders to manufacturer. Manufacturer
uses these orders and other forecast information to plan production.

The production phase is treated as a ‘black box’ where internal processes such as
production scheduling are not included in the simulation. Instead, incoming orders
(replenishment orders from distributors) are ﬁlled depending on capacity availability.
Production lead time is incorporated into transit time from plant to distributors. The
model assumes that production lead time does not depend on the size of replenishment
orders. Product quantities are moved automatically to stock and ready to ship. This
model approach allows better investigation of distribution of a new product since effects
of production process complexities such as scheduling and set-up are minimized.

The last event is collection of statistics. Fill rates, average inventory, and units
sold are computed for each facility. Individual statistics are then aggregated to get
performance for the entire supply chain. The system returns to the ﬁrst event after the

last event is complete, repeating the same sequence until the simulation period is over.

63

3.2.2 Model Verification and Validation

Sargent (2000) presents a comprehensive assessment of model veriﬁcation and
validation approaches. Rodrigues (2004) successfully used this approach. Model
veriﬁcation refers to the process of ensuring that computer program of the computerized
model is performing as intended (Law and Kelton 2000; Sargent 2000). Model validation
refers to “substantiation that a computerized model within its domain of applicability
possesses a satisfactory range of accuracy consistent with the intended application of the
model” (Sargent 2000:50). Several model validation techniques are discussed in
literature including animation, degenerate tests, extreme condition tests, face validity,
internal validity, traces, etc (Law and Kelton 2000; Sargent 2000).

According to Sargent (2000), there are four types of validity that should be upheld
in a simulation model. First, conceptual model validity determines that the theories and
assumptions underlying the conceptual model are correct and model representations of
the system are reasonable for the intended purpose of the model. To ensure conceptual
validity, theories and assumptions underlying the model are tested using mathematical
analysis and statistical methods on problem identity. The main validation techniques for
this type of validity are face validation and traces. Face validity requires that experts
evaluate the model to ensure that on the surface, it seems reasonable for the intended
purpose. In this research, model relationships are veriﬁed through consultations with
managers (case studies) and academic advisors knowledgeable in simulation design. In
trace technique, the state of the simulated system (i.e., contents of event list, state of
variables, certain statistical counters, etc.) is displayed just after each event occurs and is

compared with hand calculations to see whether the program is operating as intended

64

(Law and Kelton 2000). A log ﬁle with each activity performed (i.e., order placements,
replenishment requests, inventory shipments, etc.) is created and any errors detected in
the model are evaluated and ﬁxed. Thus, conceptual model validity is achieved.

Second, computerized model verification ensures that the computer programming
and implementation of the conceptual model is correct. Computerized model veriﬁcation
was accomplished by testing if the model was programmed correctly in ARENA
simulation language. Traces validation technique discussed above is used. In addition,
each program path is subjected to extreme conditions to test whether it performs as
expected. The tests were successful, thus, the model is veriﬁed.

Third, operational validity determines that the model’s output behavior has
sufﬁcient accuracy for the intended purpose. All validation techniques can be used to
establish this validity. In this research, four techniques are used, viz., Degenerate Tests,
Extreme Condition Tests, Internal Validity Check, and Parameter Variability—Sensitivity
Analysis. The degeneracy of the model’s behavior is tested by appropriate selection of
values of the input and internal parameters. For example, an increase in lead time in one
stage of the model resulted in an increase in backorders and reduced service levels since
it mimics a constrained environment. Thus, operational validity is established. Extreme
condition tests are done to test if model is plausible for those conditions. The model
yielded expected results, thus, operational validity is conﬁrmed. Internal validity is
determined by conducting several model replications to establish the amount of internal
stochastic variability. High variability may lead to questionable model results.
Coefﬁcient of Variation (CV) is used to assess variability (Law and Kelton 2000). In

Parameter variability—sensitivity analysis, values of the input and internal parameters of a

65

model are changed to determine the effect upon the model’s behavior and its output.
Inventory and lead time parameters are altered to check if the resulting effects reﬂect a
real system.

Finally, data validity ensures that necessary data for model building, model
evaluation and testing, and model experiments are adequate and correct. To assess data
validity, collected data was tested using internal consistency checks, in consultation with

experts (advisors and company managers) and screening for outliers.

3.2.3 Experimental Factors

This research examines six independent factors: logistics strategy, launch
strategy, demand variability, competitive response, product advantage, and production
capacity. Demand variability, competitive response, and product advantage are modeled

as one variable — Demand Characteristics (refer to “Demand Characteristics” section).

1. Logistics Strategy

Logistics strategy is modeled at four levels, representing two alternative
strategies: Push vs. Pull and Fat vs. Lean. From case study interviews, it was established
that logistics strategy tends to be characterized by the way inventory is deployed and how
much inventory is maintained in the distribution system. Push strategy is characterized
by forward inventory deployment. Under this strategy, manufacturers push inventory to
retailers in anticipation of sales. To mimic this, target and maximum inventory levels at
retail facilities are higher than at distribution center. At the retail locations, target and

maximum inventory are modeled as 6 days while at distribution centers, target and

66

maximum levels are modeled as 4 days. These represent two order cycle times for
retailers and one order cycle time for distribution center.

Pull strategy represents a response-based strategy where minimal inventory is
held at retail locations. Under this strategy, companies withhold deployment of inventory
to the distribution system until receipt of orders. Pull strategy is modeled by having
higher target inventory levels and reorder point at distribution centers compared to retail
facilities. Accordingly, it is modeled as reverse of push system — at the retail location,
reorder point and target level are 4 days while at the distribution center, reorder point and
target level are 6 days. Reorder point and target level are kept the same to reﬂect an
order up to policy in both Pull and Push system. Having same total order cycle time (10
days) for both strategies enables better tests of inventory deployment effects.

Fat strategy involves maintaining inventory levels that are twice (or more) of the
inventory required to meet one order cycle. Here, companies maintain minimum
inventory level sufﬁcient to cover one order cycle. This reﬂects a supply chain system
with high safety or buffer stock. Thus, in Fat strategy, reorder point is modeled as 3 and
4 days at retail and DC, respectively, and target levels are 7 days at both retail and DC
locations. This means that there is sufﬁcient safety stock to cover one order cycle at
retail level (3 days lead time) and DC level (4 days lead time).

Lean strategy varies from “Fat” strategy in that target levels are kept the same as
reorder points. This reﬂects a “lean” supply chain where minimal inventory is held at
any location in the distribution system. Thus, reorder and target inventory levels at retail
and DC are 3 and 4 days, respectively, which represents maximum length of one order

cycle in each location.

67

The four levels enable assessment of effects of deploying inventory to retail
locations versus withholding inventory until orders are received. At the same time, it
enables assessment of effects of maintaining limited inventory versus keeping larger
inventory in the distribution system.

In this study, Push and Fat logistics strategies are surrogates of an anticipatory
strategy. This is because in a Push strategy, new product inventory sufﬁcient to meet two
order cycles’ requirement is ‘pushed’ to the retail level. In other words, not only is the
inventory pushed into the distribution system in anticipation of sales, but sufﬁcient buffer
stock is maintained at retail level to minimize stockouts. In a Fat strategy, sufﬁcient
buffer inventory is maintained at each supply chain stage. For example, this study
models a situation where target levels at distribution and retail facilities are set to one
week (sufﬁcient to meet at least two order cycles) and reorder point to one order cycle’s
worth of inventory. Since anticipatory strategy involves (i) forward deployment of
inventory and (ii) higher levels of safety stock at each facility in the distribution system,
Push and Fat strategies provide the best estimates of a new product launch performance
when an anticipatory strategy is adopted.

On the other hand, Pull and Lean strategies mimic a response-based strategy. Pull
strategy is modeled in terms of limited new product inventory deployment to the retail
level (sufﬁcient to meet only one order cycle, with a day’s worth of safety stock). In a
Lean strategy, target level is set equal to one order cycle (in days), which represents very
lean inventory in the entire distribution system. Therefore, Pull and Lean strategies

represent a situation where (i) new product inventory is only deployed into distribution

68

system according to demand requirements and (ii) limited cycle and safety inventory is

maintained in the entire system. This is representative of a responsive logistics strategy.

2. Launch Strategy

Launch strategy was modeled at two levels: national and rollout. National launch
represents a system where new product is introduced in all market segments
simultaneously. There are ten market segments in the simulation. Therefore, with
national launch, new product was deployed in all ten market segments at once. Rollout
launch strategy involves sequential launch where new product is introduced to select
markets and then introduced in other markets sequentially. To model rollout strategy,
new product was introduced in market segments 1 and 6 in the ﬁrst month, then to 2 and
7 in the third month, 3 and 8 in the ﬁfth month, 4 and 9 in the seventh month, and 5 and
10 in the ninth month, as shown in the table below. Similar stochastic distributions were
maintained in each market segment for both national and rollout strategies.

Table 3.1: Product Launch by Month and Market Segment

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Markets where product is launched

Month National Launch Rollout Launch

1 All segments 1 and 6

2 All segments 1 and 6

3 All segments 1, 2, 6, and 7

4 All segments 1, 2, 6, and 7

5 All segments 1, 2, 3, 6, 7, and 8

6 All segments 1, 2, 3, 6, 7, and 8

7 All segments 1, 2, 3, 4, 6, 7, 8, and 9

8 All segments 1, 2, 3, 4, 6, 7, 8, and 9

9 All segments All segments

10 All segments All segments

11 All segments All segments

12 All segments All segments

 

69

3. Demand Characteristics
(1) Demand Uncertainty

Demand uncertainty is modeled in terms of demand variability, which represents
variations from daily average demand. High demand variability represents higher
uncertainty because when demand variability is high, ﬁrms ﬁnd it difﬁcult to fulﬁll
customer orders and to forecast and plan production. Daily demand is generated
independently for each customer segment using TRIA(Min, Mode,Max) distribution in
Excel. Daily average demand was modeled as 100 units. Low variability is modeled as
TRIA(75,100,125) while high variability is modeled as TRIA(0,100,200). In ARENA,
TRIA(75,100,125) represents a daily demand mode of 100 units, daily minimum demand
of 75 units, and a maximum daily demand of 125 units. The low and high variability
distributions represent 0.1 and 0.4 coefﬁcient of variation (CV), respectively. This is
typical in consumer goods sector. For example, according to Waller at al. (1999), daily
demand variability in consumer products sector is low and may vary from C.V. = 0.1 to
0.3. Moreover, C.V. = 0.4 represents the highest coefﬁcient possible for a mean demand
of 100 while maintaining a symmetrical distribution. These extreme variations enable

better assessment of demand variation effect.

b) Competitive Response

Daily demand is reduced gradually from time of launch to the end of the year to
reﬂect intense competitive response. According to previous studies, new product
introduction triggers retaliatory behavior by the incumbent ﬁrms (Robinson 1988;
Kuester et al. 1999; Debruyne et a1. 2002). When retaliation is intense, demand for new

product is greatly reduced. To mimic a situation where competitive response is intense,

7O

average demand at both low variability and high variability situations is reduced by 65%
from the time of launch to end of the year as Table 3.2 shows.

Table 3.2: Effect of Competitive Response on New Product Demand

Month Effect Factor
1.00
0.95
0.90
0.85
0.80
0.75
0.70
0.65
0.60
0.55
0.50
0.45

 

 

 

 

 

 

 

 

 

 

 

 

 

5:3\OOO\10\UIJ>UJN~

 

 

 

 

c) Product Advantage

To model effect of new product advantage, demand is gradually increased from
time of launch to the end of simulation period. Product advantage reﬂects customer
value, which is the customers’ perceived beneﬁts and convenience derived from the new
product for the price they pay. Demand for a product with more advantages is stipulated
to grow faster and in a larger scale than demand for a low advantage product. To mimic
a situation where product advantage is high and consequently demand is growing rapidly,
new product demand is modeled as increasing from the time of launch up to the end of
simulation. Demand is increased by a factor each month as shown in Table 3.3 below.
The effect factors reﬂect information from interview with managers, which showed that
new product demand can increase by more than four times in the ﬁrst year of launch in

consumer goods industry.

71

Table 3.3: Effect of High Product Advantage on New Product Demand

 

 

 

 

 

 

 

 

 

 

 

 

 

Month Effect Factor
1 1.00
2 1.35
3 1.70
4 2.00
5 2.30
6 2.50
7 2.70
8 2.90
9 3.00
10 3.10
11 3.20
12 3.30

 

 

 

 

4. Firm Resources

In this research, production capacity is used as a surrogate for ﬁrm resources (i.e.,
an indirect measure of resources available to a ﬁrm to pursue its strategies). Capacity is
modeled at two levels: 70 percent and 100 percent. 100 percent indicates that a ﬁrm
fulﬁlls all orders while 70 percent capacity indicates that a company is able to fulﬁll only
70 percent of customer orders. The rationale is that small ﬁrms will have limited
capacity to meet demand relative to large ﬁrms. These percentages are considered

representative of typical ﬁrm capacities in relation to product demand.

3.2.4 Fixed Parameters

In this simulation, an order-up-to policy is used. The target level and reorder
points vary depending on logistics strategy used. Transit lead time between manufacturer
and distributors are stochastic with a mode of 3 days, a minimum of 2 days, and a
maximum of 4 days. The transit time from distributors to each retail facility is stochastic

with a mode of 2 days, a minimum of 1 day, and a maximum of 3 days. These transit

72

times reﬂect typical lead times in the durable and non-durable consumer goods sector.
The review period is modeled as 30 days, again representing typical review period in

durable and non-durable consumer goods industry.

3.2.5 Performance Variables

Four performance variables are examined in this research: order ﬁll rate, case ﬁll
rate, average system inventory, and proﬁtability. These performance measures are
widely used in extant literature (Chow et al. 1994; Glasserman and Wang 1998; Song
1998; Doloi and Jaafari 2001). Order ﬁll rate indicates how well a company is able to
complete customer orders. An order with an item missing will be considered incomplete.
Therefore, order ﬁll rate measures how many orders are shipped complete. Case ﬁll rate
captures the impact of stock outs over time (Rodrigues 2004). It indicates the number of
units shipped divided by the total units requested in each order. Order and case ﬁll rates
reﬂect a company’s ability to fulﬁll its delivery promises to customers.

Average system inventory indicates cumulative end of day inventory position
(storage and transit) divided by number of days at the distribution and retail locations
(Rodrigues 2004). The amount of inventory in the system affects inventory carrying cost
and risks associated with excess inventory. Therefore, it is an important performance
indicator.

The fourth measure is proﬁtability. It is calculated by subtracting total cost from
revenue. Product price (unit price) was set at $10, an arbitrary value. Four costs are
computed: production, inventory carrying, transportation, and stockout penalty.

Production is modeled as a “black box” -— production factors such as scheduling, set-

73

up/changeover time, etc., are not modeled. This is to ensure that scheduling and other
production dynamics and constraints to do not affect assessment of distribution and
inventory deployment processes in the model. Average system inventory is used to
compute production cost per unit taking into consideration production economies of
scale. Production cost per unit is reduced as volume increases to reﬂect production
economies. Production costs rates are shown in Table 3.4 below.

Table 3.4: Production Cost Rate

 

 

 

 

 

 

Ave System Inv Cost/unit
55000 $6.50
310000 $6.40
515000 $6.30
520000 $6.20
Else $6.00

 

 

 

 

Transportation and inventory carrying costs vary between distributors and
retailers. The costs are based on averages in the consumer goods sector as reported in
The GMA 2005 Logistics Survey (IBM Business Consulting Services, 2005) and from
interviews. According to the GMA 2005 survey, total logistics costs as a percentage of
sales averages 6.9 percent in consumer goods industry. Of this, outbound transportation
accounts for 40 percent. Thus, the cost of transportation as a percentage of sales is 2.76
percent. Based on the GMA 2005 survey, transportation cost would average $0.276.
Since shipments from plant to Distributors are likely to be in bulk (i.e., full truck load), it
is expected that average transportation cost will be lower than the average. However,
shipments from distributor to retailers are mostly less than truck load; therefore, average
transportation is likely to be higher than industry average. Consequently, shipment cost

from plant to distributors is modeled as $0.20 while from distributors to retailers is

74

modeled as $0.40. To account for transportation economies of scale, cost per unit is
decreased with increase in shipment volume as shown in Table 3.5 below.

Table 3.5: Transportation Cost Rate

 

 

 

 

 

 

 

 

 

 

Transit inventory DC cost/unit Retail cost/unit
51000 $0.20 $0.40
52000 $0.19 $0.38
S3000 $0.18 $0.36
54000 $0.17 $0.34
55000 $0.16 $0.32

Else $0.15 $0.30

 

 

According to the GMA 2005 survey, inventory carrying cost in the consumer
goods industry averages 25 percent. However, inventory carrying cost at retail level is
expected to be higher than at distribution level because retailers own inventory at a higher
cash value than distributors (Lambert and Pohlen 2001). Therefore, inventory carrying
cost is modeled as 30 percent at retail level and 15 percent at distributor level. Although
high volume is associated with gains from economies of scale, inventory carrying cost is
modeled as increasing on a per unit cost basis as inventory volume increases. As Zinszer
(1983) points out, there are different costs associated with different inventories.
Inventory carrying cost should therefore vary depending on type of inventory. For
example, the risk component associated with speculative stock is higher than cycle
(transaction) inventory due to ﬁnancial risks in case price declines and product demand
falls as well as additional handling and storage costs (Zinszer 1983).

Similarly, risk component for a new product will be higher than for products with
established demand. This is because new product is more susceptible to failure leading to
greater write-off cost, product demand decline leading to lower sales and possibly price
cuts, and limited secondary markets in case product fails in current market. Therefore,

inventory carrying cost for a new product should increase, rather than decrease, with

75

increases in inventory in stock. Risk costs may be so high that they erode any gains from
economies of scale. Consequently, companies maintaining higher new product inventory
will have higher carrying cost than those with minimal inventory. Table 3.6 shows
inventory carrying cost rates used in the model.

