_______________________________________________________________
22 July 2015
ProQuest
Table of contents
1. An empirical analysis of warehouse measurement systems in the context of supply chain implementation 1
22 July 2015
ii
ProQuest
Document 1 of 1
Page 1 of 12
ProQuest
process fall into one of three primary dimensions: utilization, productivity, and effectiveness.' These authors also
stated that for any measure to be effective, it must be assessed across eight separate criteria: validity,
robustness, usefulness, integration, economy, compatibility, level of detail, and behavioral soundness.5 A
consortium of companies and academic institutions, under the guidance of Pittiglio, Rabin, Todd, and McGrath
(PRTM), developed a comprehensive set of agreed-upon supply chain metrics that can be used as standards
and can pass assessment using the eight criteria stated above.6 These measures fall into one of four
categories: customer satisfaction/quality, time, costs, and assets. The warehouse measures used in this
research represent all four of these categories.
A customer focus is of paramount importance when developing performance measures. For the purposes of this
research, these customers are industrial buyers and receivers, not consumers. Many organizations focus solely
on what can be called internal, or conformance, measures such as order fill and inventory turns. While these are
extremely important because of their downstream impact on the customer, they must not be measured alone.
Other metrics focusing on customer reaction to service and cost levels must be incorporated into a
comprehensive performance measurement system. Understanding how order fill and inventory turns influence
the customer's reaction is far more important than the internal measure alone. Developing the relationships
between internal measurement performance and external customer reactions allows a firm to estimate the
relationships between service and revenue, thus allowing logistics to be a competitive advantage in the
management of the supply chain.
Research has indicated that companies that use a supply chain strategy might use different types of
performance metrics than firms that do not utilize the concept of the supply chain.7 No research was found
indicating whether firms believe their measures for evaluating performance are effective, regardless of SCM
implementation. Therefore, this research will attempt to empirically determine the differences, if any, between
performance measurement systems used by firms that use a SCM strategy and those that do not, and to
assess the relationship between performance measure effectiveness and level of SCM implementation.
Research Purpose
This research is exploratory. It attempts to evaluate the extent of SCM implementation among firms and the
nature and effectiveness of their performance metrics. Supply chains include many types of firms, from
manufacturers to transportation carriers to retailers. Because of this complexity, this research will focus on only
one type of firm-warehousesin order to simplify data collection and analysis. This choice was made for two
reasons: First, warehouses play a critical intermediate role between supply chain members, affecting both
supply chain costs and service. Second, the performance metrics shown in Table 1 are very applicable to the
operations in a warehouse and capture its cost and service impacts on the supply chain.
RESEARCH QUESTIONS AND HYPOTHESES
A critical aspect of this research was the requirement for segregating respondent firms into those that have
implemented a supply chain strategy and those that have not. Research conducted by Mercer Management
Consultants identified four constructs necessary for the presence of supply chain management: (1) strategy; (2)
integrated processes; (3) technology and information; and (4) structure, people, and culture.8 The Mercer
research identified several questions to be used for each construct to identify both its importance and status of
implementation. This research utilized only one question per construct because it was intended to segregate the
respondents only by status of implementation. Appendix A contains a definition of each construct and the
question used to classify firms on each construct. The survey instrument asked the respondents to indicate on a
scale of zero to seven the level of implementation of supply chain management for each construct within their
firms. A zero response meant implementation was not planned; responses of one through seven indicated that
implementation of supply chain management was planned, in progress, or fully implemented, respectively. The
respondent base was then divided into those that were not planning to implement supply chain management (a
response of zero) and those that were (responses one through seven). This resulted in two distinct groups of
22 July 2015
Page 2 of 12
ProQuest
respondents on which the hypotheses and analyses are based. The nonimplementing group will be referred to
as Level 0 and the implementing group will be referred to as Level 4 for the remainder of this article.
Many firms will profess to the implementation of the supply chain management concept. To determine the level
of implementation of supply chain management in the respondent base, the first objective of this research
generated Research Ouestion 1:
RQ1: What is the level of supply chain management implementation for the firms in the respondent base?
