Logistics Handbook

James F. Robeson

• Publisher: Free Press
• Year: 1994
• ISBN:
  9781439106259
• Topic: Business
• Subtopic: Supply Chain Management
• Index: Business

The Logistics Handbook encompasses all of the latest advances in warehousing and distribution. It provides invaluable "how to" problem-solving tools and techniques for all the ever-increasing logistical problems managers face -- making it the most complete and authoritative handbook to date.Special features include: * The most in-depth coverage of a wide range of topics, including information systems, benchmarking, and environmental issues* Contributions found nowhere else from the leading executives, consultants, and academics in the field, such as C. John Langley, James Heskett, and David Anderson* State of the art graphics* Information-packed appendixes of logistics publications and organizationsThis all-inclusive reference will enable the next generation of managers to thoroughly integrate their logistics operations at all levels -- strategic, structural, functional, and implementation -- into a comprehensive logistics strategy.

SECTION IV

Materials Management

Joseph L. Cavinato

Raw materials, supplies, and finished goods are the physical essence of virtually all logistics systems. Hence, thoughtful management of these materials is crucial to the success of most logistics operations. This section addresses topics related to effective materials management including forecasting, determining inventory investment, managing suppliers, planning, maintaining control, manufacturing, and packaging. Each chapter provides insightful discussion for managers seeking to understand materials management and improve the efficiency and effectiveness of their investments in materials and associated operations.

In Chapter 14, Donald B. Rosenfield identifies the types and nature of demand forecasts. Demand forecasts are important not only for determining production and stocking levels but also for making broad decisions such as those involving network configuration, suppliers, and marketing. The chapter discusses the most important demand patterns and forecasting methods and concludes with examples of effective demand forecasting.

In Chapter 15, Alan J. Stenger focuses on one of the main questions in materials management: How much inventory should I hold? The chapter discusses the delicate balance between investment and availability, describes ways to reduce inventories, and concludes with a detailed model for determining inventory quantities. Stenger’s novel framework encourages managers to precede mathematics with business considerations of the purposes for inventory.

In Chapter 16, William L. Grenoble IV discusses the practicalities of maintaining control over inventory. The major clerical issues of record keeping, control systems, and counting are reviewed along with the managerial issues of analysis and performance measurement. Grenoble provides an efficient and understandable lesson in a very broad and complex subject.

In Chapter 17, Alan J. Stenger explains distribution resource planning (DRP) using a running case study to describe the concepts and processes of this continuous, comprehensive, and integrated planning tool. While simple systems work well with traditional inventory planning techniques, complex systems require greater sophistication to improve profits and provide accurate information for other functions like manufacturing.

In Chapter 18, Joseph L. Cavinato explores current and evolving practices in purchasing. As the scope of supply chain management broadens, companies are developing long-term supply relationships. Cavinato discusses these changes and shows how purchasing can move away from its traditionally isolated role as a support function toward a strategic role that provides competitive advantage.

In Chapter 19, R. John Aalbregtse and Roy L. Harmon discuss manufacturing in a just-intime (JIT) environment, an issue that is dramatically affecting many logistics organizations. The chapter focuses on cellular manufacturing, quick changeovers, and pull scheduling. These practices allow manufacturers to improve flexibility and respond to direct customer demand, thereby reducing inventories, improving product quality, and enhancing customer service.

In Chapter 20, Diana Twede reviews the packaging required for manufacturing, shipping, handling, and storage. The chapter provides an overview of logistical packaging as well as a review of typical packaging challenges. Twede notes that most packaging in use today is archaic and that innovative approaches can reduce costs while increasing customer service by simplifying unpacking and reducing damage and waste.

CHAPTER 14

Demand Forecasting

Donald B. Rosenfield

Demand forecasting is concerned with the projection of demand in the future that is critical to successful logistics management.¹ Demand for products and services by end users, as well as by intermediate supply chain members, drives many basic strategic and operating decisions. Demand forecasting’s primary goal is to have the right products in the right positions at the right time. However, its importance goes well beyond the need to target inventories, which is the application on which most practitioners focus. For example, longer-term forecasts will impact decisions on the network of manufacturing and logistics facilities, contracts with third-party suppliers, and other long-range logistics decisions. Intermediate-term forecasts will affect policies for dealing with seasonal buildups and intermediate workforce plans and contracts.

Forecasting is multifaceted. Demand is not a simple matter of a single prediction. Forecasts cover multiple time periods, multiple products or product groups, multiple geographic areas, and often multiple customer groups. They focus on predictions of demand as well as ranges and errors in demand. It is a true art to identify those dimensions that are crucial for logistics success and to develop systems to encompass the associated needs.

The purpose of this chapter is to describe the types and nature of demand forecasts and how they affect logistics operations and strategies, to discuss the most important patterns of demand, to describe methods and systems for demand forecasting, and to present examples that illustrate principles of effective demand forecasting.

Understanding and Applying Forecasts

Operating Principles for Demand Forecasting

Effective logistics practice and strategy can be based on several forecasting principles. These principles are discussed below and include the following:

• Good forecasts will still have significant errors.

• Forecasting requires monitoring and estimation of errors.

• One should expect and account for large uncertainties.

• Any forecasting system is based on either an implicit or explicit model.

• Effective forecasting is often based on aggregate forecasts broken down into product, geographical, or other components.

• Forecast errors are correlated in time and among geographical regions, and follow predictable relationships with time and aggregate forecast level.