Table 3.6: Inventory Carrying Cost Rate

 

 

 

 

 

 

Ave system inventory DC cost/unit Retail cost/unit
£2500 $0.15 $0.30
55 000 $0.16 $0.31
57500 $0.17 $0.32
310000 $0.18 $0.33
Else $0.20 $0.35

 

 

 

 

 

Stockout penalty is modeled as ﬁxed rate based on order ﬁll rates as Table 3.7
shows. Several studies have examined stockout costs and penalties (Walter and Grabner
1975, Masters 1987, Corbett 2001). These include explicit cost such as some pre-agreed
penalty for stockout (i.e., charged by retailers) and implicit costs such as loss of customer
goodwill, lost sales when customers substitute product, etc. For a new product, it is
critical that customers get the product when they need it. Therefore, stockout costs are
high when dealing with new products and increase with higher levels of stockout both in
absolute terms and on a per unit cost basis.

Table 3.7: Stockout Penalty Cost

 

 

 

 

 

 

 

 

Order Fill Rate Penalty
295% but <100% $1,500
290% but <95% $3,000
285% but <90% $4,500

<85% $6,000

 

76

3.2.6 Number of Replications

The number of replications in a simulation study represents sample size, which
determines the accuracy of stochastic variables. Using Law and Kelton (2000: 512)
procedure to calculate the number of replications required to estimate population mean at
a speciﬁed precision, it is determined that 24 replications would be required per
experimental cell. According to this procedure, an approximate expression for the
minimum number of replications, 71:03), required to obtain an absolute error of B is given
by: n;(B) = min {i 2 n: t,-1,1-a/2\l(Sz(n)/i) S B}. na'(B) can be determined by iteratively
increasing i by 1 until a value of i is obtained for which ti-1,1-a/2\l(Sz(n)/i) S B. For
example, to estimate case ﬁll rate with an absolute error, [3, of 0.05 and a conﬁdence
level, a, of 90%, initial pilot runs were done and initial mean (u) and standard deviation
(82) were computed. A 82 = 0.02 was obtained from 24 replications. Critical values of t
give toms) = 1.7139. Using the equation {i 2 n: ti-1,1-a/2\/(Sz(n)/i) S B}, we obtain a
minimum of 24 replications (n = 24) to satisfy t,-1,1-a/2\/S2(n)/i S B requirement. A large
sample size indicates that estimated mean is closer to true mean. Thus, large number of
replications increases accuracy and conﬁdence in simulation results. In this study, 30

replications are done per experimental cell.

3.2.7 Data Analysis

Multivariate analysis of variance (MANOVA) and univariate analysis of variance
(ANOVA) are used to test the main effects and interaction effects stipulated in the
hypotheses. MANOVA is a dependence technique that measures differences for two or

more metric dependent variables based on a set of categorical (nonmetric) variables

77

acting as independent variables (Hair et al. 1998). MANOVA enables comparison of
groups formed by categorical independent variables on group differences in a set of
interval dependent variables. AN OVA technique determines whether samples from two
or more groups come from populations with equal means. ANOVA tests for differences
in means between groups, which can also be done using t-test, is preferred because it is
more efﬁcient (gives more information with fewer tests) and allows for detection of
interaction effect between two variables. In this research, there are more than two
dependent variables (performance measures) and multiple independent variables, which

indicate that use of both ANOVA and MANOVA is appropriate.

78

CHAPTER 4: RESULTS
4.0 Introduction
This chapter discusses study results. The ﬁrst part discusses MANOVA
assumptions, effects of assumption violations, and conformity of study data to the
assumptions. The next section presents multivariate and univariate results. This is
followed by results of hypotheses tests examining support or lack of support for study

hypotheses. The last section discusses results implications.

4.1 MANOVA Assumptions and Data Conformity to Assumptions

MANOVA enables comparison of groups formed by categorical independent
variables on group differences in a set of interval dependent variables (Hair et al. 1998).
To develop valid results, MANOVA requires ﬁve assumptions, namely; (i) observations
are statistically independent, (ii) dependent variables follow a multivariate Normal
Distribution, (iii) the variance-covariance matrices of the dependent variables are equal
for all treatment groups, (iv) linearity and multicollinearity among dependent variables,
and (v) absence of outliers.

One of the most stringent assumptions is independence of observations.
MANOVA is used where independent variables are categorical and dependent variables
are continuous. Violations occur when there is lack of independence among observations
for several reasons. For example, violations may occur if there is serial correlation (for
measures taken overtime), when information is gathered in a group setting where
respondents’ common experiences cause a subset of individuals to have answers that are

somewhat correlated, or when extraneous and unmeasured effects create dependence

79

among respondents. Since this study used independent simulation replications,
independence of observations requirement is satisﬁed.

Multivariate normality assumes that joint effect of two variables is normaly
distributed. Violations of this assumption have little impact when large samples are used.
Violations affect application of Box’s M test, which is used to test for equality of
covariance matrices. Generally, F-test is robust to non-normality if it is caused by
skewness rather than outliers. Skewness and Kurtosis statistics (Table 4.1) and
Frequency histograms (Figure 4.1) for dependent variables indicate that there are
departures from multivariate normality. However, past studies have shown that
MANOVA is robust to departures from multivariate normality when treatments have
equal sample sizes (Hair et al. 1998). Frequency histograms (Figure 4.1) show that there
are no outliers in the data, which satisﬁes the ﬁfth assumption.

Table 4.1 - Skewness and Kurtosis Statistics

 

 

 

 

 

 

 

 

 

 

Skewness Kurtosis
Statistic Std Error Statistic Std Error
Order Fill Rate -1.002 0.046 -0.307 0.091
Case Fill Rate -0.760 0.046 -0.616 0.091
Ave System Inventory 0.932 0.046 -0.506 0.091
Proﬁtability 0.783 0.046 -0.930 0.091

 

 

 

Equality of variance-covariance matrices concerns presence of substantial
differences in the amount of variance of one group versus another for the same variables.
However, violation of this assumption has minimal impact if groups are approximately
equal in size. In addition, the assumption may be violated without greatly affecting the
level of signiﬁcance or power. In this study, all groups have same size. Levene’s Test of
homogeneity of variances is used to test the assumption that each group (category) of

independent variables has same variance. Levene’s Test is signiﬁcant for all dependent

80

variables at p<.001, which indicates that data fails the assumption of equal variances.
Box’s M test is used to test for equality of covariance matrices. This measure is very
sensitive to departures from normality. Box’s M test results (Box’s M = 28880.4,
F=28.9, dfl =950, d12=857109, p<.001) indicate that the null hypothesis that covariance
matrices of the dependent variables are equal across groups is rejected.

Figure 4.1 - Frequency Histograms for the Dependent Variables

 

 

 

 

 
   

 

 

 

 

 

 

 
 

 

  

 

 

      

1500 1500
1250! 1250I
1000t 1000I
5‘ >.
I: O
t” 5
g 7501 a 750i
9 E
L u.
250* 250' A
o, 0'
-54 -59 ~64 -69 ~74 .79 -84 .89 -94 -99 .72 .75 .79 .91 .94 .87 .90 .93 .96 .99
Order Fill Rate Case Frl Rate
1500 1500
12501 1250!
1WD' 5 10001
‘3 9
O 7501
E, 7501 5
LL “L 500.
500‘
20 h.
5 ‘ o
0 8 8 8 8 8 8 8 8 8 8 8
O O O O O O O O O O O
O O In m to ID no in in ID ID In no
g g g § § § 8 § § 3 r~ N N N is N r~ N is N is
h '- ts v- 00 10 w '- 0 0 O '- (O (D Q
1- v- I- I- v- N N '- "' '- " N N N N
Average System Inventory Proﬁtability

 

 

 

Linearity and multicollinearity among dependent variables was also tested.
General linear hypotheses models assume linearity among variables. Therefore, it is

important to test for non-linear relationships. To determine if measures are signiﬁcantly

81

correlated, Bartlett’s test of sphericity is used. The test examines correlations among all
dependent variables and determines whether, collectively, signiﬁcant intercorrelation
exists. The result, Bartlett’s = 178,872 with 9df, p<.001, indicates that signiﬁcant
intercorrelations exist, which means that use of MANOVA is appropriate.

The foregoing discussion indicates that study data conformed to most critical
requirements and assumptions for MANOVA so the results can be accepted as valid.

Consequently, it is appropriate to use MANOVA to test study hypotheses.

4.2 Overall Results

In this study, four independent variables are tested: logistics strategy (4 levels),
launch strategy (2 levels), demand characteristics (6 levels), and production capacity (2
levels), resulting in 96 scenarios. Two alternative logistics strategy models are presented:
Pull strategy is compared with Push strategy, and Fat strategy is compared with Lean
strategy [Refer to Chapter 3, section 3.2.3, Logistics Strategy, p.66]. Therefore, although
logistics strategy is presented as having 4 levels in proﬁle plots in this Chapter, in effect,
there are two alternative levels. Similarly, demand characteristics are modeled in 6
levels. However, as discussed in Chapter 3 [Refer to Chapter 3, section 3.2.3, “Demand
Characteristics,” p.70], there are three demand levels, and for each level, there is a low
and high variability situation. Therefore, proﬁle plots in this Chapter show low and high
variability results for each demand level.

Each scenario is replicated 30 times, resulting in a total sample size of 2880
observations. There are four dependent variables: order ﬁll rate, case ﬁll rate, average

system inventory, and proﬁtability. MANOVA and univariate ANOVA tests are done

82

using SPSS. Table 4.2 gives a summary of multivariate and univariate test results. These
results are discussed in detail in subsequent sections.

Table 4.2 - Summary of Significance Test

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Multivariate Univariate Results
meets Results OFR CFR SW8“ Proﬁtability
Inventory

Logistics Strategy (LOG) *** *** *** ”at at"
Launch Strategy (LAU) *** * Ns alum *Ilul:
Demand Variability (DEM) *** *** *** *H *1”!
Production Capacity (CAPL *** *** *** "I: ah“!
LOG * LAU alumi- an" *Ilnll mu: “up
LOG * DEM can: alulnl: an“: *Ilnll *Iluk
LOG * CAP *alult uni: uni: *Iltalt at“:
LAU * DEM *** II: *Ikal: talus *Ilul:
LAU "‘ CAP *** ns NS *alul: ns
DEM * CAP *** ns NS mink auras
LOG * LAU * DEM *** * *IHI aunt: *alnl:
LOG * LAU * CAP *** ns * ns at:
LOG * DEM * CAP *** ns NS *** *alnlt
LAU * DEM * CAP ns ns Ns ns ns
LOG * LAU * DEM * CAP ns ns Ns ns ns

 

 

 

 

m Signiﬁcant at p<.001

"' Significant at p<.05

ns — Not signiﬁcant
4.3 Multivariate Results

SPSS provides results of four criteria for testing multivariate differences between
groups: Wilk’s Lambda, Pillai’s Trace, Hotelling’s Trace, and Roy’s Largest Root.
Wilk’s Lambda examines whether groups are different without being concerned with
whether they differ on at least one linear combination of the dependent variables. It is
appropriate where assumptions appear to be met. Pillai’s Trace has minor differences
from Wilk’s Lambda. It is the sum of variance that can be explained by the calculation of

discriminant variables and is robust when MANOVA assumptions are not met, especially

where sample sizes are small, cell sizes are unequal, or covariances are not homogeneous.

83

Hotelling’s Trace has minor differences from Wilk’s Lambda and is safely ignored in
most cases. Roy’s Largest Root measures the differences only on ﬁrst canonical root. It
is most appropriate when all dependent variables are strongly intercorrelated on a single
dimension but is most severely affected by violations of MANOVA assumptions. In
other words, it is the most powerﬁrl test when all assumptions are strictly met and
dependent measures are representative of a single dimension of effect. Generally, the
best measure to use is one that is most immune to violations of underlying MANOVA

assumptions. Consequently, Pillai’s Trace is reported in this study. Table 4.3 shows

 

multivariate test results.

Table 4.3 - Multivariate Test Results

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Degrees of Partial Eta Observed
Effect F Freedom Sig. Squared Power
Logistics Strategy (LOG) 1,278 12 8,349 0.000 0.647 1.000
Launch Strategy (LAU) 49,425 2,781 0.000 0.986 1.000
Demand Characteristics (DEM) 1,253 2,781 0.000 0.643 1.000
Production Capacity (CAP) 289 20 11,136 0.000 0.342 1.000
LOG * LAU 271 12 8,349 0.000 0.280 1.000
LOG * DEM 163 60 11,136 0.000 0.467 1.000
LOG * CAP 57 12 8,349 0.000 0.076 1.000
LAU * DEM 205 20 11,136 0.000 0.269 1.000
LAU * CAP 29 4 2,781 0.000 0.039 1.000
DEM * CAP 70 20 11,136 0.000 0.112 1.000
LOG "‘ LAU * DEM 27 60 11,136 0.000 0.128 1.000
LOG * LAU * CAP 4 12 8,349 0.000 0.005 0.998
LOG "‘ DEM * CAP 7 60 11,136 0.000 0.036 1.000
LAU * DEM * CAP l 20 11,136 0.116 0.002 0.920
LOG * LAU * DEM * CAP l 60 11,136 0.199 0.006 0.998

 

 

 

 

 

 

 

 

Effect column shows independent variables and interactions between them. The

next column shows F-ratio. F-ratio focuses on independent variables and seeks to

84

answers the question “is each independent variable effect signiﬁcant?” The larger the F
statistic, the greater the likelihood that mean differences are due to something other than
chance. Degrees of freedom columns indicate the degrees of freedom used to obtain
observed signiﬁcance levels of multivariate test. The next column indicates signiﬁcance
level (p-value), which is the conditional probability that a relationship as strong as the
one observed in the data would be present if null hypotheses were true (chance of making
Type I error). Typically, results that yield p<.05 are considered borderline signiﬁcant
while lower p-values, i.e., p<.001 are considered highly signiﬁcant.

The ﬁfth column shows Partial Eta Squared (partial 112), the proportion of total
variability in dependent variable that is accounted for by variation in independent
variable, excluding other factors from total non-error variation (Cohen 1973, Levine and
Hullett 2002, Pierce et a1. 2004). It is a measure of effect size. Partial n2 values range
from 0 to 1, with larger values indicating stronger effect. The last column, Observed
Power, indicates power of test or the ability to reject null hypotheses when it is actually
false (chance of making Type 11 error). High observed power values (i.e., over 0.9) are
recommended.

In MANOVA, interaction terms represent joint effect and must be examined ﬁrst.
There is one four-way interaction, four three-way interactions, six two-way interactions,
and four main effects. As Table 4.3 shows, signiﬁcant three—way interactions (at p<.001)
are: Logistics Strategy x Launch Strategy x Demand Characteristics (LOG*LAU*DEM),
Logistics Strategy x Launch Strategy x Production Capacity (LOG*LAU*CAP), and
Launch Strategy x Demand Characteristics x Production Capacity (LAU*DEM*CAP).

All two-way interactions are signiﬁcant at p<.001. They are: Logistics Strategy x Launch

85

Strategy (LOG*LAU); Logistics Strategy x Demand Characteristics (LOG*DEM);
Logistics Strategy x Production Capacity (LOG*CAP); Launch Strategy x Demand
Characteristics (LAU*DEM); Launch Strategy x Production Capacity (LAU*CAP); and
Demand Characteristics x Production Capacity (DEM*CAP). All main effects are
signiﬁcant at p<.001. The main effects are Logistics Strategy, Launch Strategy, Demand
Characteristics, and Production Capacity. These interaction and main effects are

discussed in the following sections.

4.4 Univariate Results

This section discusses univariate test results for main and interaction effects.

4.4.1 Univariate Results: Three-way Interactions
Table 4.5 presents univariate results for signiﬁcant three-way interactions.

Table 4.4 - Univariate Test for Three-way Interactions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source of De rees of . Partial Eta Observed
Variation Dependent Variable F Frgeedom Srg. Squared Power
LOGISTIC * Order Fill Rate 2 15 2,784 0.004 0.012 0.982
LAUNCH * Case Fill Rate 8 15 2,784 0.000 0.041 1.000
DEMAND Avg System Inventory 76 15 2,784 0.000 0.290 1.000
Proﬁtability 3 15 2,784 0.000 0.015 0.996
LOGISTIC 1: Order Fill Rate 1 3 2,784 0.221 0.002 0.392
LAUNCH * Case Fill Rate 3 3 2,784 0.041 0.003 0.669
CAPACITY Avg System Inventory 2 3 2,784 0.076 0.002 0.581
Proﬁtability 4 3 2,784 0.01 1 0.004 0.811
LOGISTIC . Order Fill Rate 1 15 2,784 0.718 0.004 0.526
DEMAND * Case Fill Rate 1 15 2,784 0.198 0.007 0.813
CAPACITY Avg System Inventory 4 15 2,784 0.000 0.024 1.000
Proﬁtability 6 15 2,784 0.000 0.030 1.000

 

 

LOG*LAU*DEM has signiﬁcant effect on Order Fill Rate (OFR) at p<.05 and
Case Fill Rate (CFR), Average System Inventory, and Proﬁtability at p<.001. Partial n2

values indicate that a strong effect on average system inventory (partial n2 = 0.29) and

86

weak effect on OF R and proﬁtability (partial n2 = 0.012 and 0.015, respectively). Figure
4.2 and Figure 4.3 show the proﬁle plots of LOG*LAU*DEM effect on OF R and CFR.

Figure 4.2 - Effect of LOG*LAU*DEM on Order Fill Rate ‘

 

LOG*DEM => OF R for National Launch

 

 

 

 

 

 

(0
§ ILogistics
2 w *—-4—_
2 I»———-———--°-~~-"" "Push
ﬁt .85' '
(U a
2 .80 I PU"
8
d—v .75: D
g ’.__....._fat
:5; 70' ’.—-~O’.'.
“J .65 v" _ _ _ _ ° Lean
3. i 3. i 3. 2):.
9 9 a 9. e a
>
< <
Demand Characteristic

LOG*DEM => CPR for Rollout Launch

 

 

 

 

 

 

m
é 'Logistics
2 ~ x ~ ‘ '
_ ,900——-~ —--—-O———r——-¢———1| a
2 ‘7 Push
'§ .85‘
(U o
2 .80« Pull
8
E '75' at ° Fat
I?) 70‘ ’0‘. o’.-.—’. -.
“J .65 I" _ ‘f _ _ ° Lean

a. e a. r; a. i

D ’ D D

.. 9. 9 a I a

< <
Demand Characteristic

 

 

 

 

' Demand Characteristics levels in all proﬁle plots are: LD=Low demand level (high competitive
reaction); LV=Low demand variability; Ang=Average demand level (low competitive reaction, low
product advantage); HD=High demand level (high product advantage); HV=High demand variability. Low
and high demand variability are put adjusted to each other for each demand level for ease of interpretations.