Two important elements of this research are supply chain management and measurement. Again, previous
research has indicated that firms practicing SCM might use different types of measurements. Two types of
measurements were addressed by this research: the unit of measurement and the measure. Firms will differ in
the unit of measurement that is used in assessing the effectiveness of their warehouse operations. Some will
use cases, others lines, and still others, pallets. To determine the most commonly used units of measurement
for Level 0 and Level 4 firms, the first objective of this research generated Hypothesis 1:
H1: The most popular primary units of warehouse measurement will not differ between Level 0 and Level 4
firms.
Just as the units of warehouse measurement will vary among firms and types of industry, so will the actual
types of measures. Firms might measure personnel, equipment, storage, receiving, and order fulfillment by both
performance and productivity metrics. The complexity of these measurement systems will increase as the
number of activities performed by the warehouse increases. An objective of this research was to determine if
the most commonly used types of measurements differed between Level 0 and Level 4 firms. This resulted in
Hypothesis 2 to satisfy the requirements of the first objective of this research:
H2: The most popular types of warehouse measures will not differ between Level 0 and Level 4 firms.
In some firms, warehouse measurement systems are complex and comprehensive. Yet there is sometimes a
feeling that the systems are not effective at focusing on the important aspects of warehouse operations and
customer needs. To determine the perceived level of effectiveness of the warehouse measures being used, the
second objective of this research generated Hypothesis 3:
H3: The effectiveness of current warehouse measurement systems will not differ between Level 0 and Level 4
firms.
Finally, this research wanted to determine whether or not there is a relationship between the perceived
effectiveness of warehouse measurement systems and the level of supply chain implementation. The rationale
for this objective is based on the notion that a supply-chainoriented firm will be more aware of channel partner
impacts and desires than a non-supplychain-oriented firm, thus resulting in more focused and effective
measurement systems for supply chain firms. This generated Hypothesis 4 to satisfy the requirements of the
second research objective:
H4: A firm's perceived effectiveness of its warehouse measurement system is not related to its level of supply
chain implementation.
METHODOLOGY
The overall research design required a survey of logistics and warehousing executives to identify the
characteristics of their warehouse measurement systems, level of supply chain implementation, and perceived
effectiveness of warehouse measurement systems. A survey instrument was designed and sent to members of
the Warehouse Education and Research Council.
The discussion in this section and the "Results" section will attempt to incorporate the methodological tests as
suggested by Mentzer and Flint to establish the validity of the research design and its results.9 In an attempt to
establish external realism validity and internal instrumentation validity, a pre-test of the survey instrument was
sent to ten warehouse and logistics executives. Because of the sophisticated nature of the audience, a large
pre-test sample was not necessary,'" although the sample was chosen to represent the target population as
closely as possible." A follow-up phone call was made to all the pre-test survey recipients to clarify any
22 July 2015
Page 3 of 12
ProQuest
comments that were made about the content and clarity of the survey. Nine of the pre-tests were returned, none
with significant changes to the instrument. As such, all nine pre-test responses were included in the overall data
analysis.
A nonresponse bias test was performed on the two mailing groups to establish internal selection validity. A chisquare test was performed on one categorical variable, four demographic variables, and five response variables
from the survey. None of the results were significant at the .10 level; thus nonresponse bias will not be assumed
to be a significant factor in this analysis.
The survey instrument was constructed in order to best identify the warehouse measures used by firms. The
survey consists of common warehouse measures obtained from previous research by Mentzer and Conrad,'2
Capplice and Sheffi,'3 Ackerman,l4 Jenkins,'5 and PRTM.16 The measures are divided into five categories:
order fulfillment, storage, receiving, customer satisfaction, and cost and earnings. Order fulfillment is further
broken into five subcategories: labor and equipment productivity; overall productivity; labor, equipment, and
overall utilization; labor and equipment performance; and overall performance. Receiving is divided into two
subcategories: labor, equipment, and overall productivity; and utilization and performance. Overall, the
respondents were given seventy-seven different measures and were asked which ones they used. Each section
supplied an "other" category in case a specific measure was not identified. Each section also asked the
respondents to identify how effective they perceived the measures in that section to be.
