Description
This report presents the largest and most current single compilation of findings on the extent, causes, and consumer responses to retail out-of-stock (OOS) situations in the fast-moving consumer goods industry. It is also perhaps the first study that enumerates OOS on a worldwide basis.
O
O
O
O
O
O
O
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and Consumer Responses
O
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and
Consumer Responses
A research study conducted at Emory University, Goizueta Business School, Atlanta, GA U.S.;
University of St. Gallen, Institute of Technology Management, Switzerland;
and College of Business and Administration, University of Colorado at Colorado Springs, U.S.
Principal Authors:
Thomas W. Gruen, Ph.D.
University of Colorado, Colorado Springs, U.S.
Daniel S. Corsten, Dr.
Institute of Technology Management, University of St. Gallen, Switzerland
Sundar Bharadwaj, Ph.D.
Goizueta Business School, Emory University, U.S.
This study was funded by a grant from the Procter & Gamble Company
For the Grocery Manufacturers of America, The Food Marketing Institute and CIES – The Food Business Forum
@ 2002 by the Grocery Manufacturers of America. All rights reserved. This publication may
not be reproduced, stored in any informational retrieval system or transmitted in whole or
part in any means – electronic, mechanical, photocopy, recording or otherwise – without the
express written permission of GMA. Contact the Industry Affairs Group, (202) 337-9400 or
1010 Wisconsin Avenue, NW #900, Washington, DC 20007, U.S., or email at
[email protected] for permission to reprint material from this report.
$40 GMA/FMI Member Price
$80 Non-member Price
Table of Contents
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v
Chapter 1: Introduction and Overview of Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
Study Background and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
Methodology and Description of Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Chapter 2: Overall Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
A. Extent of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
1. What is an Out-of-Stock? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
2. Overall Extent of OOS Worldwide and by Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11
3. OOS Extent by Category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12
4. Variation Rates by Time of Day and Day of Week . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14
5. Variation in OOS Rates by Promotion, Movement, and Duration of OOS . . . . . . . . . . . . .15
6. Conclusions from Analysis of the Extent of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
B. Consumer Response to OOS Situations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
1. Consumer Response Types and Impact on Retailers and Manufacturers . . . . . . . . . . . . .18
2. U.S. Consumer Study Summary Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
3. Consumer Responses Vary Across Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20
4. Consumer Responses Vary by Region Country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21
5. Worldwide Responses by Category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23
6. Drawing Comparisons Across Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
7. Implications of the Worldwide Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
8. Measuring Consumer Reactions with Item Velocity Monitoring . . . . . . . . . . . . . . . . . . . .26
9. Comparison of Consumer Responses with Previous OOS Studies . . . . . . . . . . . . . . . . . . .27
10 Final Questions Regarding Consumer Responses to OOS Situations . . . . . . . . . . . . . . . .28
C. Causes of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
1. OOS Causes by Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
2. Examination of Primary Root Causes by Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
3. Other Explanations and Attributions of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
4. Determining Root Causes of Retail Chain Characteristics . . . . . . . . . . . . . . . . . . . . . . . . .38
D. The Financial and Managerial Implications of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40
1. De?ning the Implications of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40
2. Quantifying the Losses due to OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41
3. Reported Costs of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
4. OOS Fixes and Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46
5. Examples of New, Best of Breed Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49
6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50
Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53
A. 1. Listing of Studies Examined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
2. Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55
B. Consumer Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56
C. 1. Charts of Consumer Responses by Country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
2. General Observations of Consumer Responses in Various Countries . . . . . . . . . . . . . . . .62
D. 1. Listing of Root Cause Studies Examined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63
2. Root Cause Analysis Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64
E. Probability of Complete Satisfaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65
F. Authors’ Vitae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66
Acknowledgments
The Association co-sponsors would like to thank the authors and their universities – Thomas W.
Gruen, Ph.D., University of Colorado, Colorado Springs; Daniel S. Corsten, Dr., Institute of Technology
Management, University of St. Gallen, Switzerland; and Sundar Bharadwaj, Ph.D., Goizueta Business
School, Emory University.
We would especially acknowledge the research assistance of Gunther Kucza of the University of St
Gallen in Switzerland and Ravi Nayak of Emory University in Atlanta, Georgia.
A very special thanks also goes to The Procter & Gamble Company which provided the grant for the
research for this report.
iv
Executive Summary
Overview and Objectives
Key Findings
v
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and Consumer Responses
Overview and Objectives
This report presents the largest and most current single compilation of ?ndings on
the extent, causes, and consumer responses to retail out-of-stock (OOS) situations
in the fast-moving consumer goods industry. It is also perhaps the ?rst study that
enumerates OOS on a worldwide basis.
The inputs for this report come from 52 studies that examine out-of-stocks, includ-
ing the previously published results of 16 industry and academic studies as well
as the results from an additional 36 studies proprietary to this report. To provide a
sense of the extensiveness of the studies that were used to develop this report,
consider the following:
Number of retail outlets examined: 661
Number of consumer goods categories included: 32
Number of consumers surveyed worldwide: 71,000
Number of countries represented: 29
Studies addressing extent of OOS: 40 (of 52 total studies)
Studies addressing the root causes of OOS: 20 (of 52 total studies)
Studies addressing the consumer responses to OOS: 15 (of 52 total studies)
The objective of the study has been three-fold:
To present an updated and accurate map of facts surrounding retail out-of-
stocks in the consumer goods industry.
To examine out-of-stocks worldwide, analyzing rationale for similarities
and differences.
To examine differences in ?ndings based on different methodologies of
measuring out-of-stocks.
Key Findings
THE EXTENT HAS NOT DECREASED FROM EARLIER STUDIES.
Out-of-stocks remains a large problem for retailers, distributors and manufactur-
ers in the worldwide consumer goods industry. The advances in supply chain
management, the initiatives of Ef?cient Consumer Response (ECR) and category
management, and the investments in inventory-tracking technology have not –
by and large – reduced the overall level of out-of-stocks on store shelves from
what was reported in previous studies. Out-of-stock rates vary wildly among
retailers and their outlets depending on a variety of factors, but the majority
tends to fall in the range of 5-10 percent.
More importantly, in studies that examine faster selling and/or promoted prod-
ucts, the OOS rate regularly exceeds 10 percent. The overall average OOS rate
worldwide is estimated at 8.3 percent and is illustrated on Exhibit 1.
vi Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Executive
Summary
FIRST
MOST OF THE DIRECT OOS CAUSES OCCUR AT — AND MUST BE REMEDIED AT —
THE RETAIL STORE.
The analysis shows that 70-75 percent of out-of-stocks are a direct result of retail store
practices (either underestimating demand or having ordering processes/cycles that are
too lengthy) and shelf-restocking practices (product is at the store but not on the shelf).
Exhibit I-2 divides the responsibility for OOS into its major components, and interest-
ingly, the responsibility breaks out into the following approximate general groupings:
Retail store ordering and forecasting causes (about one-half of OOS).
Retail store shelving and replenishment practices in which the product is
at the store but not on the shelf (about one-fourth of OOS).
Combined upstream causes (about one-fourth of OOS).
The report provides extensive detail behind these general summary numbers in the
section on Causes of OOS. (See Chapter 2, Section C for detailed information.)
vii Executive Summary
SECOND
E
x
h
i
b
i
t
1
Worldwide 8.3
USA 7.9
Europe 8.6
Other Regions 8.2
0.0 2.0 4.0 6.0 8.0 10.0
Overall OOS Extent (Averages)
E
x
h
i
b
i
t
2 OOS Causes: Worldwide Averages
Percentage OOS
Total Upstream Causes 28%
Store Ordering and Forecasting 47%
In the Store, Not on the Shelf 25%
NEW EVIDENCE IS PRESENTED THAT CHANGES PREVIOUS UNDERSTANDING OF THE
WAYS CONSUMERS RESPOND TO OUT-OF-STOCKS.
Our consumer data of more than 71,000 consumers surveyed show an increasing
willingness of consumers — when confronted with an out-of-stock situation — to
seek those items at an alternative outlet. These consumer studies show —
depending on the product category — that when confronted with an out-of-stock
situation, 21 to 43 percent of consumers will make that purchase at another store,
while another 7 to 25 percent will not buy the item at all.
The consumer studies show that retailers are likely to lose almost one-half of the
intended purchases when a consumer confronts an out-of-stock. This loss does
not include the impact of substituting, which generally tends toward a cheaper
substitute.
The worldwide averages across eight major categories are shown in Exhibit 3. The
report provides extensive detail behind these general summary numbers in the
section on consumer response to OOS (Chapter 2 B).
THE IMPLICATION OF THE ABOVE FINDING SUGGESTS THAT THE COST OF OUT-OF-
STOCKS TO RETAILERS IS GREATER THAN PREVIOUSLY REPORTED.
Our ?ndings show that a typical retailer loses about 4 percent of sales due
to having items out-of-stock. A loss of sales of 4 percent translates into a
earnings per share loss of about $0.012 (1.2 cents) for the average ?rm in
the grocery retailing sector, where the average earnings per share, already
is about $0.25 (25 cents) per year. (For more details behind these general
summary numbers, see the section on implications of retail OOS in
Chapter 2 D.)
viii Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
THIRD
FOURTH
E
x
h
i
b
i
t
3
Worldwide Consumer Responses to OOS
(Average across eight categories)
Do not Purchase Item 9%
Substitute — Same Brand 19%
Delay Purchase 15%
Substitute — Different Brand 26%
Buy Item at Another Store 31%
THIS EXAMINATION OF OUT-OF-STOCKS SHOWS SOME STRIKING SIMILARITIES
WORLDWIDE AS WELL AS CLEAR DIFFERENCES BY REGION.