Logistics Strategy is presented at 4 levels in all proﬁle plots in this Chapter. However, two alternative
strategies are captured: Push vs. Pull or Fat vs. Lean [Refer to Chapter 3, Section 3.2.3, p.66].

87

Figure 4.3 - Effect of LOG*LAU*DEM on Case Fill Rate

 

 

 

 

 

 

 

 

 

 

 

 

LOG'DEM => CFR for National Launch
1.00
2 ' I
g .95: 'lfgjstics
a .90 I D
c 1 Push
573) .85 4 '- - -
(D J n
2 .80 Pull
8 75
g ' 1 ° Fat
93 .701 _
“J .65 - _ _ - ° Lean
3. E 3. t 3. i
o d o d D o
.1 _r g: g: I I
< <
Demand Characteristics
DEM'LOG => CF R for Rollout Launch
1.00 , r. _ , , , - -
rn ’ ‘ -O' ‘ l
5 95 . x ’ ‘ . .
g - .,___~,._,'__..___.__._°~__~_._Logrstics
__ .901 _ '
(D ’ O n
E .85 :T‘VI' "ur" "~n-__PU3h
N
g .801 ° Pull
*‘ .7 .
E 5 ° Fat
7;; .70[ _
L“ .65 _ _ _ _ ° Lean
3. a a. a a. i
o d o o’ 0 o
_r .4 gr gr I I
< <
Demand Characteristics

 

According to these plots, under a Pull logistics strategy, OFR and CPR differ
between national and rollout launch when demand variability is high for low and high
demand levels. For example, OFR and CFR are highest when Pull logistics strategy is
used and demand variability is low, irrespective of demand level (i.e., low, high,
average). However, when Pull logistics strategy is used and demand variability is high,
OF R and CFR are substantially lower in a rollout launch than a national launch strategy.

Proﬁle plots indicate that under a Pull strategy, when demand variability increases from

88

 

low to high, OFR decreases by about 4 percent for national launch and about 8 percent
for rollout launch. Similarly, CF R decreases by about 3 percent for national launch and
about 7 percent for rollout launch when demand variability increases from low to high.
This shows that Pull logistics strategy is more sensitive to variations in demand and
launch strategy. There are no signiﬁcant differences in OF R and CFR between national
and rollout strategy for other logistics strategies (Push, Fat, or Lean), irrespective of
demand levels. Therefore, the multivariate signiﬁcance of LOG*LAU*DEM is attributed
to Pull logistics strategy.

Figure 4.4 - Effect of LOG*LAU*DEM on Average System Inventory

 

DEM'LOG => System Inv for National Launch

 

 

 

 

 

25,000
‘8 I
2 20.0001 .Logistics
— '_ -
(D O
.5, 15.000 . - - .Push
1, 10.0001 PU"
2
(U O
.§ 5.0001 Fa‘
Loil-j- | _
0 ° Lean
2. i
O O
.1 I

 

Demand Characteristic

LOG'DEM =>System lnv for Rollout Launch

 

 

2 25.000
3 20 000 Logistics
1 _ _

'g r.
E” 15.000‘ ,2. gnu-"Push
g ' ° P u
.8 10'000‘ I ,—¢ID_U

D
g 5'000I —.Fat
3 t
uui ° Lean

 

 

 

LD,LV °
LD HVt
A gD,LV.
Ang,H\A
HD LV.
HD HV

Demand Characteristic

 

 

 

89

Figure 4.5 - Effect of LOG*LAU*DEM on Profitability

 

 

   
  
   
 

 

 

 

 

 

    
  
 

 

 

 

 

LOG'DEM => Proﬁtability for Nat. Launch
3,000,000
‘3 I . .
3 2,500,000. ..L_°9_'S"°s
g 2.000.000« ° push
O) ---
«‘6 1, .00 .
2 500 0 ° Full
33, 1,000,000<
2 ° Fat
'5; 500,000 —-
L” 0' _ _ _ _ ° Lean
3. E 3. 2):. 3. E
- o o -
9 3 gr 9):: :1: E
< <
Demand Characteristic
DEM'LOG => Proﬁtability for Roll. Launch
3,000,000
2
§ 2500.000 Logistics
_ I- '
% 2,000,000' : n Push
5 1.500.000: 1;-
5 Pull
3 1.0000001
2 a Fat
.3 500,000 _
w 0_ _ _ _ _ ° Lean
3. E 3. i 3. i
o d o d o o
.J .1 g: g) I I
< <
Demand Characteristic

 

 

 

Figure 4.4 and Figure 4.5 show that irrespective of logistics strategy, national
launch results in higher average system inventory and proﬁtability than a rollout strategy
for all demand levels. Since national launch involves product introduction in all
customer segments, it is expected to results in more sales, greater average system
inventory, and higher proﬁts than a rollout strategy. At the same time, when demand is
either low or average, Push, Pull, and Fat logistics strategies result in almost the same

levels of average system inventory when rollout strategy is adopted but different levels

90

when national launch is adopted. However, when demand is high, Fat logistics strategy
results in substantially higher inventory than both Push and Pull strategies. Surprisingly,
demand variability does not affect average system inventory and proﬁtability, except
under Pull logistics strategy when demand is high. However, the effect seems negligible.

LOG*LAU*CAP signiﬁcantly affects both CFR and proﬁtability at p<.05 but not
OFR and average system inventory. However, partial n2 values (.003 and .004,
respectively) indicate that effect is very weak. Figure 4.6 indicates that when Pull
logistics strategy is used, rollout strategy results in lower CFR for a 100 percent
production capacity than a 70 percent capacity. This shows that Pull logistics strategy is
more sensitive to capacity variation and is the main contributing factor to the observed

signiﬁcant results.

Figure 4.6 - Effect of LOG*LAU*CAP on Case Fill Rate 2

 

 

 

   

 

 

 

 

 

 

 

 

LOG'LAU => CFR at 100% Cap LOG'LAU => CFR at 70% Cap
1.00 100

g 951 ' g .951
2 .901 g .90l
E E
‘50 35* '5 .851
w a
2 .80 l 2 .80 1
g Launch 3 Launch
g .75 ‘ o g 75 1
3,- Nat. “5 a I:
“J .701 m .701 mm

.65 - _ ° Roll. .65 - a ° Roll.

Push Pull Fat Lean Push Pull Fat Lean
Logistics Strategy Logistics Strategy

 

 

 

For other logistics strategies, there are no signiﬁcant differences in CFR between
the two levels of capacity under national and rollout launch strategies. Intuitively, larger

capacity allows for more orders to be processed than if capacity is constrained, which

 

2 Logistics Strategy levels represent two alternative strategies: Push vs. Pull (models whether inventory is
pushed into the distribution system is held at plant/DC), and Fat vs. Lean (captures the amount of inventory
deployed in the distribution system). [Refer to Chapter 3, Section 3.2.3, p.66].

91

results in higher sales and proﬁts. Consequently, production capacity moderates the
effect of interaction between logistics and launch strategies on CFR. Figure 4.7 and
Partial n2 values show that LOG*LAU*CAP effect is negligible.

Figure 4.7 - Effect of LOG*LAU*CAP on Profitability

 

 

 

 

 

 

 

 

 

 

LOG'LAU => Proﬁtability at 100% Cap LOG'LAU => Proﬁtability at 70% Cap
3.000.000 3,000,000
2 2,500,000- 3 2.5000001
53 8
2
i 2,000,0001 .—., 2.000.000.
c s
a 9
a 1.500.0001y——/°*—°\ a 1.500.000 {A
5 l 5 l
U ‘U _________
g 1,000,000 1* """"" ° """" ' ......... 0Launch :3 1000.000 1» ----- ' ° ~~~~~~~~ "ETC"
3:: 0 .§ a
6’1 500.0001 Nat. U13 500,000. _"Nar.
0 ° Roll. 0 _ - ° Roll.
Push Pull Fat Lean Push Pull Fat Lean
Logistics Strategy Logistics Strategy

 

 

 

LOG*DEM*CAP signiﬁcantly affects average system inventory and proﬁtability
(at p<.001), but not OF R and CFR. However, partial n2 values suggest a weak effect on
both average system inventory (partial n2=.024) and proﬁtability (partial n2=.03).
Moreover, marginal means plots (Figure 4.8 and Figure 4.9, next page) show that
LOG*DEM*CAP has weak effect on average system inventory and proﬁtability. As
expected, average system inventory and proﬁtability are slightly higher when capacity is

100% than when capacity is 70% for all logistics strategies at the same demand level.

92

Figure 4.8 - Effect of LOG*DEM*CAP on Average System Inventory

 

 

    
  
 

 

 

 

 

    
  

 

 

 

DEM'LOG => System Inv at 100% Capacity

m 25,000
if: ,Logistics
§ 20.0001 _-
(U
.3 150001 "D’s"
(T: J ._I a
2 10'000‘ O, —PUll
'3 /
7; a Fat
g 5,000 d _
Lil 0_ _ - ° Lean

a. .2. g. i a. a

o ' o o ’

.1 3 g’ (>35 I g

< <
Demand Characteristic
DEM‘LOG => System Inv at 70% Capacity
25.000

2
£3 20.0001 .ngistics
.2 15,000. 3”" ° Push
9 c--
(B
5 10.0001 ,; /—'“ ° Pull
U —
2 ' a
E 5.000 _F31
«‘5 - a
“J 0 i _ _ _ _ Lean

a. e g. e. a. a

o - -

.1 9 g» 9: g 2

< 9:
Demand Characteristic

 

 

 

Figure 4.9 - Effect of LOG*DEM*CAP on Profitability (see continuation next page)

 

 

 

 

 

 

LOG'LAU => Proﬁtability at 100% Cap
3,000,000

g 2,500,0001

5’

3 2.0000001

.5

9

a 1.500.0001/_ﬁ\

2 I

3 -------- 0 -------- -~- L h

5 1,000,000" ------- . 3""0

e l ,

u”; 500.000- mm“-
0 _ - ° Roll.
Push Pull Fat Lean

Logistics Strategy

 

 

 

93

 

 

LOG'LAU => Proﬁtability at 70% Cap

3.000.000

 

2.500.000‘

2.000.000!

Estimated Marginal Means

500.000 4

 

0

1.500.000 {A
I

1.000.000 0 ‘ ' '

‘
\--‘
--

Launch
II

 

Nat.

 

 

Push

P0”

Fat

Logistics Strategy

° Roll.
Lean

 

 

4.4.2 Univariate Results: Two-way Interactions

Table 4.5 shows univariate results of the two-way interactions.

Table 4.5 - Univariate Test for Two-way Interactions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source of . De recs of . Partial Eta Observed
Variation Dependent Variable F Frgeedom S'g' Squared Power
Order Fill Rate 8 3 2,7 84 0.000 0.008 0.988
LOGISTIC * Case Fill Rate 33 3 2,784 0.000 0.034 1.000
LAUNCH Avg System Inventory 1,670 3 2,784 0.000 0.643 1.000
Proﬁtability 85 3 2,784 0.000 0.084 1.000
Order Fill Rate 30 15 2,784 0.000 0.138 1.000
LOGISTIC * Case Fill Rate 15 15 2,784 0.000 0.073 1.000
DEMAND Avg System Inventory 2,804 15 2,784 0.000 0.938 1.000
Proﬁtability 225 15 2,784 0.000 0.548 1.000
Order Fill Rate 40 3 2,784 0.000 0.041 1.000
LOGISTIC * Case Fill Rate 65 3 2,784 0.000 0.065 1.000
CAPACITY Avg System Inventory 60 3 2,784 0.000 0.060 1.000
Proﬁtability 68 3 2,784 0.000 0.068 1.000
Order Fill Rate 3 5 2,784 0.005 0.006 0.907
LAUNCH * Case Fill Rate 11 5 2,784 0.000 0.020 1.000
DEMAND Aﬁystem Inventory 2,588 5 2,784 0.000 0.823 1.000
Proﬁtability 2,129 5 2,784 0.000 0.793 1.000
Order Fill Rate 2 1 2,784 0.122 0.001 0.340
LAUNCH * Case Fill Rate 1 1 2,784 0.239 0.000 0.217
CAPACITY Avg System Inventory 61 1 2,784 0.000 0.022 1.000
Proﬁtability 4 1 2,784 0.051 0.001 0.495
Order Fill Rate 0.4 5 2,784 0.832 0.001 0.164
DEMAND " Case Fill Rate 0.5 5 2,784 0.798 0.001 0.179
CAPACITY Avg System Inventory 287 5 2,784 0.000 0.340 1.000
Proﬁtability l3 5 2,7 84 0.000 0.023 1.000

 

 

94

 

LOG*LAU signiﬁcantly affects all dependent variables at p<.001. Partial n2
values indicate that the effect is greatest on average system inventory (.643) and weakest
on OFR (.008). As Figure 4.10 and Figure 4.11 show, under Pull logistics strategy,
national launch results in higher OFR and CFR than rollout launch. However, when
Push, Fat, or Lean strategies are used, both national and rollout launch strategies result in
same level of OFR and CFR. This ﬁnding is surprising given that both launch strategies
have different demand sizes (total market served). OFR and CFR are marginally higher
for rollout launch than national launch under a Push strategy.

Figure 4.10 - Effect of LOG*LAU on Order Fill Rate

 

 

 

 

 

1 .00
U)
C
(U
(D
2
E
E
'9
(I!
2
8
‘6 Launch
E _
'3 0
Lu National
- _ ° Rollout
Push Pull Fat Lean

Logistics Strategy

Figure 4.12 indicates that for national launch, average system inventory is highest
when Fat logistics strategy is used and lowest when Lean logistics strategy is used. This
makes intuitive sense since Fat strategy involves holding greater buffer inventory at
various stages in the supply chain. At the same time, Lean strategy has the least
inventory since only stock sufﬁcient to meet one order cycle is maintained throughout the
distribution system. For rollout launch strategy, the same pattern is observed — Fat

logistics strategy results in highest average system inventory while Lean strategy results

95

in lowest average system inventory. Surprisingly, Push and Pull logistics strategies have

the same level of average system inventory when rollout strategy is adopted but different

when national launch is adopted.

Figure 4.11 - Effect of LOG*LAU on Case Fill Rate

 

1.00
.95 '
.90 i
.85 i
.801

.75 ‘

Estimated Marginal Means

.701

 

 

.65

Launch

 

° Nat.

° Roll.

 

 

Push Phil :5: Lean

Logistics Strategy

Figure 4.12 - Effect of LOG*LAU on Average System Inventory

25.000

Estimated Marginal Means

20.000 ..

 

 

 

 

 

Push PL" Fat Lean

Logistics Strategy

As Figure 4.13 indicates, national launch strategy results in higher proﬁts than a

rollout launch strategy for all logistics strategies. Unlike rollout strategy, national launch

involves product introduction

in all market segments. Consequently, there are more

sales, and for proportional cost structures, greater proﬁts. Proﬁts are highest when Pull

logistics strategy is used in a national launch. Even though both Pull and Push logistics

96

strategies have the same level of average system inventory when rollout strategy is
adopted, Pull strategy results in higher proﬁts than Push strategy. This can be attributed
to high inventory carrying cost at the retail level in a Push strategy compared to a Pull
strategy. Moreover, Push strategy has higher stockouts than Pull strategy as indicated by
lower OFR and CFR, which results in lower sales and proﬁts. A surprising ﬁnding is that
although Fat logistics strategy results in average system inventory that is almost twice
that of a Lean logistics strategy for both national and rollout launch strategies, the
difference in proﬁtability is relatively smaller. This can be attributed to higher inventory
carrying cost for Fat logistics strategy, which signiﬁcantly reduces proﬁtability even
though service levels are high. A Lean strategy on the other hand has lower system
inventory and consequently lower inventory carrying cost. However, it results in more
stockouts as attested by low OFR and CF R levels.

Figure 4.13 - Effect of LOG*LAU on Profitability

 

 

 

 

3,000,000

3 2,500,000

é

n 2000000

.5

9

g 1.500.ooma//°\°\,.

3 ..........

E: 1,000,000:a """"" ° ° ------- uLaunc"

'fg o

LU 500000 mNaL
0 _ - ° Roll.
Push Pull Fat Lean

Logistics Strategy

LOG*DEM interaction signiﬁcantly affects all performance variables at p<.001.
Partial n2 values indicate that the effect is greatest on average system inventory (partial n2

= 0.938) and proﬁtability (partial n2 = 0.548). Figures 4.14 and 4.15 indicate that OFR

97

and CF R are highest when Pull logistics strategy is used and demand variability is low.

However, when demand variability is high, Fat logistics strategy results in highest OFR

and CF R. This makes intuitive sense. Fat strategy provides buffer inventory necessary to

cushion the effect of highly variable demand. Consequently, it is expected to result in

better performance when demand variability is high than the alternative Lean strategy.