The respondents were also asked to identify which primary unit of measurement they would be using when
responding to the measurement questions. This was done in an attempt to make the survey instrument less
complex and shorter. The options were dollar value, cartons, units/pieces, pallets, weight, lines, invoices,
orders, and other.
The next section asked the respondents to identify their perceived level of implementation on four items that are
used to define the supply chain management concept. These items were taken from the Mercer research, as
mentioned previously. Four items were used because one item alone could not define the complexity of the
supply chain concept, resulting in face/content validity of the supply chain construct.
Finally, various types of demographic data were collected to help describe the respondent base as well as serve
as a basis for analysis.
An initial mailing of the survey and cover letter were sent to 982 warehouse and logistics executives. The
effective sample size was reduced to 980 because either the intended respondent left the firm or the firm's
business was not relevant to this research. The initial mailing resulted in 169 usable responses and a second
mailing produced an additional 127 responses for a total of 296 responses, or a 30 percent response rate.
To determine trait validity, two tests were conducted.17 First, a factor analysis was conducted on the four
supply chain items and on the ten effectiveness items. These will comprise constructs to be used later in the
analysis. The four supply chain items loaded on a single factor as well as did the ten effectiveness items,
indicating that each construct is unidimensional. Cronbach's coefficient alphas were also generated for the
above constructs.18 The alpha for the supply chain construct was .942 and for the effectiveness construct was
.932. The results of these analyses appear to satisfy the requirements of construct validity.'
RESULTS
As previously reported, a total of 296 respondents are included in the analysis. To answer Research Question
1, Table 1 shows that a large majority (79.1 percent) of the respondents indicated that they either were
implementing or have implemented the concept of supply chain management as defined in this research. This
significant difference between these two groups is somewhat surprising. A demographic analysis between these
two groups showed that both are composed primarily of manufacturing firms and both showed a very similar
distribution across the various ranges of firm revenues. However, the single largest revenue category for Level
0 firms was $1 million to $50 million (30.6 percent) and for Level 4 firms was over $1 billion (36.1 percent). This
characteristic offers some explanation for the differences between these groups. Many firms have found that
22 July 2015
Page 4 of 12
ProQuest
implementing a supply chain concept can require significant investment. Smaller firms might not possess the
investment capital to pursue this strategy. Table I also gives other pertinent demographic statistics of the
respondent group. The typical respondent to the survey was a manager at a manufacturing firm, with annual
revenues exceeding $1 billion, that operates its warehouses as cost centers. This is probably representative of
the profile of the WERC membership. mentation for Level 4 firms across the four supply chain constructs. As
can be seen, the average implementation levels are very close for all four constructs and have no statistical
difference. One conclusion about this result is that all four constructs of SCM need to be implemented in concert
with one another. Although the results of Table 2 show that Strategy is the furthest along in implementation,
Process, Information, and Culture are also far along in implementation. This result can be seen in practice with
firms participating in Cooperative Planning, Forecasting, and Replenishment (CPFR) initiatives or with firms
implementing Enterprise Resource Planning (ERP) systems.
Hypotheses 1 generated the results seen in Table 3. The percentages under each column add to more than 100
percent because the survey instrument allowed each participant to select more than one primary unit of
measure. Only one statistically different primary unit of measure existed between the two groups: dollar value.
As such, Hypothesis 1 is supported by the data, and the null hypothesis is accepted. The previous definition of
SCM included the management of product, information, and cash flows. Table 3 shows primary units of
measurement for product (cartons, units/pieces, pallets, weight, and lines), information (invoices and orders),
and cash (dollar value). The results of this analysis imply that the measurement of the impact of warehousing on
cash flow is one element that distinguishes between firms that practice SCM and those that do not.
Hypothesis 2 generated the results shown in Tables 4 through 8. The acronym "UOM" shown on these tables
represents the unit of measure the respondent chose when responding to this section of the survey. In total,
22 July 2015
Page 5 of 12
ProQuest
respondents were presented with seventy-seven different measures to evaluate. Each table (4 through 8)
represents, in the researchers' judgment, the most commonly used measures by both Level 0 and Level 4 firms.