The aggregate root cause attributed to retail stores for OOS situations varies little
across regions. However, while the causes attributed to the retail store are consis-
tent in the aggregate, clear differences among the regions can be seen when it
comes to the amount of store ordering vs. forecasting vs. replenishment.
When examining consumer reactions to OOS, consumer brand substitution varies
greatly across regions. Differences in the variance of the extent of OOS can be
found in developing countries (greater variance). This study sheds considerable
light on both the worldwide differences and the worldwide similarities in terms of
extent, causes and consumer responses to OOS.
THIS STUDY INTRODUCES OOS COMPARISON MEASUREMENTS USING A NEW METHOD.
This study examined several measurements of out-of-stocks by a new method that
uses scanner data and product movement to predict and identify out-of-stock situ-
ations. Most OOS studies (including many of the ones examined for this report)
have relied on physical store audits that provide measures of out-of stocks at spe-
ci?c periods of time. However, identifying an out-of-stock through a physical
audit does not necessarily identify the true effect of that out-of-stock, nor does it
provide a precise measure of the duration of the out-of-stock. The latter consider-
ation, the duration, is managerially relevant, since the length an item is out-of-
stock indicates the true damage to the store’s sales. The ?ndings using the new
method of measurement were reasonably consistent with the store audits, and this
suggests that the new method provides reliable measures.
ix Executive Summary
SIXTH
FIFTH
1 Chapter 1: Introduction and Overview of Study
Chapter 1: Introduction and Overview of Study
Study Background and Objectives
Methodology and Description of Studies
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and Consumer Responses
2 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Introduction and Overview of Study
In the past few years, three key forces have converged to add pressure and
urgency to OOS issues. For the following three reasons, as never previously in his-
tory, the issue of out-of-stocks is of greater importance to retailers and their supply
chain partners.
First, to provide motivation to address the issue is the fact that consumers
are becoming less tolerant of OOS situations. With more information at
their ?ngertips and more available outlets and channels for purchasing,
consumers are being trained to be less accepting of OOS situations. With
worldwide consistency, consumers will increasingly shop at an alternate
outlet to ?nd the item they need. To NOT address the OOS issue is clearly
becoming more hazardous.
Second, the opportunity for direct impact when addressing the problem
has increased. As retailing continues its consolidation and becomes glob-
al, retailers ?nd solutions are becoming increasingly valuable, as they can
provide solutions for these issues on a worldwide basis.
Third, technology provides new ways to address OOS. This is providing
retailers a new-found ability to address OOS, rather than the traditionally
recommended solutions that carry the heavy ongoing costs of increased
labor or greater inventory safety stocks.
Throughout this report, extensive background information and current data are
provided relating to these primary ?ndings of our 18-month worldwide study. As
the Executive Summary highlights, OOS continues to be a problem for retailers
and their supply chain partners. Previous published studies have examined the
issue regionally, but this report shows that OOS can and must be addressed by
retailers worldwide.
OUT-OF-STOCKS AND THE RETAILER
Retailing demands extraordinary commitment to detail from its managers.
Retailing also presents its managers with multiple challenges that simultaneously
beg for attention. One of those challenges has long been keeping products that
customers want and need in stock and available.
If retailing were not extremely competitive, the implications of out-of-stock prod-
ucts would not command the attention of retail managers. In metropolitan areas
worldwide, however, retail competition is keen and continues to intensify. Given
this situation, having products in stock is becoming more and more a requirement
to play in the game.
At the same time, products continue to proliferate. According to the FMI Web
site, the number of SKUs in 2001 in an average grocery store was nearly 25,000.
This makes the task of keeping products in stock and available all that more dif?-
Chapter
One
A. STUDY BACKGROUND
AND OBJECTIVES
3 Chapter 1: Introduction and Overview of Study
cult. The retailer’s problem with out-of-stock items validates the adage that “retail
is detail.”
INTERNATIONAL ECR AND RECENT OUT-OF-STOCK RESEARCH PROJECTS
Keeping items in stock is not the sole problem of the retailer, but rather is shared
by the entire supply chain. The Ef?cient Consumer Response (ECR) initiative that
was started in 1993 in the United States by grocery retailers, distributors and man-
ufacturers of fast-moving consumer goods, sought to reduce many of the inef?-
ciencies throughout the supply chain. One of its key strategies – category man-
agement – provided a means for determining what products were most important
to the consumer and to ensure availability of these products. Through category
management, all ECR supply chain players developed practices to guide the right
mix of products more ef?ciently through the supply chain to the ultimate con-
sumer.
As the ECR movement spread worldwide in the late 1990s, it provided a forum for
common industry issues to be heard. One message that arose from all parts of the
world was concern about out-of-stock items. Since all players in the supply chain
share in the problem — and the solution — of out-of-stock items, ECR in Europe,
Asia and Latin America plus joint industry initiatives in the United States provide
venues to address and solve the problem. Despite recent efforts to stem the OOS
trends, however, the level of out-of-stock continues to haunt retailers and their
supply chain partners alike. It is clear that additional study of the extent, root
causes and consumer reactions to out-of-stock items is necessary to clarify the
problem for the industry. It is also necessary to provide insights and justi?cation
for the level of resources that can and/or should be committed to addressing out-
of-stock issues.
Currently, OOS is one of the top agenda items for non-U.S. ECR. As the ECR
organizations mature in Asia, Europe, Latin America and elsewhere, they have
begun to shift their attention from the processes and components that lead to
industry ef?ciency to more speci?c outcome objectives, such as reductions in out-
of-stocks. For example, the 2001 ECR Asia conference theme was “50/50: 50
Percent Reduction in Inventory and 50 Percent Reduction in Out-of-Stocks.” ECR
Europe is currently conducting a large pan-European OOS study that is slated to
be one of the major future discussion topics. ECR organizations that have recent-
ly conducted studies and released their ?ndings include ECR Australia, ECR
France, ECR China, and ECR Thailand. More have been proposed or are under-
way in other countries. (Additional information on ECR organizations and related
links can be found on the Internet at www.globalscorecard.net, www.ecr-academ-
ics.org, and www.ecr-journal.org.)
EXISTING RESEARCH ON OUT-OF-STOCKS
While there is a ?urry of recent activity in OOS research, the applied and academic
studies over the past several years that have examined the out-of-stock issue have shed
4 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
light on the issue from speci?c perspectives. A study was conducted by Andersen
Consulting and the Coca-Cola Retailing Research Council in 1996. This study exam-
ined 11 categories of consumer goods in 10 stores across the United States for a
month and found that on average 8.2 percent of the items in the categories examined
were out-of-stock on a typical afternoon. Additionally this study reported that 46 per-
cent of consumer purchases were at risk of purchase at another store, purchase delay,
or substituting a lower-value product when an out-of-stock situation occurred.
While the focus of the Coca-Cola study was on the grocery retailing, similar studies
were conducted in the late 1990s by industry associations representing chain drug
stores and convenience stores. In 2002, the Grocery Manufacturers of America
(GMA) published a study that focuses on the top 25 grocery categories across, with
an in-depth look at seven direct store delivery (DSD) categories. The study was the
most comprehensive in years, tracking 1,600 items in 20 stores in four major U.S.
retailers for 14 consecutive days. Additionally, one thoU.S.nd shoppers were inter-
viewed. Additional data was provided on the top 25 grocery categories from 500
stores across ?ve regions of the United States, resulting in more than 92 million
individual store/item/day observations. The study found that shoppers can not ?nd
the item they want to buy 7.4 percent of the time. Some 40 percent of these shop-
pers – when confronted with an out-of-stock situation – either postpone their pur-
chase or buy elsewhere, placing $6 billion in annual sales at risk in those top 25
categories. The study found that stock-outs can jump to as high as 17.1 percent dur-
ing store promotions and that when a product is unavailable on the shelf, a retailer
can potentially lose $75,000 annually per store.
The GMA study, the Coca-Cola/Andersen study and others have been conducted
in the United States Additionally, a smattering of proprietary studies has been
reported in business publications. Finally, in addition to the published studies
done for industry, academic research has made important contributions to the
understanding of out-of-stocks. A complete listing of all of the studies that were
used as background for this current study can be found in Appendix A.
DIFFICULTIES IN MEASURING EXTENT, CAUSES AND CONSUMER REACTION
The major limitation in studying out-of-stocks rest in the large number of factors
that affect the outcomes of any particular study. Some of the primary factors that
cause the extent of reported out-of-stocks to vary include:
De?nition of out-of-stock item. (For example, the product may be in multi-
ple places in the store, but out-of-stock at one location but not another.)
Methodology used in counting out-of-stocks (Includes frequency and tim-
ing of measures e.g. time of day, day of week and other seasonal factors.)
The velocity or speed of turnover of items examined (When only the
fastest moving items are examined, rates are higher than when all SKUs
are examined.)
The way new and discontinued SKUs are considered.
Promotions and promotional coordination among channel members.
5 Chapter 1: Introduction and Overview of Study
Similarly, a wide variety of factors also affect the causes of out-of-stocks such as:
Shifts in consumer demand.