Figure 4.14 - Effect of LOG*DEM on Order Fill Rate

Estimated Marginal Means

 

 

 

 

 

c‘ ,‘1- .0.
. J
"L istics
.90ii—-—-g.——-——‘B~——e-—-"°‘-"€3_og_
.851 ° Push
'80‘ 0 Pull
.75« —
’..n a Fat
7 —-‘P —-
0‘ .’hl-.’.ﬂ D
65" _ _ _ _ Lean
3. i 3. 3):. 3. i
- o -
9 9. g: g) g E
< <
Demand Characteristic

Figure 4.15 - Effect of LOG*DEM on Case Fill Rate

Estimated Marginal Means

 

 

 

 

 

1.00c .\
.951 “‘9'"
“*“+-‘€‘--P*-+—- Logistics
.901 K K __
(K C p’ .
.851 "1’” ‘9’ “’ ° Push
‘80‘ 0 Pull
.751 —
° Fat
.701 —o
.65 - _ _ - 0 Lean
3. a a. i a. i
9 9 ‘9» % E 2
> >
< <

Demand Characteristic

98

Push logistics strategy results in more uniform OFR and CFR levels irrespective
of demand variability. This suggests that Push logistics strategy is more resilient to
effects of demand variability compared to Pull strategy. Since Push strategy involves
deployment of higher levels of inventory at retail level, any changes in demand are
accommodated and fulﬁlled. As expected, Lean strategy results in lower OFR and CFR
than Fat strategy irrespective of demand characteristic. Although Lean strategy has least
inventory cost, ﬁrms adopting this strategy should consider effects of high stockouts
especially when demand variability is high.

Figure 4.16 - Effect of LOG*DEM on Average System Inventory

 

   
   

 

 

 

25,000
E’
3 20.000: '
2 ILogistics
.8 15.0001 , I-
g’ ' Push
2 ’2 —."---
8 10,000. .I a pull
2' .....
a; 5,000 ° Fat
LU -

0 ° Lean

 

Ang LVi
Ang.HVi
HD LV 1

Demand Characteristic

Figure 4.16 (above) and Figure 4.17 (next page) show that differences in average
system inventory and proﬁtability between different logistics strategies are most marked
when demand is high. Moreover, Pull logistics strategy results in slightly lower average
system inventory and proﬁtability if demand variability is high than if it is low. This
reinforces an earlier observation that Pull logistics system is more sensitive to demand

variability than other logistics strategies.

99

Figure 4.17 - Effect of LOG*DEM on Proﬁtability

 

 

 

 

 

 

3.000.000
tn
5 2,500,000.
g “" Logistics
E 2,000,0001 --
g, 0 Push
is 1,500,000.
2 c:
1: Pull
g 1,000,000. —
.E °
‘3 500,000 _fat
LU
0 _ - _ - 0 Lean
a a. a. a. z a.
9 9 ‘9,» a. E E
>
< <

Demand Characteristic

LOG*CAP signiﬁcantly affects all performance variables at p<.001. Partial n2
values range from .041 to .068, which shows that the effect on all variables is weak. As
Figure 4.18 and Figure 4.19 show, effect of capacity on OFR and CFR varies with type of
logistics strategy. When Push strategy is used, increasing capacity from 70 percent to
100 percent results in about 5 percent increment for both OFR and CFR. In addition,
under Lean strategy, OFR and CPR are higher when capacity is 100 percent than when it
is 70 percent. However, when either Pull strategy or Fat strategy is used, increasing
capacity results in negligible increases in OFR and CFR.

These results indicate that Push and Lean strategies are more sensitive to capacity
variation than Pull and Fat logistics strategies. This can be attributed to the fact that Pull
strategy is sufﬁciently ﬂexible to accommodate changes necessitated by capacity
variation while Fat strategy has sufﬁcient buffer inventory to cushion off effects of
limited capacity. Push strategy is not ﬂexible enough since replenishing and maintaining

large inventory at retail level is strongly affected by constraints at production level. Lean

100

strategy has inventory to meet demand in one order cycle. Therefore, any constraint
created by capacity limitation is likely to affect ﬁll rates more severely.

Figure 4.18 - Effect of LOG*CAP on Order Fill Rate

 

 

 

 

 

1.00

U)

C

{0

(D

12

To

C

E»

(1)

IE

.8.

E Capacity

6 '. _

L“ \a ° 100%

311""
.65 _ _ ° 70%
Push Pull Fat Lean
Logistics Strategy

Figure 4.19 - Effect of LOG*CAP on Case Fill Rate

1.00

 

.95!

.901’

 

.851 ‘i.
.801

_75 . Capacity

Estimated Marginal Means

.701 ° 100%

.65 - - ° 70%
Push Pull Fat Lean

 

 

 

Logistics Strategy

Generally, average system inventory is lower when capacity is 70 percent than
when capacity it is 100 percent for all logistics strategies (Figure 4.20). Proﬁtability is
also higher when capacity is 100 percent (Figure 4.21). However, for Pull and Fat
logistics strategies, the difference in proﬁtability between 100 percent and 70 percent

capacity levels is negligible. This implies that type of logistics strategy adopted during

101

new product launch may either exacerbate or mitigate capacity limitation problems,
which affects performance.

Figure 4.20 - Effect of LOG*CAP on Average System Inventory

 

 

 

 

 

25.000

‘2 20.0001

m

m

2

3 15.0001

9

(U

2

3 10.000:

76 C a‘

.§ 33“”

U) l

w 5000 ”100%
o _ - ° 70%
Push Pull Fat Lean

Logistics Strategy

Figure 4.21 - Effect of LOG*CAP on Proﬁtability

 

 

 

 

 

3.000.000

to 2,500,000

5

a)

E 2,000,0001

.2

5 1.5000001

3 1M] .

in 1.000.0001 Capacrty

E.

m o

L“ 500.0001 100%
0 - - ° 70%
Push Pull Fat Lean

Logistics Strategy
LAU*DEM signiﬁcantly affects OFR at p<.05 and other performance variables at
p<.001. Partial n2 values for OFR (.006) and CFR (.02) indicate that the effect is very
weak. Indeed, proﬁle plots in Figure 4.22 and Figure 4.23 indicate that differences in

OFR and CFR between national and rollout strategies at different demand levels are

negligible.

102

Figure 4.22 - Effect of LAU*DEM on Order Fill Rate

 

 

 

 

1.00
'2 .951
to
Q)
2 .901
3 ‘~ N I
.5 ~ - "
2) .85'
‘2“
.801
3 Launch
:0 . _
_§ .75 "NL
*-' a
.13 .701
.65 _ - _ _ °Roll.
a a g a. z a
Q ‘ D ’
4 9 2’ g) 9 E
< <

Demand Characteristic

Figure 4.23 - Effect of LAU*DEM on Case Fill Rate

 

 

 

 

 

1.00
8 ~95l -. '
2 .901
E
g) .851
N
5 .801
3, Launch
m I
g .75 ”Nat
Lii '70i ---. '
.65 _ _ _ °Roll.
z. a g a a a
‘ o
9 3 g, 9):: I E
< <

Demand Characteristic

Figure 4.24 and Figure 4.25 show that LAU*DEM signiﬁcantly affects average
system inventory and proﬁtability at p<.001. Partial n2 values indicate a strong effect on
average system inventory and proﬁtability (0.823 and 0.793, respectively). As expected,
low demand results in lowest average system inventory and proﬁtability while national
launch results in higher average system inventory and proﬁtability than a rollout strategy.
When demand is high, the difference in average system inventory and proﬁtability

between national and rollout launch strategies are greater than when demand is average

103

or low. Proﬁle plots indicate that demand variability has negligible, if any, effect on
average system inventory and proﬁtability.

Figure 4.24 - Effect of LAU*DEM on Average System Inventory

25.000

 

20.000l

 
  
  

15.0001

10'000‘ Launch

 

° Nat.

° Roll.

Estimated Marginal Means

 

 

HD LV 1
HD, HV

5
Q
o:
>
<

Ang. LV1

Demand Chacteristic

Figure 4.25 - Effect of LAU*DEM on Profitability

 

 

 

 

 

 

3.000.000
2 2.5000001
N
0
E 2.0000001
.2
5 1,500,0001
L h

g 1.000.0001 aunc
'3: I a
3 500.000. __-Nat'

0 _ _ _ _ ° Roll.

3 i g: i 5 ’1’

' O
9 S g to)» I E
< <

Demand Characteristic

LAU*CAP signiﬁcantly affects average system inventory (at p<.001) but does not
have signiﬁcant effect on other performance variables. Examination of marginal means
plots (Figure 4.26) indicates that the effect is negligible. Moreover, partial 112 value

(.022) suggests that the effect is weak. National launch involves inventory deployment to

104

more markets than a rollout strategy. Consequently, average system inventory is higher
for national launch than rollout launch strategy for same level of capacity.

Figure 4.26 - Effect of LAU*CAP on Average System Inventory

 

 

 

 

 

 

25.000

2 20,0001

8

2

2 15,0001

9’

(U

2 l

'8 10,0001

3 ICapacity

‘75

LU 5.0001 a 100%
0 ° 70%
National Rollout

Launch Strategy

DEM*CAP signiﬁcantly affects average system inventory and proﬁtability
(p<.001). However, the effect is stronger on average system inventory (Partial n2 = .340)
than proﬁtability (Partial n2 = .023). When demand is either low or average, both
capacity levels (100 percent and 70 percent) result in almost same levels of average
system inventory and proﬁtability. However, when demand is high, average system
inventory and proﬁtability are higher for 100 percent than a 70 percent capacity as
Figures 4.27 and 4.28 indicate. This ﬁnding makes intuitive sense. High demand makes
it difﬁcult to ﬁrlﬁll customer orders if capacity is limited. Therefore, capacity limitations
are expected to affect performance most severely when demand is high than when

demand is low.

105

Figure 4.27 - Effect of DEM*CAP on Average System Inventory

 

   
 
 

 

 

 

25,000
m
§ 20.0001
2
— ---rll
.2 15.0001 1”-
9 I
‘2“
8 10.0001 Capacity
E -___ _
g 5.000 c’100%
Lu ——-
0 _ _ - - ° 70%
a a g1 a a a
a e .. a a a
< <

Demand Characteristic

Figure 4.28 - Effect of DEM*CAP on Proﬁtability

 

 

 

3.000.000
2
3 25000001
2
E 2.000.0001
E
9
g 1,500,0001
Ca 't
g 1.000.0001 ”ac"
E D
'5; 500,000! _--1°°°/°
LU
c’70%

 

 

 

LD.LV 0
LD HV1
Ang,LV1
Ang. HV1
HD,LV1
HD,H‘.’

Demand Characteristic

4.4.3 Univariate Results: Main Effects

Table 4.7 provides univariate test results for main effects. Except for launch
strategy, all independent variables signiﬁcantly affect performance variables at p<.001.
However, Partial n2 values indicate that effect sizes vary, with demand characteristics
showing strongest effect on average system inventory and proﬁtability. Overall, logistics

strategy has strong effect on all dependent variables.

106

Table 4.6 - Univariate test for Main Effects

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source of . Degrees of . Partial Eta Observed
Variation Dependent Variable F Freedom Slg' Squared Power
Order Fill Rate 6,411 3 2,784 0.000 0.874 1.000
L 0 GI STI C Case F Ill Rate 2,705 3 2,784 0.000 0.745 1.000
Avg System Inventory 40,246 3 2,784 0.000 0.977 1.000
Proﬁtability 2,906 3 2,784 0.000 0.758 1.000
Order Fill Rate 4 1 2,784 0.043 0.001 0.525
LAUNCH Case Fill Rate 2 1 2,784 0.135 0.001 0.320
Avg System Inventory 24,939 1 2,784 0.000 0.900 1.000
Proﬁtability 68,499 1 2,784 0.000 0.961 1.000
Order Fill Rate 35 5 2,784 0.000 0.060 1.000
DEMAND Case F111 Rate 85 5 2,784 0.000 0.133 1.000
Avg System Inventory 229,197 5 2,784 0.000 0.998 1.000
Proﬁtability 176,547 5 2,784 0.000 0.997 1.000
Order Fill Rate 225 1 2,784 0.000 0.075 1.000
Case Fill Rate 354 l 2 784 0.000 0.113 1.000
CAPACITY ’
Avg System Inventory 4,004 1 2,784 0.000 0.590 1.000
Proﬁtability 267 1 2,784 0.000 0.088 1.000

 

 

 

 

 

 

 

 

 

 

Logistics strategy strongly impacts performance as demonstrated by partial n2
values that range from 0.745 to 0.977. Figure 4.29 indicates that Pull strategy results in
higher OF R and CFR than Push strategy and Fat strategy results in higher OFR and CPR
than Lean strategy. Figure 4.30 shows that average system inventory is highest when Fat
strategy is used and lowest under Lean strategy. Pull strategy results in slightly higher
inventory than Push strategy. Figure 4.31 indicates that Pull strategy results in slightly
higher proﬁtability than other strategies while Lean strategy has the lowest.

Surprisingly, although Lean strategy results in average system inventory level that
is almost half that of Fat strategy, and substantially lower ﬁll rates, it results in proﬁts
that are fairly close to those of other strategies. This can be attributed to signiﬁcant
savings in inventory cost and associated efﬁciencies. Pull strategy results in signiﬁcantly
better performance than Push strategy. For example, Pull strategy improves OFR and
CFR by 6 percent and proﬁtability by 8 percent while average system inventory increases

by 4 percent compared to a Push strategy.

107

Figure 4.29 - Effect of Logistics Strategy on Order and Case Fill Rate

 

 

 

 

 

 

 

 

 

 

 

_— W
100% l
95% 1 — a ‘ ' L """ 5 *
90% 1%" -/ \ ‘ ~
3 85% -L_- -- — —— — —-— -— - -- ~ .
E 80% 1— ————— 2 ,,___..__-_ ~——~———- -——»——-- ~ ,-
“- 75% J- _ . -7 __. _ _ _.- . 1 - . '
70%12» .2— — —- — —— 2 —--—~-—-
65% . .
Push Pull Fat Lean l
Logistics Strategy |

 

Figure 4.30 - Effect of Logistics Strategy on Average System Inventory

 

25,000

 

____ _l

i—o—Systemhv
20,000 ~-— 2 2 —_‘.__ _ _._ ...._ ._,____L, .

 

 

 

 

Push Pull Fat Lean
Logistics Strategy

 

 

 

Figure 4.31 - Effect of Logistics Strategy on Profitability

 

 

$3,000,000 . E :ij
$2 500000 1‘ — *——————._._ ___-__.2f:t*:'.°’°“‘§
$2,000,000 1.2.- _- .2 -2 ___- ___,___~_____ __ _ ,
1e 1
a; $1,500,000
D .
$1,000,000 - ,
$500,000 —»-— -— ___._ _.--. - ._.______-_ _.__ r

l i
so ., - -— — —-——-—-»——~— ————2————— —— — .1

Push Pull Fat Lean
Logistics Strategy

__+ _‘LL
>1
4

 

 

 

Although Lean logistics strategy reduces inventory cost, it results in lowest ﬁll
rates and proﬁtability. In this study, Lean strategy was modeled to mimic a situation
where deployed inventory is only sufﬁcient to cover one order cycle. Results indicate

that increasing buffer inventory at each supply chain facility may substantially improve

108

performance. For example, maintaining buffer inventory sufﬁcient to cover one order
cycle at both distribution and retail levels (Fat strategy) on average improves OFR by 26
percent, CFR by 11 percent, proﬁtability by 16 percent, and increases average system
inventory by 68 percent. However, the marked increase in average system inventory
implies that inventory carrying cost increases substantially. Therefore, the challenge is to
identify appropriate mix of buffer inventory in the system that results in optimal proﬁts
and ﬁll rates while maintaining inventory carrying cost at reasonable levels.

Launch strategy signiﬁcantly affects OFR at p<.05, average system inventory at
p<.001, and proﬁtability at p<.001 but not CFR. The effect on average system inventory
and proﬁtability is very strong as attested by high partial n2 values (0.9 and 0.961,
respectively) but effect on OFR is very weak, indeed negligible as demonstrated by low
partial 112 (0.001) and marginal means plot (Figure 4.32).

Figure 4.32 - Effect of Launch Strategy on Order Fill Rate

 

 

 

 

 

 

 

 

National Rollout
Launch Strategy

_ ._. ___.4

 

As expected, national launch strategy results in greater average system inventory
(Figure 4.33) and proﬁtability (Figure 4.34) than rollout strategy. National launch
improves proﬁtability by 45 percent over rollout launch for a 19 percent increase in

system inventory (over rollout strategy). This suggests that although there are risks such

109

as higher costs of inventory incase of product failure, national launch strategy is better
than a rollout strategy in terms of proﬁtability.

Figure 4.33 - Effect of Launch Strategy on Average System Inventory

 

25,000 _g .3

200001 — _ __ _ _ _+:S:YSt_emlnv
i

 

15,000 , — ~___._H
10,000 ~—- 4— ,. 9* Wm _ i
5.000 —2 2 _- _. %_ __*_ ___ ,.
° ' 1

National Rollout i

 

Units I

 

 

 

Launch Strategy

 

Figure 4.34 - Effect of Launch Strategy on Profitability

 

$3,000,000 ,
$2,500,000 .
$2,000,000 1

”
raj $1,500,000 . --
3 $1,000,000 -
$500,000 -2 __‘.__ _. ___-___.___ -

 

 

 

 

 

 

National Rollout i
Launch Strategy

 

 

Demand Characteristics signiﬁcantly affects all dependent variables at p<.001.
Partial n2 values indicate that demand characteristic strongly impacts average system
inventory and proﬁtability (partial n2 = 0.998 and 0.997, respectively) but has weak effect
on OFR and CFR (partial n2 = 0.06 and 0.133, respectively). Figure 4.35 shows that
demand variability affects OF R and CF R but not average system inventory (Figure 4.36)
and proﬁtability (Figure 4.37). High demand variability results in lower OFR and CFR

(decrease of about 3%) for all demand levels. There is negligible difference in OFR and

110

CFR between low, average and high demand levels, implying that demand size has
negligible effect on ﬁll rates.

Figure 4.36 and Figure 4.37 indicate that there are marked differences in average
system inventory and proﬁtability at different demand levels. Intense competitive
response leads to reduced demand, sales and proﬁts. As a result, ﬁrms adjust their
inventories in the distribution system to reﬂect reduced customer orders and sales. Thus,
lower average system inventory and proﬁts are expected when competitive response is
intense because increased competition reduces demand for the product (reduced market
share), assuming proportionate costs.