Table 4 shows that thirteen of thirty-five order fulfillment measures were used most frequently by both types of
firms. Of these thirteen, eleven were statistically different at the .10 level. In other words, more Level 4 firms
used the eleven most popular measures than did Level 0 firms. Most of the differences appear to be a result of
Level 4 firms' ability to measure or quantify an entire process rather than just a part of it. For example, under
Labor and Equipment Productivity, Level 0 and Level 4 firms both measured total UOM picked/total labor hours
picking. However, more Level 4 firms were measuring total UOM picked/total labor hours. The difference is in
the latter part of each measure: total labor hours picking versus total labor hours. Total labor hours is more
inclusive of all labor activity, while picking hours is only a part of total labor hours. The differences under Overall
Performance can be attributed to the fact that Level 4 firms are more likely to use measures of total process
time, satisfaction/quality, and asset productivity than are Level 0 firms. These represent three of the four
categories of supply chain metrics, as previously introduced.
Table 5 shows that respondents heavily used 25 percent of the Storage Measures suggested in the survey
instrument, with all three being statistically different. All three measures are representative of asset productivity.
More Level 4 firms are concerned with measuring asset productivity than are Level 0 firms. This might be a
result of being able to implement Integrated Processes and Technology and Information (defined in Appendix
A), which allow firms to better track and manage inventories at a facility level as well as at customer and
supplier levels.
22 July 2015
Page 6 of 12
ProQuest
Only one difference existed between Level 0 and Level 4 firms for receiving measures. This can be seen in
Table 6. Similar to the result in Table 4 for picking measures, significantly more Level 4 firms measured labor
hours receiving as a percentage of total labor hours worked. This measure better reflects receiving as a part of
an entire process in the warehouse. A focus on process is a characteristic of Level 4 firms.
Table 7 shows that Level 4 firms are more likely to employ customer satisfaction measures than would Level 0
firms. A statistical significance exists among three of the four most widely used measures. Interestingly, the only
measure that is not statistically significant appears to be the only one that is reactive in nature, relying on
measuring customer complaints (implying an order error). The other three measures are more proactive in
nature since they rely on customer perception of warehouse process outputs and should be able to detect
potential process failure before it occurs.
The ability or desire to measure the impact of warehouse cost on a firm's revenue and profitability appears to be
more characteristic of Level 4 firms. Table 8 shows that significantly more Level 4 firms measure profitability
impacts than do Level 0 firms. Level 4 firms appear to be concerned with identifying the warehouse as an
investment center, rather than just a cost center. This is important when reporting warehouse activity to upper
management. Because a majority of the measures showed a statistical difference between Level 0 and Level 4
firms, it can be concluded that the data reject Hypothesis 2.
Survey respondents were asked to rate the overall effectiveness of the measures in the five major categories
just presented in Tables 4 through 8. Hypothesis 3 proposes that Level 4 firms will rate their measures as being
no more or less effective than will Level 0 firms. Table 9 contains the results of this analysis. With two
exceptions, the differences seen between the two groups concerning measure effectiveness agree with the
differences seen in Tables 4 through 8. This also means that the data reject Hypothesis 3. Under order
fulfillment measures, only Labor and Equipment Performance is not statistically different between the two
groups. Table 4 shows that significantly more Level 4 firms use this measure than do Level 0 firms. However,
no difference exists between the two groups as to their perception of the effectiveness of this measure. The
second exception is Utilization and Performance. While the difference shown in Table 9 is not significant, the
difference shown in Table 6 is significant. Again, more Level 4 firms are using this measure than are Level 0
22 July 2015
Page 7 of 12
ProQuest
firms, but both groups have similar perceptions of the effectiveness of this measure.
Finally, Hypothesis 4 generated the results seen in Table 10. The multiple regression model tested is stated:
Measure Effectiveness = f(Strategy, Process, Information, Culture)
The four independent variables are the items in the supply chain implementation construct described earlier.
The relationship implied in this model states that as a firm continues to develop a supply chain strategy, its
perception of its measure effectiveness improves. This is a logical continuation of Hypothesis 3. The results
from this hypothesis, shown in Table 9, imply that firms that are closer to full supply chain implementation
perceive their measures to be more effective than do firms that have not adopted a supply chain philosophy.