Promotional planning periods.
Sophistication of the supply chain and channel practices.
Standard channel problems, such as demand ampli?cation (“bullwhip
effect”).
Allocation of shelving to SKUs based on case size, as opposed to product
movement (which constrains and affects ordering practices).
Finally, while there are only a small number of primary actions that consumers
can take when confronted by an out-of-stock situation, several factors affect the
likelihood of action that will be taken in any given situation, such as:
Category of products examined, due to varying willingness and ability to
substitute, e.g., product loyalty.
Geographic proximity of competitors.
Overall extent of out-of-stocks (A decision to substitute or not is depend-
ent on the total number of substitutions that a consumer will need to make
in a particular shopping trip.)
STUDY OBJECTIVES
Because there are so many variables, existing studies have had dif?culty making
predictions beyond the speci?c categories, outlets, situations or regions studied.
While several existing published studies have been made available, there has
never been a synthesis of this material.
Based on the issues discussed above, this study has three central objectives.
Triangulate from a variety of studies to develop an overall range of the extent, root
causes and consumer responses to out-of-stocks.
Examine the out-of-stock issue from a global perspective analyzing differences
and similarities across national boundaries.
Present and examine the differences in measurement of out-of-stocks when using
the traditional audit method vs. estimates out-of-stocks from store scanner data.
The basic process used for the study followed ?ve general steps.
1. Collect and review published and unpublished OOS studies worldwide.
2. Collect and review related research on OOS from academic and applied
sources.
3. Delineate ?ndings from research.
4. Isolate limiting factors.
5. Synthesize ?ndings and determine areas of consensus, trends, key ?ndings.
More speci?cally, to develop this report, information was collected and synthe-
sized from the following general sources:
B. METHODOLOGY AND
DESCRIPTION OF STUDIES
USED IN THIS STUDY
6 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Previously published industry reports and studies of out-of-stocks.
New data provided from two large-scale consumer studies conducted in
1999-2000 (one in U.S. and a second identical study conducted in 19
countries outside of North America).
New data provided from studies of three retailers’ scanner and inventory
data conducted in 1999-2001.
New data provided from a series of traditional store audit studies conduct-
ed in 1998-2000. (See Appendix A, Part 1.)
Various academic articles published from 1962-2001 on out-of-stock stud-
ies. (See Appendix A, Part 2.)
Industry press and articles that addressed and/or reported on other out-of-
stock studies. (See Appendix A, Part 1.)
The academic and industry studies provided background and theory regarding the
way out-of-stocks has been measured, the likely consumer responses to out-of-
stocks, and the value of addressing the issue at the retail level. The majority of the
academic studies focused on consumer responses and provided important theoret-
ical and categorical approaches to examining consumer response data. The indus-
try studies were examined to provide baselines for evaluating the information we
would then examine from the new studies. The review of the industry studies led
us to systematically arrange the information contained in all studies into the fol-
lowing categories:
Methodology.
Categories examined.
Extent of out-of-stocks found.
Consumer responses.
Root causes identi?ed and assigned.
Efforts examined / suggested to address out-of-stocks, the costs and
returns.
The logic of the arrangement is straightforward. First the methodology was
reviewed to determine any likely limitations or concerns faced when examining
the data from the study. This methodology also provided a way to categorize the
studies. Second, the categories examined were listed in order to make compar-
isons among the studies that examined the same or similar categories. Consumer
responses to OOS situations tended to vary widely among categories, thus catego-
ry identi?cation is a key variable.
Following general categorization, examination of the extent of out-of-stocks in the
report was the logical place to begin, since it answers the question: “Is there a
problem?” After identifying the extent, the logical next question is: “Does the
OOS matter?” This is answered by examining the consumers’ responses to OOS
situations. The search for the cause to the problem leads to the next question:
7 Chapter 1: Introduction and Overview of Study
“Who is responsible for causing the problem?” This leads to the ?nal questions:
“Can and should it be ?xed? If so, how?”
The above paragraph explains the general format for the presentation of the detail
of the ?ndings. Next came the examination by region in the world (four regions),
by category and by methodology. Chapter 2 presents the ?ndings from the studies
examined for this report.
9 Chapter 2: Overall Findings
Chapter 2: Overall Findings
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and Consumer Responses
A. Extent of OOS
1. What is an Out-of-Stock?
2. Overall Extent of OOS Worldwide and by Region
3. OOS Extent by Category
4. Variation Rates by Time of Day and Day of Week
5. Variation in OOS Rates by Promotion, Movement, and Duration of OOS
6. Conclusions from Analysis of the Extent of OOS
B. Consumer Response to OOS Situations
1. Consumer Response Types and Impact on Retailers and Manufacturers
2. U.S. Consumer Study Summary Findings
3. Consumer Responses Vary Across Categories
4. Consumer Responses Vary by Region Country
5. Worldwide Responses by Category
6. Drawing Comparisons Across Countries
7. Implications of the Worldwide Analysis
8. Measuring Consumer Reactions with Item Velocity Monitoring
9. Comparison of Consumer Responses with Previous OOS Studies
10 Final Questions Regarding Consumer Responses to OOS Situations
C. Causes of OOS
1. OOS Causes by Region
2. Examination of Primary Root Causes by Process
3. Other Explanations and Attributions of OOS
4. Determining Root Causes of Retail Chain Characteristics
D. The Financial and Managerial Implications of OOS
1. De?ning the Implications of OOS
2. Quantifying the Losses due to OOS
3. Reported Costs of OOS
4. OOS Fixes and Implications
5. Examples of New, Best of Breed Scenarios
6. Conclusion
10 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Overall Findings
A. WHAT IS THE EXTENT OF OOS?
After examining 40 studies analysts found that the average OOS rate worldwide is
8.3 percent. While this is the average, the extent reported in each study varied
not only by differing management practices, but also by what is measured. Thus,
this section presents an examination of the extent or scope of out-of-stocks based
on several sub-analyses. These include:
What is the de?nition of an out-of-stock, and how is it measured and cal-
culated?
What is the overall extent of OOS?
How does this vary by
• Region
• Category
• Time of day / day of week
• Promotion
• Product movement
• Brand
• Duration?
What is an acceptable level of OOS?
The de?nition of what makes an OOS affects the extent that gets reported in stud-
ies. While many variations exist, recent studies tend to settle on a consumer-based
de?nition. Even with agreement to use a consumer perspective, two general alter-
native de?nitions emerge based on the method of measurement.
As the ?rst and most accepted approach, the OOS rate is measured as a percent-
age of SKUs that are out-of-stock on the retail store shelf at a particular moment
in time (i.e., the consumer expects to ?nd the item but it is not available). In gen-
eral, studies using this approach begin with the selection of one or more cate-
gories to examine. Next, a sample of stores from a single retail chain is selected,
and a series of physical audits is conducted at the retailer at speci?c times during
the day over a speci?ed period of time. For each category, the OOS rate is calcu-
lated as the average percentage of the SKUs not in stock at the time of the audits.
Normally, the OOS rate is reported for each category individually and then the
categories are averaged (normally unweighted average) to create and report an
overall rate for the study. Due to the number of studies that have used this
approach, a major advantage of using this method is the availability of excellent
baselines. The limitations to this type of measurement include the:
Arbitrary nature of selection of the categories.
Frequency and timing of the audits.
Duration of the study.
Human error that can and does enter from many sources.
Chapter
Two
1. WHAT IS AN OUT-
OF-STOCK
11 Chapter 2: Overall Findings
A second and alternative consumer-based de?nition of an OOS is the number of
times a consumer looks for the SKU and does not ?nd it. The percentage rate is cal-
culated as the number of times the consumer does not ?nd the SKU divided into the
sum of the times the consumer does ?nd the SKU plus the number of times the con-
sumer does not ?nd it. Instead of relying on physical audits, the second approach is
measured through the use of models that determine OOS rates from store scanner
and inventory data. This view provides the advantage of determining the extent of
out of stocks that actually matter to the retailer and the upstream supply chain mem-
bers. The major limitation of this method is that the OOS rates are estimates based
on historical sales patterns, and thus can only be calculated for SKUs that sell with a
minimum frequency (thus cannot detect OOS for very slow moving products). Few
studies have used this method, and therefore baselines do not readily exist. In this
report, the data from three studies that used this method are provided.
Exhibit 4 below presents the worldwide averages. 2. OVERALL EXTENT OF
OOS WORLDWIDE AND
BY REGION
E
x
h
i
b
i
t
4
World Average
40 Studies
8.3
Average
4.9
Low
12.3
USA
11 Studies
7.9
5.6
11.5
NW Europe
13 Studies
7.2
4.2
11.5
SE Europe
9 Studies
10.8
7.0
16.3
Other Regions
7 Studies
8.2
3.3
9.8
High
0.0 3.0 6.0 9.0 12.0 15.0 18.0
Worldwide OOS Extent
Percent
12 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
The average OOS rate for all 40 studies that reliably reported OOS extent was 8.3
percent. The average of the reported highs in the studies was 12.3 percent, and the
average of the lows was 4.9 percent. This is similar to, though slightly higher than,
the primary U.S. benchmark developed in the 1996 Coca-Cola Research Council
sponsored study that was 8.2 percent, which was calculated as the simple average
rate of eight categories that ranged from 3.9 percent to 11.1 percent. The 2002
GMA study reported an average OOS rate of 7.4 percent with DSD categories rang-
ing from 3.2 percent (milk) to 11.2 percent (prepackaged bread).