Figure 4.35 - Effect of Demand Characteristics on Order and Case Fill Rate

 

l 100% - l
95% 12 .2 - ~— —';1—-—.—_.__—1*-.v*—""——'-“l

 

 

 

‘- Q..-.

‘ 90%
85%

 

 

 

 

 

E 30% L-— -_ _._-. _- -._._-._-_- _“J—o—OFR‘H’

“' 75% 1.___+___._ «— -_“_ ##H —1',;_'_," ' _a'i

70% ‘-————— — ______, ~——-- —- -———- -— -———l

65% -, -~7 —— ~ —‘”r"——"—*—r"—~‘—‘ —.— —— -——-1——~ — —— 2 l
LD,LV LD.HV Ang,LV Ang.HV HD.LV HD.HV

Demand Characteristic i

 

When a new product has many advantages that customers like, demand grows
substantially, more sales are made, and proﬁts increase. In addition, ﬁrms expand their
inventories at retail and distribution locations so as to cope with increased demand. For
example, Figures 4.36 and 4.37 show that in high product advantage situations, average
system inventory and proﬁtability increase by 159 percent and 164 percent, respectively,
over low product advantage situations.

This suggests that ﬁrms should place greatest emphasis in communicating the

advantages of a new product to customers so as to realize increased growth.

111

Furthermore, companies introducing new products must develop a mechanism of wading
off competitive response effect. For example, ﬁrms could increase advertising and
promotions budget, ensure inventory availability, etc.

Figure 4.36 - Effect of Demand Characteristics on Average System Inventory

 

 

 

   
 
 

ﬂ 25,000
20,000 . _2. _L__ __._ ____ i
£15,000 «~— m2 _ H , i
i510.000 i— ——--> @333,in
5,000 «2 — _2 --._ 1

o

 

 

 

LD,LV LD,l-N Ang.LV Ang,HV HD.LV HD.HV
Demand Characteristic

.__.__ _J

 

Figure 4.37 - Effect of Demand Characteristics on Proﬁtability

 

 

T—$3._&50,000
$2.500,ooo1— -2. __ _. __.._..__.._-__.
$2,000,000 — - —

10!

«a $1,500,000 -

3
$1,000,000

$500,0001

so i i T
i LD,LV LD,HV Ang,LV Ang,HV HD.LV HD.HV

Demand Characteristic

 

 

 

 

 

 

 

Production capacity signiﬁcantly affects all dependent variables at p<.001. Partial
n2 values indicate that effect is greatest on average system inventory (partial n2 = 0.59).
As expected, capacity constraints limit the number of orders that can be processed and
ﬁlled, which lowers ﬁll rates, sales, and ultimately proﬁts. Consequently, as Figures
4.38, 4.39 and 4.40 indicate, OFR, CFR, average system inventory, and proﬁtability are
higher when capacity is 100 percent than when it is 70 percent. However, performance

differences between the two levels of capacity are not high, especially proﬁtability. For

112

example, order ﬁll rate decreases by less than 5 percent when capacity decreases from
100 percent to 70 percent. This suggests that expanding production capacity from 70
percent to 100 percent does not necessary yield better results.

Figure 4.38 - Effect of Production Capacity on Order and Case Fill Rate

 

 

 

 

 

 

 

 

 

 

? 100% i—o—OFRi

95% ~57 -'-__r-__._.._ - _____,__._._..:_in, t. .. -CFRi

8 90% --,__ , .k’—* 7-~—- —— ":17.“

1g 85% 7* —— — "_i— _-, i

.5 80% -——- -—-— ——— —-- - E 1

‘u. 75% ._-_. _._. ~__.____. ___ __ _ _ __.. .
700/0 .,,,__W __~, _ ._ _ _ _ _. _ _ . __,

55% 1

100% 70% '

Production Capacity !

 

Figure 4.39 - Effect of Production Capacity on Average System Inventory

 

§“7 I

 

 

 

 

 

 

25.00?
20.000 1 — 2 2 2.---» .2 4+§£teii§2
1% 15,000 — - ______,__ _______--____ l
is
10,000 .- -——- -— —§ —-- ———-:¢-—— _—
5,000 .N —. — ._ __ _ _.L._ __ L_. L ‘
o
, 100% 70%
Production Capacity

 

 

Figure 4.40 - Effect of Production Capacity on Profitability

 

$3,003,000 1
$2,500,000 12— — 222—.“ 2 22_Li___.___p.‘o_,.,2- I
$2,000,000 1l~ 2 - 2 .2 .- _ - __ - __ _-__
$1,500,000 1 _ _..-.._2_._--. .2- .._. ._L___ _. __ ___ 1
$1,000,000 t— --~» ._ —~— ——_. -._--_ ___.-._, _ __ ,___L___. 1
$500,000 .. - 2_ ._.--. -. _ 2...-.- _ ____g

$0

us}

 

 

 

 

100% 70% 1
Production Capacity

113

4.5 Hypotheses Testing
In this sub-section, study hypotheses are evaluated in light of preceding analysis
of results so as to determine and discuss their statistical support or lack of support. A

summary of hypotheses tests is provided in Table 4.5 (end of sub-section).

4.5.1 Hypotheses H1a

This hypothesis states that, “Logistics strategy directly and positively impacts new
product performance.” All tests indicate that Logistics strategy signiﬁcantly and strongly
impacts performance. Multivariate and univariate test results indicate that effects of
Logistics strategy on OFR, CFR, average system inventory, and proﬁtability are all
statistically signiﬁcant at p<.001. Moreover, evaluation of partial n2 values and marginal
means plots of Logistics strategy effect on performance variables (Figures 4.30, 4.31, and

4.32) suggests a strong effect. Therefore, Hypothesis H1a is supported.

4.5.2 Hypotheses Hlb

This hypothesis states that, “Response-based logistics strategy has a greater
positive impact on new product performance than anticipatory logistics strategy.”
Logistics strategy was examined at four levels: Push, Pull, Fat, and Lean. Past studies
stipulate that ﬁrms adopt either anticipatory or response-based logistics strategy. In this
study, Push and Fat strategies represent anticipatory strategy (characterized by sufﬁcient
deployment of inventory in the distribution channels in anticipation of sales) while Pull
and Lean represent a response-based strategy (involves deployment of inventory into the

distribution system only upon receipt of customer orders). Interaction and main effect

114

plots (Figures 4.2 — 4.21, 4.30 — 4.32) indicate that response-based logistics strategy has
greater positive impact on new product performance than anticipatory strategy. Under a
response-based strategy, ﬁll rates and proﬁts are relatively higher than in anticipatory

strategy. Therefore, Hlb is supported.

4.5.3 Hypotheses H2a

This hypothesis states that, “Launch strategy directly and positively impacts new
product performance.” Multivariate results (Table 4.3) indicate that the effect of Launch
strategy is statistically signiﬁcant at p<.001. However, univariate results (Table 4.6)
show that the effect is statistically signiﬁcant on order ﬁll rate (p<.05), average system
inventory (p<.001), and proﬁtability (p<.001) but not on case ﬁll rate. In addition, partial
n2 values show that the effect is strongest on average system inventory and proﬁtability

but very weak on OFR. Therefore, H2a is partially supported.

4.5.4 Hypotheses H2b
This hypothesis states that, “A rollout launch strategy has a greater positive

”

impact on new product performance than a national launch strategy. Examination of
marginal means plot (Figure 4.33) shows that there is no signiﬁcant difference in order
ﬁll rate between national and rollout launch strategy. Although rollout strategy results in
lower average system inventory (Figure 4.34) than a national launch, proﬁtability is
substantially lower for rollout strategy compared to a national launch. Consequently,

there is no sufﬁcient evidence to indicate that rollout strategy has greater positive impact

on performance than national launch strategy. Therefore, H2b is not supported.

115

4.5.5 Hypotheses H3

This hypothesis states that, “The interaction between logistics strategy and launch
strategy has a direct and positive impact on new product performance.” Multivariate
results (Table 4.3) show that the effect of interaction between logistics and launch
strategies is statistically signiﬁcant at p<.001. Univariate results (Table 4.5) also show
that interaction between logistics and launch strategies is statistically signiﬁcant at

p<.001 for all performance variables. Therefore, H3 is supported.

4.5.6 Hypotheses H4a

This hypothesis states that, “High demand variability negatively impacts new
product performance.” Demand variability is modeled as variation from mean demand
for all demand levels (low, average, high). According to main effect analysis, demand
variability signiﬁcantly affects OFR and CFR (Figure 4.36) but not average system
inventory and proﬁtability (Figures 4.37 and 4.38). OFR and CFR are lower when
demand variability is high than when it is low irrespective of level of demand. However,
there is no noticeable difference in average system inventory and proﬁtability between

low and high demand variability. Therefore, H4a is partially supported.

4.5.7 Hypotheses H4b

This hypothesis states that, “High competitive response negatively impacts new
product performance.” Although demand characteristics signiﬁcantly affect new product
performance as multivariate (Table 4.3) and univariate (Table 4.6) results show, the

effect of competitive response, represented by low demand level, is signiﬁcant on

116

average system inventory (Figure 4.37) and proﬁtability (Figure 4.38) but not on order
and case ﬁll rates (Figure 4.36). When demand is low (intense competitive response),
proﬁtability and average system inventory are lower than when demand is either average
or high. However, there is no noticeable change in order and case ﬁll rates. It is
expected that with decreased demand, ﬁll rates should either increase or not change much
since demand is reduced, which means less challenge of fulﬁlling it. To the extent that
proﬁtability decreases and there is no noticeable change in OFR and CPR with increased

competitive reaction, H4b is supported.

4.5.8 Hypotheses H5a

This hypothesis states that, “The greater the demand variability for a new
product, the less the impact of?

(i) Logistics strategy on new product performance.

(ii) Launch strategy on new product performance.

(iii) Logistics and launch strategies interaction on new product performance.”

This hypothesis stipulates that demand variability moderates effect of logistics
strategy, launch strategy, and their interaction on new product performance. Examination
of three-way interaction between logistics strategy, launch strategy, and demand
characteristics (Figures 4.2, 4.3, 4.4, and 4.5) indicate that demand variability moderates
the effect on OFR and CFR but not on average system inventory and proﬁtability. In
addition, proﬁle plots of two-way interactions between logistics strategy and demand
characteristics (Figures 4.14, 4.15, 4.16, and 4.17), and launch strategy and demand

characteristics (Figures 4.22, 4.23, 4.24, and 4.25) show that demand variability has a

117

signiﬁcant moderating effect for OFR and CFR but negligible effect for average system

inventory and proﬁtability. Therefore, H5a is partially supported.

4.5.9 Hypotheses H5b

This hypothesis states that, “The greater the competitor response to the new
product introduction, the less the impact of'

(i) Logistics strategy on new product performance.

(ii) Launch strategy on new product performance.

(iii) Logistics and launch strategies interaction on new product performance.”

According to this hypothesis, competitive response moderates the impact of
logistics strategy, launch strategy, and their interaction on new product performance.
Competitive response is modeled in terms of decreasing new product demand throughout
the year. Multivariate results (Table 4.3) show that the effect of demand characteristics is
statistically signiﬁcantly at p<.001. Univariate results (Table 4.6) also show that the
effect on all performance variables is statistically signiﬁcant at p<.001. However,
examination of two-way interactions between demand characteristics and logistics
strategy (Figures 4.14, 4.15, 4.16, and 4.17), and demand characteristics and launch
strategy (Figures 4.22, 4.23, 4.24, and 4.25) indicate that these interactions signiﬁcantly
affect average system inventory and proﬁtability but have negligible effect on OFR and
CFR. Furthermore, examination of three-way interaction between logistics strategy,
launch strategy, and demand characteristics (Figures 4.2, 4.3, 4.4, and 4.5) on
performance yields similar results — that interaction effect is signiﬁcant on average

system inventory and proﬁtability but not on OFR and CFR. Intense competitive reaction

118

decreases sales and proﬁts but may not necessarily affect customer service. Therefore,

H5b is partially supported.

4.5.10 Hypotheses H6

This hypothesis states that, “New product advantage directly and positively
impacts new product performance.” New product advantage was modeled in terms of
high demand (i.e. demand for the new product more than triples by the end of the year the
product is introduced). Therefore, new product advantage is represented by high demand
level under demand characteristics variable. According to marginal means plots (Figures
4.36, 4.37 and 4.38), high product advantage results in substantial increases in average
system inventory and proﬁtability. The increase in inventory and proﬁtability is expected
as demand increases. However, there is no marked increase in OFR and CFR. That
order and case ﬁll rates do not decrease despite increases in demand and challenges of
fulﬁlling it compared to average demand situations suggest that, indeed, high product

advantage directly and positively impacts performance. Therefore, H6 is supported.

4.5.11 Hypotheses H7

This hypothesis states that, “F irms’ manufacturing and distribution capacity
directly and positively impacts new product performance.” Capacity is used as surrogate
for ﬁrm size. Production capacity is modeled at two levels, 100 percent and 70 percent,
representing large and small/medium sized ﬁrms, respectively. Distribution capacity was
not modeled. Modeling distribution capacity would have made the model design more

complicated while adding little value to the study since production capacity captured the

119

size factor. Multivariate results (Table 4.3) indicate that effect of capacity is statistically
signiﬁcant at p<.001. In addition, univariate results (Table 4.6) show that the effect is
signiﬁcant at p<.001 for all performance variables. According to marginal means plots,

new product performance is better when capacity is 100 percent than when it is 70

percent. Therefore, H7 is supported.

Table 4.7: Summary of Hypotheses Tests

 

 

 

 

 

 

 

 

 

 

 

 

 

impacts new product performance.

HYPOTHESES RESULTS
H1a: Logistics strategy directly and positively impacts new product performance. Supported
Hlb: Response-based logistics strategy has a greater positive impact on new Su o rt ed
product performance than anticipatory logistics strategy. pp
H2a: Launch strategy directly and positively impacts new product performance. Partially Supported
H2b: A rollout launch strategy has a greater positive impact on new product
performance than a national launch strategy. Not supported
H3: The interaction between logistics strategy and launch strategy has a direct
. . . Supported
and posrtive impact on new product performance.
H4a: High demand variability negatively impacts new product performance. Partially Supported
H4b: High competitive reaction negatively impacts new product performance. Supported
H5a: The greater the demand variability for a new product, the less the impact of:
(i) Logistics strategy on new product performance. Partially Supported
(ii) Launch strategy on new product performance. Partially Supported
(iii) Logistics and launch strategy interaction on new product performance. Partially Supported
H5b: The greater the competitor response to the new product introduction, the
less the impact of:
(i) Logistics strategy on new product performance. Partially Supported
(ii) Launch strategy on new product performance. Partially Supported
(iii) Logistics and launch strategy interaction on new product performance. Partially Supported
H6: New product advantage directly and posrtively impacts new product Supported
performance.
H7: Firms manufactunng and distribution capacrty directly and posrtively Supported

 

120

 

4.6 Discussion of Results

The main objectives of this research are to examine how logistics and launch
strategies affect new product performance, to examine the interaction between the two
strategies during new product introduction, and to identify the combination of logistics
and launch strategies that results in optimal new product performance given prevailing
competitive and demand situations. To realize these objectives, the research examines
effects of logistics strategy, launch strategy, their interaction, demand variability,
competitive response, product advantage, and production capacity on new product
performance. Analysis of results suggests several ﬁndings.

First, logistics strategy strongly impacts new product performance. Two sets of
logistics strategy alternatives were examined. These represented anticipatory and
response-based logistics strategies. First, results show that logistics strategy has the
strongest effect overall of all independent variables. Partial n2 values for effect of
logistics strategy on all dependent variables are greater than 0.745, indicating large effect
size. None of the other independent variables has high partial n2 values for all dependent
values tested. Results also show that interaction between logistics and other independent
variables such as launch strategy, demand, and capacity signiﬁcantly affects new product
performance. Therefore, it can be argued that logistics strategy is the most critical
determinant of new product success among the tested variables, which underscores
stipulations in extant literature that logistics is a key determinant of new product
performance.

Results indicate that Pull strategy outperforms Push strategy. A comparison of

Push and Pull strategy indicates that Pull strategy results in better customer service

121

(higher ﬁll rates) and proﬁtability than Push strategy for comparable average system
inventory. This implies that although forward deployment of inventory may reduce
stockout problem, it results in higher inventory carrying cost and shortages when there is
capacity constraints. Moreover, Pull strategy provides greater ﬂexibility in inventory
deployment. Fat strategy results in better customer service (higher ﬁll rates) than Lean
strategy. Average system inventory under Fat strategy is 68 percent higher than under
Lean strategy yet proﬁtability for Fat strategy is only 16 percent higher. This implies that
under Fat strategy, ﬁrms incur high costs of inventory which signiﬁcantly erodes proﬁt
gains. Under a Lean strategy, inventory is kept to a minimum. This results in lower
costs. However, stockouts are more frequent. Consequently, although Fat strategy
provides better customer service, it has high risks and opportunity costs that make Lean
strategy more suitable for new product introduction. Generally, maintaining sufﬁcient
inventory to cover one order cycle or slightly higher provides many performance beneﬁts
than maintaining large safety stock levels. Therefore, ﬁrms can realize better new
product performance if they adopted a Lean strategy coupled with 8 Pull strategy, than a
Fat strategy coupled with Push strategy.

Second, launch strategy impacts new product performance. It impacts average
system inventory and proﬁtability but not customer service. Results indicate that national
launch outperforms rollout launch in terms of proﬁtability. However, national launch
demands more inventory deployment and greater coordination between supply chain
partners since it covers more market segments. As a result, it is associated with higher
average system inventory and more sales. If product succeeds, national launch provides

opportunities for a ﬁrm to build strong market presence and realize greater sales and

122

proﬁts. However, if product fails, ﬁrms adopting this strategy incur high cost of
inventory including storage, handling, write-off, and obsolescence. In other words,
national launch exposes ﬁrms to greater risks than rollout strategy but opportunity loss in
terms of lost sales if product succeeds is quite high. Consequently, whenever possible,
ﬁrms should consider launching products nationally especially for high advantage
products. National launch strategy is more appropriate when demand is fairly stable,
product advantage is high, and competitive response likely to be slow.