Several regression models were run on the data. Table 10 shows that the first model, including all respondents,
is significant at the .0043 level. This implies that there is a relationship between level of supply chain
implementation and perceived measure effectiveness. The R-square statistics were low for the models. These
statistics are a measure of the linear relationship between the dependent and independent variables and
indicate the strength of the relationship. A low R-square also implies that there are other variables, not included
in the model, that account for variation in the dependent variable. This makes intuitive sense since
measurement effectiveness could also be influenced by such variables as data integrity, complexity of the
supply chain, or measurement construction. However, the independent variables included in the models in this
research were significant, indicating that they do have a positive relationship with the dependent variable. As
such, Hypothesis 4 is rejected by the data. Several demographic variables were also captured by the survey.
Regression models were run on these variables to determine if certain sub-groups of respondents were
responsible for the significance of the relationship. Table 10 also shows these results. The categories under
each demographic variable are different from those shown in Table 1. Several categories had to be collapsed
into others to provide enough degrees of freedom to allow the regression results to be unbiased. As the data
show, manufacturing firms having revenues between $50 million and $499 million, and measuring their
operations as cost centers, show a significant relationship between supply chain implementation and perceived
measure effectiveness. Disappointing, however, are the relatively low R-square statistics for each model. This
implies that even though there is a relationship between the dependent and independent variables in the model,
there are other factors that account for perceived measure effectiveness. One possible explanation for this is
that a supply chain orientation might capture only the "integration" criterion in Capplice and Sheffi's
categorization of measure effectiveness.
22 July 2015
Page 8 of 12
ProQuest
Page 9 of 12
ProQuest
training in the required skills for supply chain management; performance measurements that enforce supply
chain management performance; regular examination of performance gaps and root cause analysis; inclusion of
customers and suppliers in performance measurement; and a plan to measure change.
Footnote
ENDNOTES
Footnote
See, for example: Lisa Harrington, "Logistics, Agent for Change: Shaping the Integrated Supply Chain,"
Transportation and Distribution (January 1995), p. 30; Tom Davis, "Effective Supply Chain Management," Sloan
Management Review (Summer 1993), p. 37; Todd R. Hammel, and Laura Rock Kopczak, "Tightening the
Supply Chain," Production and Inventory Management Journal (2 quarter 1993), p. 65; Charles Scott and Roy
Westbrook, "New Strategic Tools for Supply Chain Management," International Journal of Physical Distribution
and Logistics Management, No. 1 (1991), p. 23; and Graham Stevens, "Integrating the Supply Chain,"
International Journal of Physical Distribution and Materials Management, Vol. 19, No. 8 (1989), p. 3. 2 CSC
Consulting, Inc. Efficient Healthcare Consumer
Footnote
Response (November 1996), p. 2. ' Kurt Salmon Associates, Inc., Efficient Consumer Response: Enhancing
Consumer Value in the Grocery Industry, (Washington, D.C.: The Food Marketing Institute, 1993).
1 Chis Caplice and Yossi Sheffi, "A Review and Evaluation of Logistics Metrics," The International Journal of
Logistics Management, Vol. 5, No. 2 (1994), p. 18.
T Ibid., p. 14.
Footnote
I Pittiglio, Rabin, Todd, and McGrath, Integrated Supply Chain Performance Measurement A Multi-Industry
Consortium Recommendation (Westin, Ma, 1994).
' See, for example: Battaglia and Tyndall, op. cit.; and Garland Chow, Trevor D. Heaver, and Lennart E.
Henriksson, "Logistics Performance: Definition and Measurement," International Journal of Physical Distribution
and Logistics Management, Vol. 24, No. I (1994), pp. 17-28.
Footnote
' Barry Jennings, "Integrated Supply Chain Management Best Practice: Introduction and Survey," (August
1995), unpublished presentation.
9 John T. Mentzer and Daniel J. Flint, "Validity in Logistics Research," Journal of Business Logistics, Vol. 18,
No. I (1997), pp. 199-216.
22 July 2015
Page 10 of 12
ProQuest
" Shelby 0. Hunt, Richard D. Sparkman Jr., and James B. Wilcox, "The Pretest in Survey Research: Issues and
Preliminary Findings," Journal of Marketing Research, 19 (May 1982), pp. 269-273.