Keep in mind that the 40 studies examined here used slightly different measurement
methods and different people, measured different categories, and examined differ-
ent durations and different daily and weekly factors. All of these can affect the
measurement. However, when all of the various factors are considered together, the
averages regress to an uncanny similarity, and this provides a sense that the ?ndings
are reliable in the aggregate, and the differences can easily be explained by vari-
ances in categories, methods and regions.
For this study, Europe was split into its northern and western region (Norway, Denmark,
Sweden, France, Belgium, Netherlands, Germany, Switzerland, Austria) and into its
southern and eastern region (Portugal, Spain, Greece, Poland, Hungary, Czech
Republic, Slovakia). Countries within each of these two areas showed similarities in
OOS rates, and differences between the two regions were substantial. Northwest
Europe showed the lowest OOS rates, while Southeast Europe showed the highest.
The Northwest examination was limited due to a lack of detailed studies from the UK,
and Finland. Summary extent numbers were reviewed from four additional studies
from the UK and the extents fell in line with those reported elsewhere in the region.
Unfortunately, study analysts were unable to review the studies in detail and thus did
not include them in the calculations in this request.
OOS rates in other regions (South America and Asia) were lower in average,
although the extents varied as much or more than other regions, and the small num-
ber of studies does not provide a complete representation of these regions.
OOS is often measured by category. A category is a microcosm of the retail store,
and category management principles encourage a focus on retail performance by
category. Of the 40 OOS studies that examined the extent of OOS, 14 of these pro-
vided reliable OOS data by category. Additional studies measured OOS by catego-
ry, but only reported the composite ?ndings and did not report by category. In total,
18 categories provided OOS results except for the GMA DSD study, which detailed
the top 25 categories. However, in only six of these 18 categories did data come
from three or more studies. Thus, the averages were computed and the OOS rates
were reported for these six categories only. Exhibit 5 illustrates the averages and
ranges of OOS for the six categories. Exhibit 6 provides a chart of the category aver-
ages only. Note that the average of the six categories is slightly lower than the over-
all worldwide average based on 40 studies.
3. OOS EXTENT
BY CATEGORY
13 Chapter 2: Overall Findings
E
x
h
i
b
i
t
5
Toilet Tissue
3 Studies
6.6
Average
6.1
Low
7.1
Diapers
6 Studies
7.0
1.9
12.0
Hair Care
6 Studies
9.8
7.0
16.0
Laundry
9 Studies
7.7
2.1
15.6
Salty Snacks
5 Studies
5.3
1.4
8.0
Fem Hygiene
4 Studies
6.8
1.9
10.2
World Average
40 Studies
18 Categories
8.3
4.9
12.3
High
0.0 3.0 6.0 9.0 12.0 15.0 18.0
Out-of-Stock Extent by Category
Percent
Overall OOS Extent (Averages)
E
x
h
i
b
i
t
6
Worldwide Avg
18 Categories
8.3
Fem Hygiene 6.8
Salty Snacks 5.3
0.0 2.0 4.0 6.0 8.0 12.0 10.0
OOS Averages by Category
Percent
Toilet Tissue 6.6
Hair Care 9.8
Laundry 7.7
Diapers 7.0
14 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Thirteen studies measured and reported variations in OOS rates by time of day
and/or day of the week. In general, there are two clear conclusions. First, consis-
tent across all studies are patterns that showed increases in OOS rates in the early
evening hours as opposed to morning or early afternoon. The highest OOS
occurred in the evenings (after 8:00 p.m.), while the lowest were during the early
afternoon. Morning rates were lower due to overnight restocking practices, slight-
ly higher than those after noon, and lower than in the evenings. The conclusion
from these ?ndings is that ordering decisions and replenishment patterns as deter-
mined by store management have an effect on OOS rates.
Second, consistent across all studies are weekly patterns where OOS rates rose
and fell on different days of the week. The 2002 GMA DSD study, which meas-
ured time-of-day and day-of-week stock-outs, also showed that same pattern,
whether or not the retailer is directly responsible for keeping the shelves stocked.
Exhibit 7 shows how the OOS rates vary during days of the week.
All of the studies that reported daily OOS rates showed the same general pattern
of decreasing rates throughout the week, but a large rate on Sunday (and the
resulting carry-over to Monday). This pattern re?ects both retail strategy and eco-
nomic realities. Assuming the weekend to be the heaviest shopping days, re-order-
ing and deliveries occur on Monday and Tuesday. Another reason that Monday
has a high average is that in some countries stores are closed on Sunday and
restocking does not begin until Monday.
4. VARIATION IN OOS
VARY BY TIME OF DAY
AND WEEK
Overall OOS Extent (Averages)
E
x
h
i
b
i
t
7
0.0 2.0 4.0 6.0 10.0 8.0 12.0
OOS by Day of Week
Percent
Sunday 10.9
Saturday 7.3
Friday 8.7
Thursday 9.1
Wednesday 9.8
Tuesday 10.0
Monday 10.9
(Average of 13 Studies)
15 Chapter 2: Overall Findings
Throughout the week, restocking and preparations for the Saturday and Sunday
promotions lead to lower OOS rates. Saturday, despite being the heaviest shop-
ping day, has the lowest OOS as retailers employ extra labor and can ?ll-in using
safety stocks for promoted items. In countries where stores are open on Sundays,
labor is normally at a lower level, and safety stocks for high demand items begin
to be depleted. Thus, any incorrect demand estimation becomes manifested in
OOS increases.
Promotional Effects.
In general, the studies that reported OOS rates on promoted and non-promoted
items consistently showed OOS rates to be higher on the promoted items. In
some cases the differences were minor while in others the differences were sub-
stantial. Although the promoted items should be receiving attention from the
retail store management, all studies that report promotional effects ?nd substan-
tially greater OOS on promoted items than everyday items.
While the differences vary among studies, in general a 2:1 ratio of promoted vs.
non-promoted OOS rates was found. Examples of this in publicly reported stud-
ies include the ECR France study (where promoted items have a 75 percent
greater OOS rates the 1996 Coca Cola U.S. study (where OOS levels of promoted
items were approximately double of non-promoted items), and the 2002 GMA
DSD study (where OOS levels of promoted items were approximately double of
non-promoted items). Several of the proprietary studies examined for this report
found similar results.
One study found that the increase in the amount of discount offered by a promo-
tion corresponded with the OOS rate. Another study highlighted a related prob-
lem where promotional decisions (and the resulting last-minute advertising
changes) based on responses to competitors led to increased OOS when the tim-
ing of the changes were too late to be included in the normal order cycle.
Velocity of Product Movement.
Somewhat overlapping with promoted items, studies that exclusively examined
fast-moving items found higher OOS rates (13 percent-15 percent) than those that
examined entire categories that include both fast-moving and slow-moving items
(8.3 percent average). This translates to a 50-80 percent higher OOS rate for fast
moving vs. all products. The GMA DSD study found that, on average, the top 10
percent of the fastest moving items accounted for 45 percent of the out-of-stocks.
The studies that examine the fast moving items used a different methodology
(scanner data analysis vs. visual audits), and thus some of the difference could be
due to variances in measurement. However, it is clear that the faster-moving
items — promoted or not — have higher OOS rates than slower-moving items.
5. VARIATION BY
PROMOTION, MOVEMENT
AND DURATION OF OOS
16 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Product and Brand Effects on OOS Rates.
The sparse brand-level data available for this analysis was not adequate to make
solid conclusions about speci?c brands within categories. However, it was clear
that the faster-moving items also had more incidences of OOS, although the dura-
tion was not necessarily longer. Thus, in any category, the faster-moving SKUs are
going to incur more frequent OOS, regardless of the brand. The implication of
this – and the value of addressing the faster moving SKUs – is that the faster
movers suffer disproportionately more due to OOS than do slower-movers.
Duration of OOS.
Data on duration of OOS, while sparse, is very interesting. Based on a study of
13 stores in the U.S. by Data Ventures, a U.S. software service provider, the fol-
lowing results were found. When products become OOS, only about 20 percent
are replenished in less than eight hours while a similar percentage remain OOS
for more than three days. Duration is a critical though under-used measure for the
extent of OOS. The traditional measure of OOS (the percentage of SKUs not on
the shelf at a particular point in time) does not provide the measure that is most
meaningful from the perspective of the consumer. When the duration of the OOS
item is considered along with the extent, then a better picture for managerial
action emerges.
All of the above issues (promotion effects, velocity and duration) indicate that both
retail store management systems and practices contribute to OOS extents. While
this will be discussed in more detail with the other implications, it is important to
note that there are two ways to address the higher OOS rates on faster-moving prod-
ucts. First, retailers can pay more attention to high velocity products to ensure that
they get reordered and restocked more frequently. Second, following category man-
agement principles, retailers can examine a category and eliminate some slower-
movers and allocate more shelf space to faster-movers. According to Broniarczyk et
al.’s category management research (1998, Journal of Marketing Research, Vol. 35,
pp. 166-176) sales and customer satisfaction for the category increases following a
reduction in SKUs from a category review.