Third, interaction between logistics and launch strategies impact new product
performance. Results indicate that under Push strategy, national and rollout strategies
result in almost the same level of customer service. However, under 3 Pull strategy,
national launch outperforms rollout strategy in terms of customer service and
proﬁtability. There are no differences in level of customer service between national and
rollout strategies for both Fat and Lean strategies. However, Fat strategy had
signiﬁcantly higher average system inventory and proﬁts under national than rollout
strategy. Similarly, Lean strategy results in higher average system inventory and
proﬁtability under national launch than under rollout launch. This ﬁnding suggests that
when launching new product, ﬁrms must consider the simultaneous impact of logistics
and launch strategies.

Generally, national launch supported by anticipatory logistics strategy provides
many opportunities to leverage economies of scale in production, transportation, and
warehousing. These cannot be realized under rollout strategy supported by response-
based logistics. However, there are high risks such as inventory obsolescence and write-

off in case product fails when national launch supported by anticipatory logistics is

123

adopted. National launch supported by response-based logistics strategy seems to yield
better performance in terms of proﬁtability. Therefore, when launching new products,
companies should consider the simultaneous impact of logistics and launch strategies,
and adopt a combination of strategies that results in optimal performance.

Fourth, demand characteristics have strong impact on performance. In this study,
demand variability, competitive response, and product advantage are considered. Results
indicate that demand variability affects OFR and CFR but not average system inventory
and proﬁtability. High demand variability results in lower customer service. The effect
is greater when strategies that results in lower service levels (lower OFR and CFR) are
used. For example, the effect is greater under a Push strategy than a Pull strategy, and
under a Lean strategy that a Fat strategy. This is because when demand is highly
variable, ﬁrms’ ability to plan for and deploy inventory to fulﬁll customer orders in a
timely manner is greatly reduced. However, when demand variability is low, ﬁrms can
plan in advance for expected orders and deploy inventory accordingly. At the same time,
demand variability has greater effect on service when rollout launch is adopted than when
national launch is adopted. Demand variability has signiﬁcant interaction with capacity.
Results indicate that high demand variability under capacity constraints results in poorest
new product performance. When launching new product launch, it is important that ﬁrms
evaluate demand uncertainty and adopt strategies that minimize its negative effects.

Results indicate that when competitive response is intense thereby reducing new
product demand, ﬁrms lose sales and realize lower proﬁts. Moreover, if a ﬁrm deploys
sufﬁcient inventory in the distribution system in anticipation of sales, and intense

competitive reaction causes demand for the new product to shrink signiﬁcantly, then the

124

ﬁrm is left with large inventory cuhninating in high inventory carrying and write-off
costs. Therefore, competitive response to new product introduction must be addressed
and measures taken to counter competitor actions at the earliest. Firms must aggressively
counter competitor actions through measures such as heavy investment in marketing and
promotion as well as efﬁcient logistics operations to ensure product availability when
customers demand it.

When demand is high due to high product advantages, firms realize signiﬁcantly
higher proﬁts irrespective of logistics and launch strategies used. High product advantage
is associated with more customer orders, more sales, and ultimately more proﬁt as new
product gains greater acceptance. Firms can gain from economies of scale in production,
transportation, and warehousing given the large and growing demand associated with
high product advantage. Therefore, anticipatory and national launch strategies seem
appropriate since they provide better opportunities to leverage economies of scale. It is
also important for ﬁrms to consider simultaneous effect of other factors that interact with
demand when launching new product that has high product advantages.

Finally, production capacity directly impacts new product performance and
interacts signiﬁcantly with logistics strategy. For example, increasing production
capacity results in higher customer service when Lean strategy is used but has negligible
effect under Fat logistics strategy. However, interaction between capacity and launch
strategy has negligible effect on new product performance. Large capacity provides for
greater inventory and proﬁts than smaller capacity irrespective of launch strategy used.
Capacity constraints limit the number of orders that can be processed and ﬁlled, which

lowers customer service and proﬁts. Therefore, logistics strategies characterized by

125

greater stock-outs are affected more by capacity limitations than those characterized by
lower stock-outs. Fat strategy is characterized by high inventory in the distribution
system. Such inventory acts as a buffer when orders cannot be processed promptly.
Indeed, companies often use inventory to mitigate capacity problems. Therefore, when
launching new product, it is important that ﬁrms consider implications and possibilities
of expanding capacity as demand grows.

Research ﬁndings are summarized in Tables 4.8 and 4.9 below. Implications of
these ﬁndings for managers and researchers are discussed in the next chapter.

Table 4.8 — Summary of Results Matrix 1

 

 

 

 

Logistics Strategy
Anticipatory Response-based
- High inventory carrying cost - Lower inventory carrying cost.
» High sales, proﬁts, and market share if - Expensive due to increased orders and
product succeeds. small, frequent shipments (inefﬁcient).
L High costs due to overstock, write-off, P Insufﬁcient capacity to support
National Launch etc, if product fails. ﬂexible order fulﬁllment in multiple
Strategy - Economies of scale in production, markets.
transportation, and warehousing. - Stockouts in many markets
L Appropriate for high product advantage - Lower service levels.
situations. Better in high competitive response
High risk when demand variability is situations.
high.
- High inventory carrying cost. - More targeted.
- Lost sales if product succeeds. - Higher service levels.
Rollou t Launch - High inventory cost. if product fails. - Lower cost of implementation.
High risk of competitors capturing the . .
Strategy - High transportation cost due to small,
markets not served. .
frequent shipments.
Appropriate when demand is highly
variable.

 

 

 

 

 

126

Table 4.9 — Summary of Results Matrix 2

 

Logistics Strategy

Launch Strategy

 

Anticipatory

Response-based

National Launch

Rollout Launch

 

- High cost due to
excess inventory.

- Some economies

. Lower inventory.

» Lower inventory
costs — lower risks

r Lower service
level than rollout.

- Higher inventory

~ Higher customer
service levels.

. Lower inventory,

 

 

 

 

of scale in of write-off, cost. lower cost.
transportation, obsolescence, etc. . .
High warehousing etc. _ ~ Enables countering - More appropriate
Competitive ’ - More appropriate of competitors in to reduce risk
Response - Greater market tltirar: anticipatory all markets. inelase product
presence. 5 a egy. _ Greater market a1 5.

r High risk of presence. - High risk of
inventory, i.e. competitors
obsolescence, capturing markets
write-off, etc. not served.

1 High service levels - Higher stockouts — - Higher service -Lower service
due to higher greater challenges levels. levels — no

. safety stoclc (12:12:33,011 as P Higher challenging :szlréomies of
Hg” Demand - Higher costs. increases to implement as '
Variability ' variability -Lower proﬁts

increases.
- Higher costs.

- Higher inventory More stockouts . Higher customer - Lower service
costs. and lower service service than rollout levels.

- More sales, market levels. strategy. - Lower inventory,
share, proﬁts. . Higher proﬁts with - Higher inventory. causing more

. lower costs. . stockouts.
H' - Gains from Econormes of scale
igh Product . . . i . . . .
Advantage econonues of scale - Higher risk of lost gains in ordering, - Higher lost sales.
(ordering, sales. transportation,
production, ’ Expensive to warehousrng, etc.
transportation, . .
. implement, lacks - More appropriate.
warehousmg). .
economies of scale
beneﬁts.
- Allows for - Lower service ~ Higher inventory - Lower inventory
. production levels cost. cost.
Production economies.
Capacity - Allows for greater - Lower proﬁt.

 

 

economies of scale.

 

 

 

127

 

CHAPTER 5: IMPLICATIONS AND CONCLUSION
5.0 Introduction
This chapter presents implications of the research ﬁndings. The ﬁrst section
discusses managerial implications. This is followed by research contribution. Then a
discussion of study limitations and directions for future research is presented. Finally, a

conclusion summarizing main issues in the Chapter is presented.

5.1 Managerial Implications

The research ﬁndings have several managerial implications. First, when planning
and launching a new product, it is important for managers to engage logistics personnel
as early as possible in the process. Study results indicate that logistics strategy strongly
impacts all performance variables, which means logistics processes are a critical
determinant of new product performance [Refer to p.105-106]. However, in some
companies, logistics personnel are not involved in new product development process until
late in the process (Tracy 2004). In other cases, as was observed during interviews,
personnel with limited expertise in logistics such as marketing and operations managers
are asked to develop, plan, and execute logistics activities during new product launch.
These managers lack the expertise required to develop requisite logistics plans. Engaging
logistics personnel early in the new product launch helps the company to identify
logistics resource requirements in advance and make appropriate plans to acquire them or
allocate them. In addition, logistics professionals possess expertise and perspective

necessary to provide sound input regarding decision trade-offs during launch.

128

Companies can also employ logistics service provides to help them with planning
logistics operations during launch.

Secondly, managers should evaluate logistics strategies and adopt a strategy that
results in best performance in different market and competitive situations. Results
indicate that response-based strategy is better than anticipatory strategy [Refer to p.107-
108]. For example, Pull strategy, which is response-driven, outperforms Push strategy,
which is anticipatory-driven in terms of proﬁtability. At the same time, although Fat
strategy results in better customer service than Lean launch, it is associated with higher
inventory carrying cost and associated management and handling costs. Indeed, Lean
strategy has greater proﬁts per unit than Fat strategy, suggesting that if slightly higher
inventory than that tested in the model were kept as safety stock in Lean strategy, it
would yield better results than Fat strategy. Moreover, response-based strategy not only
lowers inventory carrying cost but also reduces risks associated with new product failure.
The strategy encourages better response to customer orders while maintaining low
average system inventory. Consequently, managers should identify their companies’
primary goals and seek the strategy that best allows realization of those goals. For
example, if primary goal is proﬁtability, a response-based strategy is preferred while if
superior customer service is the main goal, then an anticipatory strategy is recommended.

Third, logistics strategy interacts strongly with other independent variables. For
example, a National launch supported by Fat logistics strategy results in higher average
system inventory than a rollout launch supported by F at logistics strategy. In addition, a
national launch supported by Pull logistics strategy results in higher proﬁtability than a

national launch supported by other logistics strategies [Refer to p.96-97]. This means

129

that managers should take into consideration the simultaneous impact of logistics
strategy, launch strategy, and other independent factors when planning and implementing
a new product introduction.

Where competitive response is intense, managers should adopt a rollout launch
supported by response-based logistics strategy. This allows the company to respond
more appropriately to customer needs and provides necessary ﬂexibility to adjust
inventory deployment as market demands. However, when product advantage is high
and demand is growing rapidly, a national launch supported by response-based logistics
strategy is recommended since it yields higher proﬁtability. Yet, since risks associated
with anticipatory strategy such as obsolescence and write-off costs in case of product
failure are minimized when demand is growing rapidly, an anticipatory logistics strategy
may be considered. The strategy allows ﬁrms to leverage economies of scale in
production and distribution, which results in reduced cost and high proﬁts.

Fourth, managers should carefully consider the costs and risks associated with
national launch strategy against rollout launch strategy. Launch strategy does not
signiﬁcantly affect customer service as measured by order and case ﬁll rate [Refer to
Figure 4.32, p.109]. Thus, considerations regarding which launch strategy to adopt
should focus on the average system inventory and associated costs, and proﬁtability
implications [Refer to p.109-110]. National launch requires greater inventory
commitment and resources than rollout strategy. If new product is successful, national
launch provides an opportunity for ﬁrms to establish early market dominance, realize
greater sales and proﬁts, and recoup new product development cost sooner. However, if

product fails, national launch exposes the ﬁrm to higher costs of handling excess

130

inventory that is not sold, write-off costs, transshipment cost where necessary, and high
capital costs for inventory held in the distribution system. In such cases, rollout strategy
is preferable.

However, it is important to note that the type of launch strategy to adopt may be
dictated by supply chain partners. For example, major retailers with national presence
would prefer stocking a new product in all markets. Such retailers may inﬂuence the
launch strategy adopted. Furthermore, retail shelve allocation is done periodically. For a
ﬁrm launching new product, it would mean whenever possible, tinting of launch would
be made to coincide with the period when shelve spaces are allocated at retail levels.
This may encourage ﬁrms to launch nationally rather than through sequential rollout.
Whichever the case, national launch offers more opportunities than rollout strategy and is
recommended.

Fifth, demand characteristic signiﬁcantly impacts new product performance and
moderates the effect of logistics strategy and launch strategy on performance. Demand
variability affects customer service but does not affect average system inventory and
proﬁtability. Generally, customer service is lower when demand variability is high and
higher when demand variability is low [Refer to Figure 4.35, p.111]. This means that
managers must seek ways to mitigate demand variability problem when launching new
product. Study results indicate that response-based logistics strategy is more sensitive to
demand variability than anticipatory strategy [Refer to p.97—98]. Anticipatory logistics is
characterized by greater safety stock in the distribution system, which means that ﬁrms
can utilize this stock to hedge against shortages occasioned by demand variability.

Response-based logistics strategy does not provide this option. Instead, ﬁrms rely on

131

system ﬂexibility to respond to varying customer orders. Therefore, to deal with demand
variability, managers should evaluate logistics strategies in the context of other factors
such as launch strategy, competitive environment, etc., so as to adopt the strategy that,
collectively, minimizes negative effects of demand variability.

Sixth, managers should develop mechanisms to anticipate and appropriately
counter competitor actions when launching new product. Competitive response has a
negative effect on new product performance [Refer to p.111-112]. As study results
indicate, reduced demand due to competitor response erodes proﬁts and undermines
product success. Thus, managers could increase advertising and promotions in markets
where competitive response is intense. In addition, managers may consider deploying
more inventories in markets facing more intense competition than those with less intense
competitor response. This ensures that customers readily access the new product.

Seventh, high product advantage provides opportunities for greater sales and
ultimately proﬁts. The main challenge when product advantage is high is to cope with
increasing demand. A response-based logistics strategy results in higher proﬁtability
while an anticipatory strategy results in higher average system inventory when product
advantage is high [Refer to p.99-100]. Although response-based strategy results in higher
proﬁtability, managers should consider anticipatory strategy by determining the cost
savings from production and distribution economies against inventory costs. To the
extent that savings are high, national launch supported by anticipatory strategy may be
adopted too. Given that demand is increasing substantially, national launch is most

appropriate in high advantage situations.

132

Eighth, results show that product advantage has greater impact on demand and
product performance than competitive response in terms of proﬁtability [Refer to Figure
4.37, p.112]. For example, increase in proﬁtability when product advantage is high is
greater than relative decrease in proﬁtability when competitive response is high. Thus,
ﬁrms gain more by promoting new product advantages than they would if they invested
resources toward countering competitive response. In other words, managers should pay
greater attention to and channel more resources toward mechanisms of communicating
and promoting product advantages even as they counter competitive actions.

Finally, although large capacity enables a ﬁrm to fulﬁll its demand requirements,
study results indicate that the difference in performance between 100 percent and 70
percent capacity is minimal [Refer to Figure 4.40, p.113]. For example, average proﬁts
decrease when capacity decreases from 100 percent to 70 percent for both national and
rollout strategies. This shows that capacity expansion in case of product success should
be done cautiously since it does not improve performance substantially. Production
capacity is used as a surrogate for ﬁrm size. Generally, a company with more resources
is expected to allocate larger budget to new product launch. However, such investment
will not be meaningful if channeled towards capacity expansion. Instead, managers
should channel resources toward activities that raise demand or counter competitor
actions such as promotions.

Table 5.1 presents a matrix that summarizes the best logistics strategy, launch
strategy, and combination of logistics and launch strategy for each demand characteristic,

competitive response, product advantage, and production capacity situations. The matrix

133

provides a snap shot view of strategies that managers should pay more attention to when
launching a new product.

Table 5.1 - Best Strategy Matrix

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Logistics Launch Interaction: Logistics Strategy x Launch
Snatfgy Strategy Strategy
Level A nric. Resp. Nat. Roll. Antic. *Nat. Antic. "Roll. Resp. *Nat. Resp. ‘Roll.
Demand LOW ‘ * ‘
Variability High a a Ill
Competitive LOW * * * *
Response High :1- all in
Product LOW * it *
Advantage High 1.. s s. 1..
Production 100% t * 1' I ‘
Capacity 70% it all :1-
* Represents best strategy Antic. = Anticipatory Roll. = Rollout
Resp. = Response-based Nat. = National

As Table 5.1 shows, when demand variability is low, response-based logistics
strategy and national launch strategy should be adopted. However, when demand
variability is high, anticipatory logistics strategy and rollout launch strategy should be
adopted. Therefore, in high demand variability situations, managers should adopt a
rollout strategy supported by anticipatory logistics while in low demand variability
situations national launch supported by response-based strategy should be adopted.

When competitive response is high, rollout launch supported by response-based
logistics is most suitable. A rollout strategy allows ﬁrms to identify and target markets
with greater potential and less competitive pressure while Response-based logistics
provides ﬂexibility necessary to counter competitor action or respond appropriately to
demand with greater efﬁciency. However, when competitive response is low, national

launch supported either by anticipatory or response-based logistics strategy is preferable.

134

When product advantage is high, a national launch supported by response-based
strategy is recommended. A Response-based strategy results in higher proﬁtability than
Anticipatory strategy. It also results in higher customer service (i.e., ﬁll rates) when
demand variability is low. However, anticipatory logistics may offer cost savings
through economies of scale, which may translate to lower cost especially if new product
turnover is high. Thus, national launch supported by anticipatory logistics strategy may
be considered as a viable alternative in high product advantage situations. When product
advantage is low, rollout strategy supported by response-based logistics strategy is better.

Production capacity is a surrogate for ﬁrm size. Response-based strategy is
preferable when there are capacity limitations. Moreover, a national launch requires
larger capacity (and resource endowment) than a rollout launch, which a large ﬁrms can
sustain more than a small/medium size ﬁrm.

The preceding results and discussions lead to managerial decision options shown
in Table 5.2 and Table 5.3.