Footnote
" Donald S. Tull and Del I. Hawkins, Marketing Research Meaning, Measurement, and Method (New York;
Macmillan, 1976).
12 John T. Mentzer and Brenda Ponsford Konrad, "An Efficiency/Effectiveness Approach Logistics Performance
Analysis," Journal of Business Logistics, Vol. 12, No. 1 (1991 ), 33-62.
3 Caplice and Sheffi, op.cit. '4 Kenneth B. Ackerman, Practical Handbook of Warehousing, 31d Edition (New
York: Van Norstrand Reinhold, 1990).
Footnote
15 Creed Jenkins, Complete Guide to Modern Warehouse Management (Englewood Cliffs, NJ: Prentice-Hall,
1990). 6 Pittiglio, et.al., op.cit.
" Convergent validity, a type of trait validity, was tested here. Since several different measures were used for
each construct, convergent validity was tested using factor analysis to determine if all of the measures
converged on a common statistical factor.
" Chronbach's Alphas, also a measure of trait validity, compares how well each question correlates with the
combination of all the other questions measuring a construct.
19 Achieving construct validity means that the theoretical phenomena identified in the research have been
correctly defined and measured.
AuthorAffiliation
Mr. Kiefer is support operation manager, $1' Maintenance Department, U.S. Army, and is headquartered in
Germany. Mr. Novack, CTL, is associate professor of business logistics, The Pennsylvania State University,
University Park, Pennsylvania 16802.
The authors would like to thank the Penn State Center for Logistics Research for funding this research.
Appendix
Appendix A. Survey Questions Used to Determine Level of Supply Chain Implementation
Appendix
Strategy -- Aligning supply chain strategy with business goals; senior management commitment to supply chain
management; a total system approach; customer service strategy to meet different customer requirements;
establishing strategy alliances with key suppliers; and outsourcing non-core, non-strategic supply chain
activities.
Appendix
Integrated Processes - Use of cross-functional teams for process design and improvement; use of process
owners; use of life-cycle management into supply chain processes; integrating manufacturing, customers, and
suppliers into the design process; utilizing total corporate leverage in procurement; shifting functions to the most
efficient provider; monitoring supplier performance; integrate and balance supply, demand, and financial plans
and objectives; delivery systems with tailored service; jointly manage inbound and outbound transportation; and
regular monitoring customer satisfaction levels with feedback to all supply chain processes.
Appendix
Technology and Information -- Aligned with key business processes; minimal data redundancy visible to all;
availability of a data warehouse that is widely available; enterprise-wide planning systems; institutionalized
sharing of technology assets across divisions; use of EDI between company, customers and suppliers; system
for demand forecasting, distribution planning, production planning, and material planning that are highly
integrated; and manufacturing execution systems to track material flow and production costs as well as
providing order status information to customer service.
22 July 2015
Page 11 of 12
ProQuest
Subject: Supply chains; Warehouse management systems; Measurement; Statistical analysis; Studies;
Location: US
Classification: 9190: US; 5160: Transportation management; 5240: Software & systems; 5330: Inventory
management; 9130: Experimental/theoretical treatment
Publication title: Transportation Journal
Volume: 38
Issue: 3
Pages: 18-27
Number of pages: 10
Publication year: 1999
Publication date: Spring 1999
Publisher: Pennsylvania State University Press
Place of publication: Lock Haven
Country of publication: United States
Publication subject: Transportation
ISSN: 00411612
CODEN: TRNJAE
Source type: Scholarly Journals
Language of publication: English
Document type: PERIODICAL
Accession number: 01820444
ProQuest document ID: 204590562
Document URL: http://search.proquest.com/docview/204590562?accountid=39226
Copyright: Copyright American Society of Transportation and Logistics Spring 1999
Last updated: 2012-04-04
Database: ABI/INFORM Complete
_______________________________________________________________
Contact ProQuest
Copyright 2015 ProQuest LLC. All rights reserved. - Terms and Conditions
22 July 2015
Page 12 of 12
ProQuest