E
x
h
i
b
i
t
8
Duration of OOS
3 Days or More 19%
1 Day to
This report presents the largest and most current single compilation of findings on the extent, causes, and consumer responses to retail out-of-stock (OOS) situations in the fast-moving consumer goods industry. It is also perhaps the first study that enumerates OOS on a worldwide basis.
O
O
O
O
O
O
O
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and Consumer Responses
O
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and
Consumer Responses
A research study conducted at Emory University, Goizueta Business School, Atlanta, GA U.S.;
University of St. Gallen, Institute of Technology Management, Switzerland;
and College of Business and Administration, University of Colorado at Colorado Springs, U.S.
Principal Authors:
Thomas W. Gruen, Ph.D.
University of Colorado, Colorado Springs, U.S.
Daniel S. Corsten, Dr.
Institute of Technology Management, University of St. Gallen, Switzerland
Sundar Bharadwaj, Ph.D.
Goizueta Business School, Emory University, U.S.
This study was funded by a grant from the Procter & Gamble Company
For the Grocery Manufacturers of America, The Food Marketing Institute and CIES – The Food Business Forum
@ 2002 by the Grocery Manufacturers of America. All rights reserved. This publication may
not be reproduced, stored in any informational retrieval system or transmitted in whole or
part in any means – electronic, mechanical, photocopy, recording or otherwise – without the
express written permission of GMA. Contact the Industry Affairs Group, (202) 337-9400 or
1010 Wisconsin Avenue, NW #900, Washington, DC 20007, U.S., or email at
[email protected] for permission to reprint material from this report.
$40 GMA/FMI Member Price
$80 Non-member Price
Table of Contents
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v
Chapter 1: Introduction and Overview of Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
Study Background and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
Methodology and Description of Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Chapter 2: Overall Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
A. Extent of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
1. What is an Out-of-Stock? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
2. Overall Extent of OOS Worldwide and by Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11
3. OOS Extent by Category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12
4. Variation Rates by Time of Day and Day of Week . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14
5. Variation in OOS Rates by Promotion, Movement, and Duration of OOS . . . . . . . . . . . . .15
6. Conclusions from Analysis of the Extent of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
B. Consumer Response to OOS Situations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
1. Consumer Response Types and Impact on Retailers and Manufacturers . . . . . . . . . . . . .18
2. U.S. Consumer Study Summary Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
3. Consumer Responses Vary Across Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20
4. Consumer Responses Vary by Region Country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21
5. Worldwide Responses by Category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23
6. Drawing Comparisons Across Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
7. Implications of the Worldwide Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
8. Measuring Consumer Reactions with Item Velocity Monitoring . . . . . . . . . . . . . . . . . . . .26
9. Comparison of Consumer Responses with Previous OOS Studies . . . . . . . . . . . . . . . . . . .27
10 Final Questions Regarding Consumer Responses to OOS Situations . . . . . . . . . . . . . . . .28
C. Causes of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
1. OOS Causes by Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
2. Examination of Primary Root Causes by Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
3. Other Explanations and Attributions of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
4. Determining Root Causes of Retail Chain Characteristics . . . . . . . . . . . . . . . . . . . . . . . . .38
D. The Financial and Managerial Implications of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40
1. De?ning the Implications of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40
2. Quantifying the Losses due to OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41
3. Reported Costs of OOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
4. OOS Fixes and Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46
5. Examples of New, Best of Breed Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49
6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50
Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53
A. 1. Listing of Studies Examined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
2. Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55
B. Consumer Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56
C. 1. Charts of Consumer Responses by Country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
2. General Observations of Consumer Responses in Various Countries . . . . . . . . . . . . . . . .62
D. 1. Listing of Root Cause Studies Examined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63
2. Root Cause Analysis Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64
E. Probability of Complete Satisfaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65
F. Authors’ Vitae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66
Acknowledgments
The Association co-sponsors would like to thank the authors and their universities – Thomas W.
Gruen, Ph.D., University of Colorado, Colorado Springs; Daniel S. Corsten, Dr., Institute of Technology
Management, University of St. Gallen, Switzerland; and Sundar Bharadwaj, Ph.D., Goizueta Business
School, Emory University.
We would especially acknowledge the research assistance of Gunther Kucza of the University of St
Gallen in Switzerland and Ravi Nayak of Emory University in Atlanta, Georgia.
A very special thanks also goes to The Procter & Gamble Company which provided the grant for the
research for this report.
iv
Executive Summary
Overview and Objectives
Key Findings
v
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and Consumer Responses
Overview and Objectives
This report presents the largest and most current single compilation of ?ndings on
the extent, causes, and consumer responses to retail out-of-stock (OOS) situations
in the fast-moving consumer goods industry. It is also perhaps the ?rst study that
enumerates OOS on a worldwide basis.
The inputs for this report come from 52 studies that examine out-of-stocks, includ-
ing the previously published results of 16 industry and academic studies as well
as the results from an additional 36 studies proprietary to this report. To provide a
sense of the extensiveness of the studies that were used to develop this report,
consider the following:
Number of retail outlets examined: 661
Number of consumer goods categories included: 32
Number of consumers surveyed worldwide: 71,000
Number of countries represented: 29
Studies addressing extent of OOS: 40 (of 52 total studies)
Studies addressing the root causes of OOS: 20 (of 52 total studies)
Studies addressing the consumer responses to OOS: 15 (of 52 total studies)
The objective of the study has been three-fold:
To present an updated and accurate map of facts surrounding retail out-of-
stocks in the consumer goods industry.
To examine out-of-stocks worldwide, analyzing rationale for similarities
and differences.
To examine differences in ?ndings based on different methodologies of
measuring out-of-stocks.
Key Findings
THE EXTENT HAS NOT DECREASED FROM EARLIER STUDIES.
Out-of-stocks remains a large problem for retailers, distributors and manufactur-
ers in the worldwide consumer goods industry. The advances in supply chain
management, the initiatives of Ef?cient Consumer Response (ECR) and category
management, and the investments in inventory-tracking technology have not –
by and large – reduced the overall level of out-of-stocks on store shelves from
what was reported in previous studies. Out-of-stock rates vary wildly among
retailers and their outlets depending on a variety of factors, but the majority
tends to fall in the range of 5-10 percent.
More importantly, in studies that examine faster selling and/or promoted prod-
ucts, the OOS rate regularly exceeds 10 percent. The overall average OOS rate
worldwide is estimated at 8.3 percent and is illustrated on Exhibit 1.
vi Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Executive
Summary
FIRST
MOST OF THE DIRECT OOS CAUSES OCCUR AT — AND MUST BE REMEDIED AT —
THE RETAIL STORE.
The analysis shows that 70-75 percent of out-of-stocks are a direct result of retail store
practices (either underestimating demand or having ordering processes/cycles that are
too lengthy) and shelf-restocking practices (product is at the store but not on the shelf).
Exhibit I-2 divides the responsibility for OOS into its major components, and interest-
ingly, the responsibility breaks out into the following approximate general groupings:
Retail store ordering and forecasting causes (about one-half of OOS).
Retail store shelving and replenishment practices in which the product is
at the store but not on the shelf (about one-fourth of OOS).
Combined upstream causes (about one-fourth of OOS).
The report provides extensive detail behind these general summary numbers in the
section on Causes of OOS. (See Chapter 2, Section C for detailed information.)
vii Executive Summary
SECOND
E
x
h
i
b
i
t
1
Worldwide 8.3
USA 7.9
Europe 8.6
Other Regions 8.2
0.0 2.0 4.0 6.0 8.0 10.0
Overall OOS Extent (Averages)
E
x
h
i
b
i
t
2 OOS Causes: Worldwide Averages
Percentage OOS
Total Upstream Causes 28%
Store Ordering and Forecasting 47%
In the Store, Not on the Shelf 25%
NEW EVIDENCE IS PRESENTED THAT CHANGES PREVIOUS UNDERSTANDING OF THE
WAYS CONSUMERS RESPOND TO OUT-OF-STOCKS.
Our consumer data of more than 71,000 consumers surveyed show an increasing
willingness of consumers — when confronted with an out-of-stock situation — to
seek those items at an alternative outlet. These consumer studies show —
depending on the product category — that when confronted with an out-of-stock
situation, 21 to 43 percent of consumers will make that purchase at another store,
while another 7 to 25 percent will not buy the item at all.
The consumer studies show that retailers are likely to lose almost one-half of the
intended purchases when a consumer confronts an out-of-stock. This loss does
not include the impact of substituting, which generally tends toward a cheaper
substitute.
The worldwide averages across eight major categories are shown in Exhibit 3. The
report provides extensive detail behind these general summary numbers in the
section on consumer response to OOS (Chapter 2 B).
THE IMPLICATION OF THE ABOVE FINDING SUGGESTS THAT THE COST OF OUT-OF-
STOCKS TO RETAILERS IS GREATER THAN PREVIOUSLY REPORTED.
Our ?ndings show that a typical retailer loses about 4 percent of sales due
to having items out-of-stock. A loss of sales of 4 percent translates into a
earnings per share loss of about $0.012 (1.2 cents) for the average ?rm in
the grocery retailing sector, where the average earnings per share, already
is about $0.25 (25 cents) per year. (For more details behind these general
summary numbers, see the section on implications of retail OOS in
Chapter 2 D.)
viii Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
THIRD
FOURTH
E
x
h
i
b
i
t
3
Worldwide Consumer Responses to OOS
(Average across eight categories)
Do not Purchase Item 9%
Substitute — Same Brand 19%
Delay Purchase 15%
Substitute — Different Brand 26%
Buy Item at Another Store 31%
THIS EXAMINATION OF OUT-OF-STOCKS SHOWS SOME STRIKING SIMILARITIES
WORLDWIDE AS WELL AS CLEAR DIFFERENCES BY REGION.