Table 5.2 - Strategy Analysis Chart 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Decision Product Competitive Demand Firm Best S tra t
Determinant Advantage Response Variability Resources egy
High gaggle]
High National x Response-based
Large
Low
H' Small
igh L
High arge National x Anticipatory
Small
Low Lar e
Low 8 National x Response-based
. . Small
Profitability
. Large
High S 11
High ma Rollout x Response—based
Large
Low
Small
Low
. Large . .
High Rollout x Antrcrpatory
Small
Low Lar e
Low 5 Rollout x Response-based
Small

 

 

 

 

 

 

 

 

135

Table 5.2 shows that product advantage followed by competitive response is the
leading factor with regard to proﬁtability. Firm resources (capacity) have negligible
effect on proﬁtability. Demand variability has a weak effect on proﬁtability. Therefore,
where proﬁtability is the main consideration, managers should focus on product
advantage and to a lesser extent, competitive response. When product advantage is high,
national launch is preferred but when it is low, rollout launch is preferred. Anticipatory
strategy performs better than a response-based strategy when demand variability is high.
Therefore, in high demand variability situations, anticipatory strategy may be considered.

Table 5.3 - Strategy Analysis Chart 2

 

Decision Demand Firm Product Competitive

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Determinant Variability Resources Advantage Response Best Strat g3
High HIEL
Low . . .
Large . National x AntiCipatory
High
Low
. Low
High .
. High
High Low
Small . Rollout x Anticipatory
High
Low
. Low
Servrce H' h
High L‘i
High ow National x Response-based
High
Low
Low
Low H' h
High [.2511
Low . Rollout x Response-based
High
Low
Low

 

 

 

 

 

 

 

 

Table 5.3 shows that when primary objective is to provide high customer service
(i.e., high ﬁll rates), managers should pay more attention to demand variability and ﬁrm’s
resource endowment. Product advantage and competitive response have negligible
impact on ﬁll rates. Results indicate that anticipatory logistics strategy results in better
performance than response-based strategy when demand variability is high. Therefore,

when demand variability is high and a ﬁrm has large capacity or resource endowment,

136

national launch supported by anticipatory strategy is recommended. However, when

demand variability is low, response-based logistics strategy is preferable.

5.2 Research Contribution

This study makes several research contributions. First, whereas past studies have
stipulated most propositions tested in this study, none of them is empirically tested.
Therefore, the study makes an important contribution by empirically testing individual
and simultaneous effects of logistics and launch strategies on new product performance.
In addition, the study examines interaction between these strategies and demand,
competitive, product, and ﬁrm characteristics. These are issues previously discussed but
never empirically tested in a single study.

Second, the study examines simultaneous effect of different determinants of new
product performance. Past studies tend to examine these determinants in isolation.
However, these factors impact performance both individually and interactively.
Examining interaction effects provides a sound basis for managerial decision making and
an important foundation for future research.

Third, simulation research has not been widely applied in new product research.
Yet, simulation offers many beneﬁts for in depth analysis of new product development
and launch situations that are difﬁcult to analyze using other techniques. Indeed, most
studies on new product launch have tended to be qualitative. This study demonstrates
that applying simulation approach can greatly enhance our understanding of major factors
and relationships in new product launch. Future research can build on the study ﬁndings

and methodology to advance new product launch knowledge.

137

Finally, this study makes suggestions that managers will ﬁnd useful. These
suggestions are based on thorough and sound tests of main and interaction effects. Study
assumptions are based on interviews with managers and careful evaluation of extant

literature. Thus, the ﬁndings have a direct application in managerial decision making.

5.3 Limitations and Opportunities for Future Research

First, in simulation research, researchers make certain assumptions to limit the
scope of analysis so as to make detailed investigation of phenomena in question. For
example, in this study, assumptions are made regarding inventory policies, transit lead
times, and stochastic demand distribution. However, these assumptions limit the scope of
issues and situations examined, and resulting ﬁndings may be speciﬁc to certain scenario.
Although assumptions are based on interviews with managers and extant literature, they
may be too limiting. For example, the assumption that demand distribution and transit
lead times follow a triangular distribution pattern may not apply in some industries and
product launch situations. Therefore, future research should examine different industries
and demand distributions.

Second, this study modeled a simple supply chain network consisting of one
manufacturer, two distributors, and ten retailers dealing with one product. This may not
capture major dynamics at play during new product launch especially for large companies
with multiple networks of suppliers, distributors, and products. A more comprehensive
supply chain network that includes suppliers and more products should be considered in

future research.

138

Third, production was treated as a “black box” where incoming orders are ﬁlled
without regard to production factors such as scheduling, set-up/change-over times, etc.
Complexities in production process might provide additional insights regarding new
product launch dynamics. For example, variations in production lead time, raw material
availability, and process efﬁciency might affect distribution effectiveness for a new
product. It might also have different effects if there is product variation. Thus, future
research should consider production parameters in detail.

Fourth, this study is based on consumer goods industry. However, individual
industries tend to have different dynamics that may not reﬂect events in other industries.
For example, lead times and demand patterns in consumer goods industry may be
different from those in high-tech sector. Future research should investigate other
industries too.

Fifth, retail and market segments were modeled as portraying similar
characteristics. For example, markets segments were modeled as similar in size with
similar demand distributions. However, markets can have different characteristics, which
may require companies to institute strategies tailored to individual markets. Future
research may consider testing variations in market segments and demand characteristics.

Sixth product value density may affect distances to which a new product is
deployed or the market coverage for a new product. It would be interesting to see how
launching new products that have low value density compares with high value density
products from a logistics viewpoint. Future research should examine value density.

Finally, this study considers four performance variables. However, other

performance variables such as cycle time, market share gains, and customer loyalty may

139

provide more insights and show different results. Therefore, future research should

consider other ways of measuring new product performance.

5.4 Conclusion

The preceding discussion pinpoints key managerial implications, research
contributions, and opportunities for future research. This study empirically tested and
conﬁrmed some propositions stipulated in extant literature. It also highlighted important
ﬁndings that managers and researchers will ﬁnd useful. The study makes important
contributions in advancing new product launch knowledge, especially as far as logistics
processes and strategies are concerned. Managers should consider the managerial

suggestions while future research should build on these ﬁndings.

140

APPENDICES

141

APPENDIX A: QUESTIONNAIRE

MICHIGAN STATE UNIVERSITY
Department of Marketing & Supply Chain Management

INTERVIEW QUESTIONS

These questions are designed to provide a general idea about your company’s new
product launch processes and about a recent launch activity. You need not do speciﬁc
research on any of the questions prior to the interview.

1.

General Questions
a) At what level do you manage new product launch in your company or SBU?
b) Could you walk me through your company’s new product launch process?
c) Do you have a launch process diagram? If so, could you provide me with a
copy?
(1) How do you manage capacity (manufacturing/distribution) during launch?

Think about the most recent new product launch in your company or SBU:
a) Product characteristics
i. What were the product’s main advantages over competitor brands?
ii. What did you expect customers to think of your product?
iii. What was the nature of demand for the product and brand?

b) Launch process
i. What launch strategy did you adopt/apply? Why?
ii. What was the scale of entry (size, in how many markets)? Why?
iii. What inﬂuenced your entry timing? Why?
iv. Did you face any capacity problems?

c) Logistics processes
i. What logistics strategy did you adopt? Why?
ii. What was the nature of your distribution channel/supply chain
network?

d) Market characteristics
i. Generally, what was the nature of competition in the industry?
ii. How did competitors react to your newly launched product?
iii. How predictable is customer demand in the industry?

e) Performance
i. Generally, how do you measure success/failure?
ii. Was the launch a success or failure? Why?

142

APPENDIX B: COVER LETTER

[Date ..................... ]
[Dr./Mr./Ms. .................. ]
[Address ................... ]
[Dear ....................... ]

I am a doctoral candidate in the Department of Marketing and Supply Chain Management
at Michigan State University. I am working on my dissertation, in which I investigate
logistics processes during new product launch. Speciﬁcally, I examine how the
integration of launch and logistics strategies under various competitive and demand
situations impacts ﬁrms’ new product performance. I would like to explore this question
at two different ﬁrms, one in durable and another in non-durable consumer goods. I
selected your ﬁrm because it not only deals with durable/non-durable consumer goods,
but also ﬁrm has a long history of successfully introducing new products to the market.

In order to complete my dissertation, I need your assistance. I would like to interview
you at your company location. I have attached a brief summary of my research that
describes what I am investigating and why this may be beneﬁcial to you and your ﬁrm.
Also attached are the questions I intend to use to guide the interview, which will last
about one hour. As a VP. /director /manager /supervison you have considerable
experience and knowledge about your company’s product launch processes. Therefore,
your input will be most helpﬁil.

I will maintain your anonymity and will keep your responses completely conﬁdential. I
will also be happy to send you a research results summary report.

If you agree to participate, I would like to schedule an appointment with you at your most
convenient time. Please contact me by email: nvaga@bus.msu.edu or by phone: (517)
353-6381.

I know your time is valuable. I greatly appreciate your consideration and look forward to
hearing from you soon.

Sincerely,

Gilbert N. Nyaga

PhD Candidate

Department of Marketing and Supply Chain Management
The Eli Broad College of Business

Michigan State University

N3 70 North Business College Complex

East Lansing, MI 48824-1122

Phone: (517) 353-6381

Fax: (517) 432-1112

E-mail: nvaaa@bus.msu.edu

143

APPENDIX C: RESEARCH IN BRIEF

Title: An Investigation of the Simultaneous Impact of Launch and Logistics Strategies
on New Product Performance

What is the research about?

This research investigates how ﬁrms’ launch and logistics strategies impact new product
performance under different competitive and customer demand environments. New
product introduction is the most expensive, risky and time-consuming phase of new
product development process. Consequently, it signiﬁcantly impacts new product
success. To reduce the costs incurred during introduction, companies prioritize inventory
deployment and adopt ﬂexible logistics systems to mitigate demand uncertainty risks.
However, the impact of such an approach on product and ﬁrm performance is not fully
understood. Moreover, its effectiveness under various competitive and customer demand
environments is not known. Thus, two important questions arise from this: what launch
and logistics strategies result in minimal product introduction costs and optimal new
product performance under different competitive and demand conditions? What
combinations of launch and logistics strategies result in best new product performance?

How is the research conducted?

I will conduct the research in two phases. First, I will interview managers at two
companies in durable and non-durable consumer goods industries. The select companies
should have a long history in developing and introducing new products to market. The
purpose of the manager interviews is to understand new product launch processes and
dynamics. I will then develop a product launch process diagram and supply chain
network that mirrors these companies’ launch processes. Second, based on the diagram
and information about competition and customer demand from the interviews, I will
design a simulation model to generate data that I will then use to empirically test the
research questions. Data from interviews are important because they provide real-system
baseline for the simulation model, thereby making it more realistic.

How does it beneﬁt you and your company?

First, I will provide a summary report of research ﬁndings. In addition, I will be happy to
present the results verbally to you/your company if you wish. Secondly, I will make
recommendations regarding how companies can leverage logistics functions to make new
product introductions more successful. Furthermore, I will prepare an easy to apply
framework for identifying appropriate launch and logistics strategies to adopt in different
competitive and demand conditions. Such a framework will be useful to you as decision-
maker and to your company’s competitiveness. Thirdly, since I will base the simulation
model on the information you provide, I expect the results to be closely applicable to
your company. Thus, the ﬁndings could have immediate beneﬁts to you and your
company’s operations.

What do I need from you?

I request to interview at least three managers in your company, representing senior level,
middle level, and lower level management. The managers should be responsible for new
product development, product introduction, or logistics/distribution operations. I have
enclosed the interview questions. I will keep your responses conﬁdential and will
maintain your anonymity.

144

APPENDIX D: CONSENT FORM

MICHIGAN STATE UNIVERSITY
Department of Marketing & Supply Chain Management

INTERVIEW CONSENT FORM
Dear Sir/Madam,

You have been selected to participate in this research because we believe you have
experience and knowledge about new product launch processes in your company that will
be insightﬁrl to the study. The research investigates the simultaneous impact of logistics
strategies, launch strategies, competition, and customer demand on new product
performance. First, case studies of select companies are done to enable the researchers
understand how new product launch is conducted in real-world. Second, a simulation
model based on case studies ﬁndings will be developed to generate data that will be used
to examine and test the performance outcomes.

Your company has been selected for the case study purpose because it has along history
of developing and launching new products. Your knowledge of launch processes in your
ﬁrm will be helpful in designing a representative simulation model. Participation in the
study will require you to answer open ended questions about a recent new product launch
in your company. Your responses will be kept conﬁdential and will not be linked to you
personally.

The interview will take approximately one hour. You can refuse to answer any question
or stop the interview at any time. No audio-recorder will be used. Participation in the
study will not result in any negative consequences for you. Your privacy will be
protected to the maximum extent allowable by law.

If you have questions about the study, please contact: Dr. David J. Closs, The John H.
McConnell Chair in Business Administration, N370 Business College Complex, East
Lansing, MI 48824; phone: (517) 355-6381; email: closs@bus.msu.edu.

In case you have questions or concerns regarding your rights as a study participant, or are
dissatisﬁed at any time with any aspect of this study, you may contact - anonymously, if
you wish — Peter Vasilenko, Ph.D., Director of the Human Subject Protection Programs
at Michigan State University: (517) 355-2180, fax: (517) 432-4503, email: irb@msu.edu,
or regular mail: 202 Olds Hall, East Lansing, MI 48824.

****************Ilﬂhl‘*****************************************************

I have read this consent form and I agree to take part in this research.

 

 

Participant Signature Date

145

REFERENCES

Abbot, Andrew (1990), “A primer on sequence methods,” Organization Science, 1(4),
375-392.

Alderson, W. (1950), “Marketing efﬁciency and the principal of postponement,” Cost
and Profit Outlook, 3, 1-3.

Ali, Abdul (1994), “Pioneering versus incremental innovation: review and research
propositions, Journal of Product Innovation Management, 11(1), 46-61.

Amit, Raphael and Paul J. H. Schoemaker (1993), “Strategic assets and organizational
rent,” Strategic Management Journal, 14(1), 33-46.

Barney, Jay B. (1991), “Firm resources and sustained competitive advantage,” Journal of
Management, 17(1), 99-120.

Biggadike, E. R. (1979), Corporate Diversiﬁcation: Entry, Strategy, and Performance,
Cambridge, MI: Harvard University Press.

Bonoma, Thomas V. (1985), “Case research in marketing: opportunities, problems, and a
process,” Journal of Marketing Research, 22 (May 1985), 199-208.

Bowersox, Donald J. and David J. Closs (1989), “Simulation in logistics: A review of
present practice and a look to the future” Journal of Business Logistics, 10(1),
133-148.

, Patricia J. Daugherty, Cornelia L. Droge, Dale S. Rogers, and D. L. Wardlow
(1989), Leading edge logistics: Competitive positioning for the 1990s, The
Council of Logistics Management, Oak Brook, IL.

, David J. Closs, and Theodore P. Stank (1999), 21" Century Logistics: Making
supply chain integration a reality, Council of Logistics Management.

, Theodore P. Stank, and Patricia J. Daugherty (1999), “Lean launch: Managing
product introduction risk through response-based logistics,” Journal of Product
Innovation Management, 16(6), 557-568.

, David J. Closs, and M. Bixby Cooper (2002), Supply Chain Logistics
Management, McGraw-Hill/Irwin: New York, p.558.

Bowman, Douglas and Hubert Gatignon (1995), “Determinants of competitor response
time to a new product introduction,” Journal of Marketing Research, 32(1), 42-
53.

Bronnenberg, Bart J. and Carl F. Mela (2004), “Market roll-out and retailer adoption of
new brands,” Marketing Science, 23(4), 500-518.

146

Bucklin, L. P. (1965), “Postponement, speculation, and the structure of distribution
channels,” Journal of Marketing Research, 2, 26-31.

Burghout, Wilco (2004), “A note on the number of replication runs in stochastic trafﬁc
simulation models” [http://www.infrathh.se/ctr/publikﬁationer/ctr2004__0l .pdf].

Calantone, Roger J ., C. Anthony Di Benedetto, and Theodore P. Stank (2005),
“Managing the supply chain implications of launch,” in Kahn, Kenneth B.
(editor), The PDMA Handbook of new product development, Hoboken, NJ: John
Wiley & Sons.

Chatterjee, Sayan and Birger Wemerfelt (1991), “The link between resources and type of
diversiﬁcation: theory and evidence,” Strategic Management Journal, 12(1), 33-
48.

Choffray, Jean Marie and Gary L. Lilien (1984), “A decision-support system for
evaluating sales prospects and launch strategies for new products,” Industrial
Marketing Management, 15, 75-85.

Chow, Garland, Trevor D. Heaver and Lennart E. Henriksson (1994), “Logistics
performance: Deﬁnition and measurement,” International Journal of Physical
Distribution and Logistics Management, 24(1), 17-29.

Closs, David J .; Anthony S. Roath; Thomas J. Goldsby; James A. Eckert; and Stephen M.
Swartz (1998) “An empirical comparison of anticipatory and response-based
strategies,” International Journal of Logistics Management, 9(2), 21-34.

Colby, B. N., S. Kennedy and L. Milanesi (1991), “Content analysis, cultural grammars,
and computers,” Qualitative Sociology, 14, 373-387.

Cooper, Robert G. (1993), Winning at new products: Accelerating the process from idea
to launch, Reading, MA: Addison-Wesley.

and Elko J. Kleinschmidt (1987), “New products: What separates winners
from losers?” Journal of Product Innovation Management, 4(3), 169-184.

 

Corbett, Charles J. (2001), “Stochastic inventory systems in a supply chain with
asymmetric information: Cycle stocks, safety stocks, and consignment stock,”
Operations Research, 49(4), 487-501.

Daugherty, Patricia J ., Theodore P. Stank and Alexander B. Ellinger (1998), “Leveraging
logistics/distribution capabilities: The effect of logistics service on market share,”
Journal of Business Logistics, 19(2), 35-51.

Day, George S. (1994), “The capabilities of market-driven organizations,” Journal of
Marketing, 58(4), 37-52.

147

and Robin Wensley (1988), “Assessing advantage: A framework for
diagnosing competitive superiority,” Journal of Marketing, 52(2), 1-20.

Debruyne, Marion, Rudy Moenaert, Abbie Grifﬁn, Susan Hart, Erik Jan Hultink, and
Henry Robben (2002), “The impact of new product launch strategies on
competitive reaction in industrial markets,” Journal of Product Innovation
Management, 19(2), 159-170.

Di Benedetto, C. Anthony (1999), “Identifying the key success factors in new product
launch, Journal of Product Innovation Management, 16(6), 530—544.

Doloi, Hemanta and Ali J aafari (2001), “Conceptual simulation model for strategic
decision evaluation in project management,” Logistics Information Management,
15(1/2), 88-104.

Droge, Cornelia, J ayanth J ayaram and Shawn Vickery (2000), “The ability to minimize
the timing of new product development and introduction: An examination of
antecedent factors in the North American automobile supplier industry,” Journal
of Product Innovation Management, 17(1), 24-40.

Eisenhardt, Kathleen M. (1989), “Building theories from case study research,” Academy
of Management Review, 14(4), 532-550.

Ellram, Lisa M. (1996), “The use of the case study method in logistics research,” Journal
of Business Logistics, 17(2), 93-138.

Fahy, John (2001) The role of resources in global competition, London & New York:
Routledge.

Fawcett, Stanley E., Roger J. Calantone and Sheldon R. Smith (1997), “Delivery
capability and ﬁrm performance in international operations,” International
Journal of Production Economics, 51, 191-204.

Fisher, Marshall L. (1997), “What is the right supply chain for your product?” Harvard
Business Review, 75 (March/April), 105-116.

Gatignon, Hubert, Erin Anderson, and Kristiaan Helsen (1989), “Competitive reactions to
market entry: Explaining interﬁrrn differences,” Journal of Marketing Research,

26(1), 44-55.

, Barton Weitz and Pradeep Bansal (1990), “Brand introduction strategies and
competitive environments,” Journal of Marketing Research, 27(4), 390-401.

, and J can-Marc Xuereb (1997), “Strategic orientation of the ﬁrm and new
product performance,” Journal of Marketing Research, 34(1), 77-90.

148

Glasserman, Paul and Yashan Wang (1998), “Leadtirne-inventory trade-offs in assemble-
to-order systems,” Operations Research, 46(6), 85 8-871.

Green, Donna H. and Adrian B. Ryans (1990), “Entry strategies and market performance:
Causal modeling of a business simulation,” Journal of Product Innovation
Management, 7(1), 45-58.

, Donald W. Barclay, and Adrian B. Ryans (1995), “Entry strategy and long-term
performance: Conceptualization and empirical examination,” Journal of
Marketing, 59(4), 1-16.

Grifﬁn, Abbie, and Albert L. Page (1996), “PDMA’s success measurement project:
recommended measures by project and strategy type,” Journal of Product
Innovation Management, 13(6), 478-496.

Grifﬁn, Larry J. (1993), “Narrative, event-structure analysis, and causal interpretation in
historical sociology,” The American Journal of Sociology, 98(5), 1094-1133.

Grifﬁs, Stanley E., Martha Cooper, Thomas J. Goldsby, and David J. Closs (2004),
“Performance measurement: Measure selection based upon ﬁrm goals and
information reporting needs,” Journal of Business Logistics, 25(2), 95-118.

Grimm, Curtis M., Thomas M. Corsi and R. D. Smith (1993), “Determinants of strategic
change in the LTL motor carrier industry: A discrete choice analysis,”
Transportation Journal, 32(4), 56-61.

Guiltinan, Joseph P. (1999), “Launch strategy, launch tactics, and demand outcomes,”
Journal of Product Innovation Management, 16(6), 509-529.

Gummesson, Evert (2000), Qualitative methods in management research, Thousand
Oaks, CA: Sage Publications.

Gunasekaran, Angappa, C. Patel and Ercan Tirtiroglu (2001), “Performance measures
and metrics in a supply chain environment,” International Journal of Operations
and Production Management, 21(1/2), 71-87.

Hair, Joseph F ., Rolph B. Anderson, Ronald T. Tatharn, and William C. Black (1998),
Multivariate Data Analysis, 5th Ed, Upper Saddle River, NJ: Prentice Hall.

Harrington, Thomas C., Douglas M. Lambert, and Jay U. Sterling (1992), “Simulating
the ﬁnancial impact of marketing and logistics decisions, International Journal of

Physical Distribution and Logistics Management, 22(7), 3-12.

Hart, Susan and Nikolaos Tzokas (2000), “New product launch “mix” in growth and
mature markets,” Benchmarking, 7(5), 389-405.

149

Haughton, Michael A. (2002), “Measuring and managing the learning of route re-
optimization of delivery vehicle drivers,” Journal of Business Logistics, 23(2),
45-66.

Hefferman, Troy (2004), “Trust formation in cross-cultural business-to-business
relationships,” Qualitative Market Research, 7(2), 114-125.

Heil, Oliver P. and Rockney G. Walters (1993), “Explaining competitive reactions to new
products: An empirical signaling study,” Journal of Product Innovation
Management, 10(1), 53-65.

Heise, D. R. (1989), “Modeling event structures,” Journal of Mathematical Sociology,
14(2), 139-169.

Henard, David H. and David M. Szymanski (2001), “Why some new products are more
successful than others,” Journal of Marketing Research, 38(3), 362-375.

Hultink, Erik Jan, Abbie Grifﬁn, Susan Hart, and Henry S. J. Robben (1997), “Industrial
new product launch strategies and product development performance” Journal of

Product Innovation Management, 14(4), 243-257.

, Abbie Grifﬁn, Henry S. J. Robben, and Susan Hart (1998), “In search of generic
launch strategies for new products” International Journal of Research in
Marketing, 15(3), 269-285.

, and Susan Hart (1998), “The world's path to the better mousetrap: myth or
reality? An empirical investigation into the launch strategies of high and low
advantage new products,” European Journal of Innovation Management, 1(3),

106-122.

, and Henry S. J. Robben (1999), “Launch strategy and new product performance:
An empirical examination in The Netherlands,” Journal of Product Innovation
Management, 16(6), 545-556.

and Fred Langerak (2002), “Launch decisions and competitive reactions: an
exploratory market signaling study,” Journal of Product Innovation Management,
19(3), 199-212.

IBM Business Consulting Services (2001), “The GMA 2005 Logistics Survey: Supply
Chain Performance in Food, Grocery, and Consumer Products,” An IBM Institute
for Business Value Executive Brief.

J oshi, B. D., R. Unal, N. H. White, and W. D. Morris (1996), “A framework for the
optimization of discrete event simulation models,” Proceedings of the 17‘h ASEM
National Conference, Dallas, TX.

150

Kelton, David W., Randall P. Sadowski and Deborah A. Sadowski (2003), Simulation
with Arena, 3rd Edition, New York: McGraw-Hill.

Kleinschmidt, Elko J. and Robert G. Cooper (1991), “The impact of product
innovativeness on performance,” Journal of Product Innovation Management,
8(4), 240-251.

Kuester, Sabine; Christian Homburg; and Thomas S. Robertson (1999), “Retaliatory
behavior to new product entry,” Journal of Marketing, 63(4), 90-106.

Lambert, Douglas M. and Terrance L. Pohlen (2001), “Supply Chain Metrics,”
International Journal of Logistics Management, 12(1), 1-19.

Lambkin, Mary (1988), “Order of entry and performance in new markets,” Strategic
Management Journal, 9(Special Issue), 127-140.

Langerak, Fred, Erik Jan Hultink, and Henry S.J. Robben (2004), “The impact of market
orientation, product advantage, and launch proﬁciency on new product

performance and organizational performance,” Journal of Product Innovation
Management, 21(2), 79-94.

Law, Averill M., Michael McComas and Stephen G. Vincent (1994), “The crucial role of
input modeling in successful simulation studies,” Industrial Engineer, 26(7), 55-
59.

and w. David Kelton (2000), Simulation Modeling and Analysis, 3rd Edition,
New York: McGraw-Hill.

Lee, Yikuan and Gina Colarelli O'Connor (2003), “New product launch strategy for
network effects products,” Journal of the Academy of Marketing Science, 31(3)
241 -25 5.

Li, Tiger and Roger J. Calantone (1998), “The impact of market knowledge competence
on new product advantage: Conceptualization and empirical examination,”
Journal of Marketing, 62(4), 13-29.

Lieberman, Marvin B. and David B. Montgomery (1998), “F irst-mover (dis)advantages:
Retrospective and link with the resource-based view,” Strategic Management
Journal, 19(12), 1111-1125.

Lilien, Gary L. and Eunsang Yoon (1990), “The timing of competitive market entry: an
exploratory study of new industrial products,” Management Science, 36(5), 568-
585.

Lynch, Daniel F., Scott B. Keller, and John Ozment (2000), “The effects of capabilities
and strategy on ﬁrm performance,” Journal of Business Logistics, 21(2), 47-68.

151

Marginson, David E. W. (2002), “Management control systems and their effects on
strategy formation at middle-management levels: evidence from a UK.
organization,” Strategic Management Journal, 23(1 1), 1019-1031.

Maria, Anu (1997), “Introduction to modeling and simulation,” Proceedings of the 1997
Winter Simulation Conference.

Marucheck, Ann and Marilyn McClelland (1992), “Planning capacity utilization in an
assemble-to-order environment,” International Journal of Operations and
Production Management, 12(9), 18-38.

Masters, James K. (1987), “Analysis of the life-of-type buy decision,” Journal of
Business Logistics, 8(2), 40-56.

Michigan State University Global Logistics Research Team [MSUGLRT] (1995), World
Class Logistics: The Challenge of Managing Continuous Change, Oak Brook, IL:
Council of Logistics Management.

Miles, Matthew B. and A. Michael Huberman (1994), Qualitative data analysis, 2nd
edition, Thousand Oaks, CA: Sage Publications.

Miles, Raymond E. and Charles C. Snow (1994), F it, Failure and the Hall of Fame: How
companies succeed or fail, New York: Macmillan.

Montoya-Weiss, Mitzi M. and Roger J. Calantone (1994), “Determinants of new product
performance: A review and meta-analysis,” Journal of Product Innovation
Management, 1 1(5), 397-417.

Olavarrieta, Sergio and Alexander B. Ellinger (1997), “Resource-based theory and
strategic logistics research,” International Journal of Physical Distribution &
Logistics Management, 27(9/10), 559-587.

Ottum, Brian D. (1996), “Launching a new consumer product,” in Rosenau, Milton D.,
Abbie Grifﬁn, George A. Castellion, and Ned F. Anschuetz (1996) The PDMA
Handbook of New Product Development, New York: John Wiley & Sons, p. 381-
394.

Pagh, Janus D. and Martha C. Cooper (1998), “Supply chain postponement and
speculation strategies: How to choose the right strategy,” Journal of Business
Logistics, 19(2), 13-33.

Patterson, Kirk A., Curtis M. Grimm, and Thomas M. Corsi (2004), “Diffusion of Supply
Chain Technologies,” Transportation Journal, 43(3), 5-23.

152

Penrose, Edith (1959), “The theory of the growth of the ﬁrm,” in Foss, N. J. (Ed) (1997)
Resources, ﬁrms, and strategies: A reader in the resource-based perspective,
Oxford & New York: Oxford University Press.

Pentland, Brian T. (1999), “Building process theory with narrative: From description to
explanation,” Academy of Management Review, 24(4), 711-724.

Peteraf, Margaret A. (1993), “The cornerstones of competitive advantage: 8 resource-
based view,” Strategic Management Journal, 14(3), 179-191.

Porter, Michael E. (1980), Competitive Strategy: Techniques for analyzing industries and
competitors, New York: The Free Press.

Raj agopalan, Nandini and Gretchen M. Spreitzer (1997), “Toward a theory of strategic
change: A multi-lens perspective and integrative framework,” Academy of
Management Review, 22(1), 48-79.

Ray, Gautam, Jay B. Barney, and Waleed A. Muhanna (2004), “Capabilities, business
processes, and competitive advantage: Choosing the dependent variable in

empirical tests of the resource-based view,” Strategic Management Journal,
25(1), 23-37.

Ramasesh, Ranga (1993), “A logistics-based inventory model for J IT procurement,”
International Journal of Operations and Production Management, 13(6), 44-58.

Rees, Loren Paul, Allen G. Greenwood, and Fernando C. Siochi (2002), “A best-ﬁrst
search approach for determining starting regions in simulation optimization,” IIE
Transactions, 34(3), 283-295.

Riessman, Catherine K. (1993), Narrative Analysis, Newbury Park: Sage Publications.

Robertson, Thomas S., J ehoshua Eliashberg, and Talia Rymon (1995), “New product
announcement signals and incumbent reactions,” Journal of Marketing, 59(3), 1-
1 5.

Robinson, William T. (1988) “Marketing mix reaction to entry,” Marketing Science, 7(4),
368-385.

Rodrigues, Alexandre M. (2004), “Simulation of the impact of forecasting accuracy on
supply chain performance: The bias effect,” PhD Dissertation, Michigan State
University.

Sabherwal, Rajiv and Daniel Robey (1993), “An empirical taxonomy of implementation

processes based on sequences of events in information system development,”
Organization Science, 4(4), 548-576.

153

Sargent, Robert G. (2000), “Veriﬁcation, validation, and accreditation of simulation
models,” Proceedings of the 2000 Winter Simulation Conference.

Shankar, Venkatesh (1999), “New product introduction and incumbent response
strategies: Their interrelationship and the role of multi-market contact,” Journal
of Marketing Research, 36(3), 327-344.

Smith, Ken G., Curtis M. Grimm, Ming-Jer Chen, and Martin J. Gannon (1989),
“Predictors of response time to competitive strategic actions: Preliminary theory
and evidence, Journal of Business Research, 18(3), 245-258.

Song, J ing-Sheng (1998), “On the order-ﬁll-rate in a multi-item, base-stock inventory
system,” Operation Research, 46(6), 831-845.

Stank, Theodore P., Thomas J. Goldsby, Shawnee K. Vickery and Katrina Savitskie
(2003), “Logistics service performance: Estimating its inﬂuence on market
share,” Journal of Business Logistics, 24(1), 27-56.

Stevenson, William B. and Danna N. Greenberg (1998), “The formal analysis of
narratives of organization change,” Journal of Management, 24(6), 741-762.

Stock, Gregory N., Noel P. Greis, John D. Kasarda (1999), “Logistics, strategy and
structure: A conceptual framework,” International Journal of Physical
Distribution & Logistics Management, 29(4), 224-239.

Stock, James R. and Douglas Lambert (2001 ), Strategic Logistics Management, 4th
Edition, New York: McGraw-Hill/Irwin.

Terwiesch, Christian, Christoph Loch, and Niederkoﬂer Martin (1998), “When product
development performance makes a difference: A statistical analysis in the
electronics industry,” Journal of Product Innovation Management, 15(1), 3-15.

Tholke, J urg M.; Erik Jan Hultink, and Henry S. J. Robben (2001), “Launching new
product features: A multiple case examination,” Journal of Product Innovation
Management, 18(1), 3-14.

Tracey, Michael (2004), “A holistic approach to new product development: New insights,
Journal of Supply Chain Management, 40(4), 37-55.

Urban, G. L. and J. R. Hauser (1993), Design and marketing of new products, Englewood
Cliffs, NJ: Prentice-Hall.

Van der Zee, D. J. and J. G. A. J. van der Vorst (2005), “A modeling framework for

supply chain simulation: Opportunities for improved decision making,” Decision
Sciences, 36(1), 65-95.

154

Van Hoek (1999), “Postponement and the reconﬁguration challenge for food supply
chains” Supply Chain Management, 4(1), 18-34.

Venkateswaran, J. and Y. J. Son (2004), “Impact of modeling approximations in supply
chain analysis — an experimental study,” International Journal of Production
Research, 42(15), 2971-2992.

Venkatraman, N. and John E. Prescott (1990), “Environment-Strategy coalignment: An
empirical test of its performance implications,” Strategic Management Journal,
11(1), 1-23.

Waller, Matt, M. Eric Johnson, and Tom Davis (1999) “Vendor-managed inventory in the
retail supply chain,” Journal of Business Logistics, 20(1), 183-203.

, Pratibha A. Dabholkar, and Julie J. Gentry (2000), “Postponement, product
customization, and market-oriented supply chain management,” Journal of
Business Logistics, 21(2), 133-159.

Walter, C. K. (1975), “Stockout cost models: Empirical tests in a retail situation,”
Journal of Marketing, 39(3), 56-60.

Wemerfelt, Birger (1984), “A resource-based view of the ﬁrm,” Strategic Management
Journal, 5(2), 171-180.

Wong, Veronica (2002), “Antecedents of international new product rollout timeliness,”
International Marketing Review, 19(2), 120-132.

Yap, Chee Meng and Wm. B. Souder (1994), “Factors inﬂuencing new product success
and failure in small entrepreneurial high-technology electronics ﬁrms,” Journal of
Product Innovation Management, 11(5), 418-432.

Yin, Robert K. (2003), Case study research: Design and methods, Thousand Oaks, CA:
Sage Publications.

Yoon, Eunsang and Gary L. Lilien (1985), “New industrial product performance: The
effects of market characteristics and strategy,” Journal of Product Innovation
Management, 3(3), 134-144.

Zacharia, Zach G. and John T. Mentzer (2004), “Logistics salience in a changing
environment,” Journal of Business Logistics, 25(1), 187-210.

Zaj ac, Edward J ., Matthew S. Kraatz, and Rudi K. F. Bresser (2000), “Modeling the
dynamics of strategic ﬁt: A normative approach to strategic change,” Strategic
Management Journal, 21(4), 429-453.

155

Zinn, Walter and Donald Bowersox (1988), “Planning physical distribution with the
principle of postponement,” Journal of Business Logistics, 9(2), 117-136.

Zinszer, Paul H. (1983), “Inventory costing: A return on inventory approach to
differentiating inventory risk,” Journal of Business Logistics, 4(2), 20-39.

156

   

1111111111111