The aggregate root cause attributed to retail stores for OOS situations varies little
across regions. However, while the causes attributed to the retail store are consis-
tent in the aggregate, clear differences among the regions can be seen when it
comes to the amount of store ordering vs. forecasting vs. replenishment.
When examining consumer reactions to OOS, consumer brand substitution varies
greatly across regions. Differences in the variance of the extent of OOS can be
found in developing countries (greater variance). This study sheds considerable
light on both the worldwide differences and the worldwide similarities in terms of
extent, causes and consumer responses to OOS.
THIS STUDY INTRODUCES OOS COMPARISON MEASUREMENTS USING A NEW METHOD.
This study examined several measurements of out-of-stocks by a new method that
uses scanner data and product movement to predict and identify out-of-stock situ-
ations. Most OOS studies (including many of the ones examined for this report)
have relied on physical store audits that provide measures of out-of stocks at spe-
ci?c periods of time. However, identifying an out-of-stock through a physical
audit does not necessarily identify the true effect of that out-of-stock, nor does it
provide a precise measure of the duration of the out-of-stock. The latter consider-
ation, the duration, is managerially relevant, since the length an item is out-of-
stock indicates the true damage to the store’s sales. The ?ndings using the new
method of measurement were reasonably consistent with the store audits, and this
suggests that the new method provides reliable measures.
ix Executive Summary
SIXTH
FIFTH
1 Chapter 1: Introduction and Overview of Study
Chapter 1: Introduction and Overview of Study
Study Background and Objectives
Methodology and Description of Studies
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and Consumer Responses
2 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Introduction and Overview of Study
In the past few years, three key forces have converged to add pressure and
urgency to OOS issues. For the following three reasons, as never previously in his-
tory, the issue of out-of-stocks is of greater importance to retailers and their supply
chain partners.
First, to provide motivation to address the issue is the fact that consumers
are becoming less tolerant of OOS situations. With more information at
their ?ngertips and more available outlets and channels for purchasing,
consumers are being trained to be less accepting of OOS situations. With
worldwide consistency, consumers will increasingly shop at an alternate
outlet to ?nd the item they need. To NOT address the OOS issue is clearly
becoming more hazardous.
Second, the opportunity for direct impact when addressing the problem
has increased. As retailing continues its consolidation and becomes glob-
al, retailers ?nd solutions are becoming increasingly valuable, as they can
provide solutions for these issues on a worldwide basis.
Third, technology provides new ways to address OOS. This is providing
retailers a new-found ability to address OOS, rather than the traditionally
recommended solutions that carry the heavy ongoing costs of increased
labor or greater inventory safety stocks.
Throughout this report, extensive background information and current data are
provided relating to these primary ?ndings of our 18-month worldwide study. As
the Executive Summary highlights, OOS continues to be a problem for retailers
and their supply chain partners. Previous published studies have examined the
issue regionally, but this report shows that OOS can and must be addressed by
retailers worldwide.
OUT-OF-STOCKS AND THE RETAILER
Retailing demands extraordinary commitment to detail from its managers.
Retailing also presents its managers with multiple challenges that simultaneously
beg for attention. One of those challenges has long been keeping products that
customers want and need in stock and available.
If retailing were not extremely competitive, the implications of out-of-stock prod-
ucts would not command the attention of retail managers. In metropolitan areas
worldwide, however, retail competition is keen and continues to intensify. Given
this situation, having products in stock is becoming more and more a requirement
to play in the game.
At the same time, products continue to proliferate. According to the FMI Web
site, the number of SKUs in 2001 in an average grocery store was nearly 25,000.
This makes the task of keeping products in stock and available all that more dif?-
Chapter
One
A. STUDY BACKGROUND
AND OBJECTIVES
3 Chapter 1: Introduction and Overview of Study
cult. The retailer’s problem with out-of-stock items validates the adage that “retail
is detail.”
INTERNATIONAL ECR AND RECENT OUT-OF-STOCK RESEARCH PROJECTS
Keeping items in stock is not the sole problem of the retailer, but rather is shared
by the entire supply chain. The Ef?cient Consumer Response (ECR) initiative that
was started in 1993 in the United States by grocery retailers, distributors and man-
ufacturers of fast-moving consumer goods, sought to reduce many of the inef?-
ciencies throughout the supply chain. One of its key strategies – category man-
agement – provided a means for determining what products were most important
to the consumer and to ensure availability of these products. Through category
management, all ECR supply chain players developed practices to guide the right
mix of products more ef?ciently through the supply chain to the ultimate con-
sumer.
As the ECR movement spread worldwide in the late 1990s, it provided a forum for
common industry issues to be heard. One message that arose from all parts of the
world was concern about out-of-stock items. Since all players in the supply chain
share in the problem — and the solution — of out-of-stock items, ECR in Europe,
Asia and Latin America plus joint industry initiatives in the United States provide
venues to address and solve the problem. Despite recent efforts to stem the OOS
trends, however, the level of out-of-stock continues to haunt retailers and their
supply chain partners alike. It is clear that additional study of the extent, root
causes and consumer reactions to out-of-stock items is necessary to clarify the
problem for the industry. It is also necessary to provide insights and justi?cation
for the level of resources that can and/or should be committed to addressing out-
of-stock issues.
Currently, OOS is one of the top agenda items for non-U.S. ECR. As the ECR
organizations mature in Asia, Europe, Latin America and elsewhere, they have
begun to shift their attention from the processes and components that lead to
industry ef?ciency to more speci?c outcome objectives, such as reductions in out-
of-stocks. For example, the 2001 ECR Asia conference theme was “50/50: 50
Percent Reduction in Inventory and 50 Percent Reduction in Out-of-Stocks.” ECR
Europe is currently conducting a large pan-European OOS study that is slated to
be one of the major future discussion topics. ECR organizations that have recent-
ly conducted studies and released their ?ndings include ECR Australia, ECR
France, ECR China, and ECR Thailand. More have been proposed or are under-
way in other countries. (Additional information on ECR organizations and related
links can be found on the Internet at www.globalscorecard.net, www.ecr-academ-
ics.org, and www.ecr-journal.org.)
EXISTING RESEARCH ON OUT-OF-STOCKS
While there is a ?urry of recent activity in OOS research, the applied and academic
studies over the past several years that have examined the out-of-stock issue have shed
4 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
light on the issue from speci?c perspectives. A study was conducted by Andersen
Consulting and the Coca-Cola Retailing Research Council in 1996. This study exam-
ined 11 categories of consumer goods in 10 stores across the United States for a
month and found that on average 8.2 percent of the items in the categories examined
were out-of-stock on a typical afternoon. Additionally this study reported that 46 per-
cent of consumer purchases were at risk of purchase at another store, purchase delay,
or substituting a lower-value product when an out-of-stock situation occurred.
While the focus of the Coca-Cola study was on the grocery retailing, similar studies
were conducted in the late 1990s by industry associations representing chain drug
stores and convenience stores. In 2002, the Grocery Manufacturers of America
(GMA) published a study that focuses on the top 25 grocery categories across, with
an in-depth look at seven direct store delivery (DSD) categories. The study was the
most comprehensive in years, tracking 1,600 items in 20 stores in four major U.S.
retailers for 14 consecutive days. Additionally, one thoU.S.nd shoppers were inter-
viewed. Additional data was provided on the top 25 grocery categories from 500
stores across ?ve regions of the United States, resulting in more than 92 million
individual store/item/day observations. The study found that shoppers can not ?nd
the item they want to buy 7.4 percent of the time. Some 40 percent of these shop-
pers – when confronted with an out-of-stock situation – either postpone their pur-
chase or buy elsewhere, placing $6 billion in annual sales at risk in those top 25
categories. The study found that stock-outs can jump to as high as 17.1 percent dur-
ing store promotions and that when a product is unavailable on the shelf, a retailer
can potentially lose $75,000 annually per store.
The GMA study, the Coca-Cola/Andersen study and others have been conducted
in the United States Additionally, a smattering of proprietary studies has been
reported in business publications. Finally, in addition to the published studies
done for industry, academic research has made important contributions to the
understanding of out-of-stocks. A complete listing of all of the studies that were
used as background for this current study can be found in Appendix A.
DIFFICULTIES IN MEASURING EXTENT, CAUSES AND CONSUMER REACTION
The major limitation in studying out-of-stocks rest in the large number of factors
that affect the outcomes of any particular study. Some of the primary factors that
cause the extent of reported out-of-stocks to vary include:
De?nition of out-of-stock item. (For example, the product may be in multi-
ple places in the store, but out-of-stock at one location but not another.)
Methodology used in counting out-of-stocks (Includes frequency and tim-
ing of measures e.g. time of day, day of week and other seasonal factors.)
The velocity or speed of turnover of items examined (When only the
fastest moving items are examined, rates are higher than when all SKUs
are examined.)
The way new and discontinued SKUs are considered.
Promotions and promotional coordination among channel members.
5 Chapter 1: Introduction and Overview of Study
Similarly, a wide variety of factors also affect the causes of out-of-stocks such as:
Shifts in consumer demand.
Promotional planning periods.
Sophistication of the supply chain and channel practices.
Standard channel problems, such as demand ampli?cation (“bullwhip
effect”).
Allocation of shelving to SKUs based on case size, as opposed to product
movement (which constrains and affects ordering practices).
Finally, while there are only a small number of primary actions that consumers
can take when confronted by an out-of-stock situation, several factors affect the
likelihood of action that will be taken in any given situation, such as:
Category of products examined, due to varying willingness and ability to
substitute, e.g., product loyalty.
Geographic proximity of competitors.
Overall extent of out-of-stocks (A decision to substitute or not is depend-
ent on the total number of substitutions that a consumer will need to make
in a particular shopping trip.)
STUDY OBJECTIVES
Because there are so many variables, existing studies have had dif?culty making
predictions beyond the speci?c categories, outlets, situations or regions studied.
While several existing published studies have been made available, there has
never been a synthesis of this material.
Based on the issues discussed above, this study has three central objectives.
Triangulate from a variety of studies to develop an overall range of the extent, root
causes and consumer responses to out-of-stocks.
Examine the out-of-stock issue from a global perspective analyzing differences
and similarities across national boundaries.
Present and examine the differences in measurement of out-of-stocks when using
the traditional audit method vs. estimates out-of-stocks from store scanner data.
The basic process used for the study followed ?ve general steps.
1. Collect and review published and unpublished OOS studies worldwide.
2. Collect and review related research on OOS from academic and applied
sources.
3. Delineate ?ndings from research.
4. Isolate limiting factors.
5. Synthesize ?ndings and determine areas of consensus, trends, key ?ndings.
More speci?cally, to develop this report, information was collected and synthe-
sized from the following general sources:
B. METHODOLOGY AND
DESCRIPTION OF STUDIES
USED IN THIS STUDY
6 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Previously published industry reports and studies of out-of-stocks.
New data provided from two large-scale consumer studies conducted in
1999-2000 (one in U.S. and a second identical study conducted in 19
countries outside of North America).
New data provided from studies of three retailers’ scanner and inventory
data conducted in 1999-2001.
New data provided from a series of traditional store audit studies conduct-
ed in 1998-2000. (See Appendix A, Part 1.)
Various academic articles published from 1962-2001 on out-of-stock stud-
ies. (See Appendix A, Part 2.)
Industry press and articles that addressed and/or reported on other out-of-
stock studies. (See Appendix A, Part 1.)
The academic and industry studies provided background and theory regarding the
way out-of-stocks has been measured, the likely consumer responses to out-of-
stocks, and the value of addressing the issue at the retail level. The majority of the
academic studies focused on consumer responses and provided important theoret-
ical and categorical approaches to examining consumer response data. The indus-
try studies were examined to provide baselines for evaluating the information we
would then examine from the new studies. The review of the industry studies led
us to systematically arrange the information contained in all studies into the fol-
lowing categories:
Methodology.
Categories examined.
Extent of out-of-stocks found.
Consumer responses.
Root causes identi?ed and assigned.
Efforts examined / suggested to address out-of-stocks, the costs and
returns.
The logic of the arrangement is straightforward. First the methodology was
reviewed to determine any likely limitations or concerns faced when examining
the data from the study. This methodology also provided a way to categorize the
studies. Second, the categories examined were listed in order to make compar-
isons among the studies that examined the same or similar categories. Consumer
responses to OOS situations tended to vary widely among categories, thus catego-
ry identi?cation is a key variable.
Following general categorization, examination of the extent of out-of-stocks in the
report was the logical place to begin, since it answers the question: “Is there a
problem?” After identifying the extent, the logical next question is: “Does the
OOS matter?” This is answered by examining the consumers’ responses to OOS
situations. The search for the cause to the problem leads to the next question:
7 Chapter 1: Introduction and Overview of Study
“Who is responsible for causing the problem?” This leads to the ?nal questions:
“Can and should it be ?xed? If so, how?”
The above paragraph explains the general format for the presentation of the detail
of the ?ndings. Next came the examination by region in the world (four regions),
by category and by methodology. Chapter 2 presents the ?ndings from the studies
examined for this report.
9 Chapter 2: Overall Findings
Chapter 2: Overall Findings
Retail Out-of-Stocks:
A Worldwide Examination of Extent, Causes and Consumer Responses
A. Extent of OOS
1. What is an Out-of-Stock?
2. Overall Extent of OOS Worldwide and by Region
3. OOS Extent by Category
4. Variation Rates by Time of Day and Day of Week
5. Variation in OOS Rates by Promotion, Movement, and Duration of OOS
6. Conclusions from Analysis of the Extent of OOS
B. Consumer Response to OOS Situations
1. Consumer Response Types and Impact on Retailers and Manufacturers
2. U.S. Consumer Study Summary Findings
3. Consumer Responses Vary Across Categories
4. Consumer Responses Vary by Region Country
5. Worldwide Responses by Category
6. Drawing Comparisons Across Countries
7. Implications of the Worldwide Analysis
8. Measuring Consumer Reactions with Item Velocity Monitoring
9. Comparison of Consumer Responses with Previous OOS Studies
10 Final Questions Regarding Consumer Responses to OOS Situations
C. Causes of OOS
1. OOS Causes by Region
2. Examination of Primary Root Causes by Process
3. Other Explanations and Attributions of OOS
4. Determining Root Causes of Retail Chain Characteristics
D. The Financial and Managerial Implications of OOS
1. De?ning the Implications of OOS
2. Quantifying the Losses due to OOS
3. Reported Costs of OOS
4. OOS Fixes and Implications
5. Examples of New, Best of Breed Scenarios
6. Conclusion
10 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Overall Findings
A. WHAT IS THE EXTENT OF OOS?
After examining 40 studies analysts found that the average OOS rate worldwide is
8.3 percent. While this is the average, the extent reported in each study varied
not only by differing management practices, but also by what is measured. Thus,
this section presents an examination of the extent or scope of out-of-stocks based
on several sub-analyses. These include:
What is the de?nition of an out-of-stock, and how is it measured and cal-
culated?
What is the overall extent of OOS?
How does this vary by
• Region
• Category
• Time of day / day of week
• Promotion
• Product movement
• Brand
• Duration?
What is an acceptable level of OOS?
The de?nition of what makes an OOS affects the extent that gets reported in stud-
ies. While many variations exist, recent studies tend to settle on a consumer-based
de?nition. Even with agreement to use a consumer perspective, two general alter-
native de?nitions emerge based on the method of measurement.
As the ?rst and most accepted approach, the OOS rate is measured as a percent-
age of SKUs that are out-of-stock on the retail store shelf at a particular moment
in time (i.e., the consumer expects to ?nd the item but it is not available). In gen-
eral, studies using this approach begin with the selection of one or more cate-
gories to examine. Next, a sample of stores from a single retail chain is selected,
and a series of physical audits is conducted at the retailer at speci?c times during
the day over a speci?ed period of time. For each category, the OOS rate is calcu-
lated as the average percentage of the SKUs not in stock at the time of the audits.
Normally, the OOS rate is reported for each category individually and then the
categories are averaged (normally unweighted average) to create and report an
overall rate for the study. Due to the number of studies that have used this
approach, a major advantage of using this method is the availability of excellent
baselines. The limitations to this type of measurement include the:
Arbitrary nature of selection of the categories.
Frequency and timing of the audits.
Duration of the study.
Human error that can and does enter from many sources.
Chapter
Two
1. WHAT IS AN OUT-
OF-STOCK
11 Chapter 2: Overall Findings
A second and alternative consumer-based de?nition of an OOS is the number of
times a consumer looks for the SKU and does not ?nd it. The percentage rate is cal-
culated as the number of times the consumer does not ?nd the SKU divided into the
sum of the times the consumer does ?nd the SKU plus the number of times the con-
sumer does not ?nd it. Instead of relying on physical audits, the second approach is
measured through the use of models that determine OOS rates from store scanner
and inventory data. This view provides the advantage of determining the extent of
out of stocks that actually matter to the retailer and the upstream supply chain mem-
bers. The major limitation of this method is that the OOS rates are estimates based
on historical sales patterns, and thus can only be calculated for SKUs that sell with a
minimum frequency (thus cannot detect OOS for very slow moving products). Few
studies have used this method, and therefore baselines do not readily exist. In this
report, the data from three studies that used this method are provided.
Exhibit 4 below presents the worldwide averages. 2. OVERALL EXTENT OF
OOS WORLDWIDE AND
BY REGION
E
x
h
i
b
i
t
4
World Average
40 Studies
8.3
Average
4.9
Low
12.3
USA
11 Studies
7.9
5.6
11.5
NW Europe
13 Studies
7.2
4.2
11.5
SE Europe
9 Studies
10.8
7.0
16.3
Other Regions
7 Studies
8.2
3.3
9.8
High
0.0 3.0 6.0 9.0 12.0 15.0 18.0
Worldwide OOS Extent
Percent
12 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
The average OOS rate for all 40 studies that reliably reported OOS extent was 8.3
percent. The average of the reported highs in the studies was 12.3 percent, and the
average of the lows was 4.9 percent. This is similar to, though slightly higher than,
the primary U.S. benchmark developed in the 1996 Coca-Cola Research Council
sponsored study that was 8.2 percent, which was calculated as the simple average
rate of eight categories that ranged from 3.9 percent to 11.1 percent. The 2002
GMA study reported an average OOS rate of 7.4 percent with DSD categories rang-
ing from 3.2 percent (milk) to 11.2 percent (prepackaged bread).
Keep in mind that the 40 studies examined here used slightly different measurement
methods and different people, measured different categories, and examined differ-
ent durations and different daily and weekly factors. All of these can affect the
measurement. However, when all of the various factors are considered together, the
averages regress to an uncanny similarity, and this provides a sense that the ?ndings
are reliable in the aggregate, and the differences can easily be explained by vari-
ances in categories, methods and regions.
For this study, Europe was split into its northern and western region (Norway, Denmark,
Sweden, France, Belgium, Netherlands, Germany, Switzerland, Austria) and into its
southern and eastern region (Portugal, Spain, Greece, Poland, Hungary, Czech
Republic, Slovakia). Countries within each of these two areas showed similarities in
OOS rates, and differences between the two regions were substantial. Northwest
Europe showed the lowest OOS rates, while Southeast Europe showed the highest.
The Northwest examination was limited due to a lack of detailed studies from the UK,
and Finland. Summary extent numbers were reviewed from four additional studies
from the UK and the extents fell in line with those reported elsewhere in the region.
Unfortunately, study analysts were unable to review the studies in detail and thus did
not include them in the calculations in this request.
OOS rates in other regions (South America and Asia) were lower in average,
although the extents varied as much or more than other regions, and the small num-
ber of studies does not provide a complete representation of these regions.
OOS is often measured by category. A category is a microcosm of the retail store,
and category management principles encourage a focus on retail performance by
category. Of the 40 OOS studies that examined the extent of OOS, 14 of these pro-
vided reliable OOS data by category. Additional studies measured OOS by catego-
ry, but only reported the composite ?ndings and did not report by category. In total,
18 categories provided OOS results except for the GMA DSD study, which detailed
the top 25 categories. However, in only six of these 18 categories did data come
from three or more studies. Thus, the averages were computed and the OOS rates
were reported for these six categories only. Exhibit 5 illustrates the averages and
ranges of OOS for the six categories. Exhibit 6 provides a chart of the category aver-
ages only. Note that the average of the six categories is slightly lower than the over-
all worldwide average based on 40 studies.
3. OOS EXTENT
BY CATEGORY
13 Chapter 2: Overall Findings
E
x
h
i
b
i
t
5
Toilet Tissue
3 Studies
6.6
Average
6.1
Low
7.1
Diapers
6 Studies
7.0
1.9
12.0
Hair Care
6 Studies
9.8
7.0
16.0
Laundry
9 Studies
7.7
2.1
15.6
Salty Snacks
5 Studies
5.3
1.4
8.0
Fem Hygiene
4 Studies
6.8
1.9
10.2
World Average
40 Studies
18 Categories
8.3
4.9
12.3
High
0.0 3.0 6.0 9.0 12.0 15.0 18.0
Out-of-Stock Extent by Category
Percent
Overall OOS Extent (Averages)
E
x
h
i
b
i
t
6
Worldwide Avg
18 Categories
8.3
Fem Hygiene 6.8
Salty Snacks 5.3
0.0 2.0 4.0 6.0 8.0 12.0 10.0
OOS Averages by Category
Percent
Toilet Tissue 6.6
Hair Care 9.8
Laundry 7.7
Diapers 7.0
14 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Thirteen studies measured and reported variations in OOS rates by time of day
and/or day of the week. In general, there are two clear conclusions. First, consis-
tent across all studies are patterns that showed increases in OOS rates in the early
evening hours as opposed to morning or early afternoon. The highest OOS
occurred in the evenings (after 8:00 p.m.), while the lowest were during the early
afternoon. Morning rates were lower due to overnight restocking practices, slight-
ly higher than those after noon, and lower than in the evenings. The conclusion
from these ?ndings is that ordering decisions and replenishment patterns as deter-
mined by store management have an effect on OOS rates.
Second, consistent across all studies are weekly patterns where OOS rates rose
and fell on different days of the week. The 2002 GMA DSD study, which meas-
ured time-of-day and day-of-week stock-outs, also showed that same pattern,
whether or not the retailer is directly responsible for keeping the shelves stocked.
Exhibit 7 shows how the OOS rates vary during days of the week.
All of the studies that reported daily OOS rates showed the same general pattern
of decreasing rates throughout the week, but a large rate on Sunday (and the
resulting carry-over to Monday). This pattern re?ects both retail strategy and eco-
nomic realities. Assuming the weekend to be the heaviest shopping days, re-order-
ing and deliveries occur on Monday and Tuesday. Another reason that Monday
has a high average is that in some countries stores are closed on Sunday and
restocking does not begin until Monday.
4. VARIATION IN OOS
VARY BY TIME OF DAY
AND WEEK
Overall OOS Extent (Averages)
E
x
h
i
b
i
t
7
0.0 2.0 4.0 6.0 10.0 8.0 12.0
OOS by Day of Week
Percent
Sunday 10.9
Saturday 7.3
Friday 8.7
Thursday 9.1
Wednesday 9.8
Tuesday 10.0
Monday 10.9
(Average of 13 Studies)
15 Chapter 2: Overall Findings
Throughout the week, restocking and preparations for the Saturday and Sunday
promotions lead to lower OOS rates. Saturday, despite being the heaviest shop-
ping day, has the lowest OOS as retailers employ extra labor and can ?ll-in using
safety stocks for promoted items. In countries where stores are open on Sundays,
labor is normally at a lower level, and safety stocks for high demand items begin
to be depleted. Thus, any incorrect demand estimation becomes manifested in
OOS increases.
Promotional Effects.
In general, the studies that reported OOS rates on promoted and non-promoted
items consistently showed OOS rates to be higher on the promoted items. In
some cases the differences were minor while in others the differences were sub-
stantial. Although the promoted items should be receiving attention from the
retail store management, all studies that report promotional effects ?nd substan-
tially greater OOS on promoted items than everyday items.
While the differences vary among studies, in general a 2:1 ratio of promoted vs.
non-promoted OOS rates was found. Examples of this in publicly reported stud-
ies include the ECR France study (where promoted items have a 75 percent
greater OOS rates the 1996 Coca Cola U.S. study (where OOS levels of promoted
items were approximately double of non-promoted items), and the 2002 GMA
DSD study (where OOS levels of promoted items were approximately double of
non-promoted items). Several of the proprietary studies examined for this report
found similar results.
One study found that the increase in the amount of discount offered by a promo-
tion corresponded with the OOS rate. Another study highlighted a related prob-
lem where promotional decisions (and the resulting last-minute advertising
changes) based on responses to competitors led to increased OOS when the tim-
ing of the changes were too late to be included in the normal order cycle.
Velocity of Product Movement.
Somewhat overlapping with promoted items, studies that exclusively examined
fast-moving items found higher OOS rates (13 percent-15 percent) than those that
examined entire categories that include both fast-moving and slow-moving items
(8.3 percent average). This translates to a 50-80 percent higher OOS rate for fast
moving vs. all products. The GMA DSD study found that, on average, the top 10
percent of the fastest moving items accounted for 45 percent of the out-of-stocks.
The studies that examine the fast moving items used a different methodology
(scanner data analysis vs. visual audits), and thus some of the difference could be
due to variances in measurement. However, it is clear that the faster-moving
items — promoted or not — have higher OOS rates than slower-moving items.
5. VARIATION BY
PROMOTION, MOVEMENT
AND DURATION OF OOS
16 Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses
Product and Brand Effects on OOS Rates.
The sparse brand-level data available for this analysis was not adequate to make
solid conclusions about speci?c brands within categories. However, it was clear
that the faster-moving items also had more incidences of OOS, although the dura-
tion was not necessarily longer. Thus, in any category, the faster-moving SKUs are
going to incur more frequent OOS, regardless of the brand. The implication of
this – and the value of addressing the faster moving SKUs – is that the faster
movers suffer disproportionately more due to OOS than do slower-movers.
Duration of OOS.
Data on duration of OOS, while sparse, is very interesting. Based on a study of
13 stores in the U.S. by Data Ventures, a U.S. software service provider, the fol-
lowing results were found. When products become OOS, only about 20 percent
are replenished in less than eight hours while a similar percentage remain OOS
for more than three days. Duration is a critical though under-used measure for the
extent of OOS. The traditional measure of OOS (the percentage of SKUs not on
the shelf at a particular point in time) does not provide the measure that is most
meaningful from the perspective of the consumer. When the duration of the OOS
item is considered along with the extent, then a better picture for managerial
action emerges.
All of the above issues (promotion effects, velocity and duration) indicate that both
retail store management systems and practices contribute to OOS extents. While
this will be discussed in more detail with the other implications, it is important to
note that there are two ways to address the higher OOS rates on faster-moving prod-
ucts. First, retailers can pay more attention to high velocity products to ensure that
they get reordered and restocked more frequently. Second, following category man-
agement principles, retailers can examine a category and eliminate some slower-
movers and allocate more shelf space to faster-movers. According to Broniarczyk et
al.’s category management research (1998, Journal of Marketing Research, Vol. 35,
pp. 166-176) sales and customer satisfaction for the category increases following a
reduction in SKUs from a category review.
E
x
h
i
b
i
t
8
Duration of OOS
3 Days or More 19%
1 Day to