EVEN “GOOD” FORECASTS MAY HAVE SIGNIFICANT ERRORS

It is usually not possible to make demand estimates with great accuracy, for two reasons. First, certain factors having significant effects on demand are subject to great uncertainty. For example, in considering the stocking policies for air conditioners, the merchandiser of a major retailer suggests, from years of experience, that the single most important impact on aggregate demand is the weather in April. Many demand processes can be affected by factors that only marginally affect the need or quality of the good or service. A similar phenomenon is that demand at New England ski areas depends to a great deal on the snowfall in the metropolitan Boston area, despite the nature of the snow in the mountains. Such factors introduce significant uncertainty in the forecast of demand, and often at a time when it is difficult to plan appropriately.

The second reason is that the detail of forecasts, no matter how sophisticated they may be, reduces them to a level of significant natural uncertainty. Consider a major national retailer or manufacturer. Forecast detail at the product and store or customer zone level for an appropriate time period (say, a week) will reduce the forecast to possibly only a few units. This natural level of uncertainty is based on the Poisson process, which stipulates that the standard deviation of demand is equal to the square root of the expected demand. If a retail store expects to sell four bicycles during a sale week (even supposing that the store can account for factors that affect sales, such as economic conditions, recent store sales, the weather, and so forth, with great accuracy), then the standard deviation of that demand is two units. Because plus or minus two standard deviations is not unusual, demand can vary between zero and eight bicycles per week! And this is with a “perfect” forecasting system. Indeed, in the actual application, variations are typically three times greater than the natural Poisson variation. Two additional principles follow as a result of the extent of these uncertainties.

FORECASTING REQUIRES MONITORING AND ESTIMATION OF ERRORS

For tasks like setting inventory targets and understanding trade-offs between inventory and transportation, one needs to understand how to translate ranges of errors into policies. The errors need to be tracked and converted into the language of statistics and probability. A logistics planning system consequently should be based on likelihood and probabilities, not on the “best” forecast. Because significant errors are possible, planners need to analyze the possibilities of demand. For example, inventory target points should be adjusted for the possible ranges of the distribution, which in turn requires the estimation of errors.

EXPECT AND ACCOUNT FOR LARGE UNCERTAINTIES

Stated another way, competitive leverage requires maximizing information for—and minimizing the time required in—forecasting situations. Because forecasting is so prone to error, planning and strategy require reducing the impact of this error as much as possible. Two strategies for doing this are, first, to eliminate errors (beyond those caused by the natural error of uncertainty) by accounting for as much causal information through information technology, and, second, to minimize the time horizon over which the forecast must be made. Because forecast errors increase with time, reduction of the forecast period can effectively reduce the magnitude of the forecast error. Implementation of this strategy is based on effective information technology and logistics. In simple terms, the logistics system needs to respond as quickly as possible to demand information within the value chain. Effective retailing, for example, requires flexibility in responding to demand trends in individual stores. One retailer used a strategy of holding merchandise at the central warehouse during promotions until early sales information indicated which stores were moving ahead of plan. This reduction of response time is a major competitive weapon in both logistics and manufacturing, and is essential for minimizing the effects of uncertainty.

ALL FORECASTING IS BASED ON EITHER AN IMPLICIT OR EXPLICIT MODEL

Forecasting demand requires understanding and determining the underlying process that produces this demand. Like any aspect of logistics strategy, the forecasting decisions make implicit assumptions about the drivers of demand. What appears to be a simple or straightforward assumption about future demand may be unrealistic and significantly constrain the logistics system. For example, a constant forecast is precisely that, and there is no allowance for the effect of major external events or long term business patterns. Linkage of a forecast to some other process (for example, tennis court usage at a resort in relationship to projected guest stays) makes a very clear assumption about the drivers of that process.

The implication is that effective forecasting requires (1) an understanding of the drivers of demand, and (2) the development of a model for that demand. This model can be fairly simple, but it serves two important purposes: first, it reduces forecast errors as much as possible; second, it improves the forecaster’s understanding about the drivers of the demand process.

The underlying models for the demand process can incorporate statistical features, such as trends or cycles, as well as dependencies on other variables, such as the dependence of automobile parts on the production of automobiles.

BREAKING AGGREGATE FORECASTS INTO SMALLER COMPONENTS

Forecasts require a great deal of detail: different geographical regions, different products, or different channels. Building up forecasts from data at these detailed levels is subject to a great deal of uncertainty. From a statistical point of view, because the coefficient of variation decreases as the total demand increases (the so-called laws of large numbers), an aggregate forecast will be more accurate on a percentage basis. From this perspective, it makes sense for an organization to develop an aggregate forecast for all products and all channels or geographic regions, and then disaggregate this forecast into its separate components according to some specific measure. This...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Credits
• Dedication
• Contents
• Preface
• Section I Perspectives on Logistics Management Bernard J. La Londe, Section Editor
• Section II Strategic Logistics Planning Kevin A. O’Laughlin, Section Editor
• Section III Logistics Quality and Productivity Douglas M. Lambert
• Section IV Materials Management Joseph L. Cavinato
• Section V Transportation Management C. John Langley, Jr.
• Section VI Distribution Facilities Management Thomas W. Speh
• Section VII International Logistics Management David L. Anderson, Section Editor
• Section VIII Logistics Information Systems Donald J. Bowersox
• Section IX Logistics Organizations and Human Resources Management Jonathan L. S. Byrnes
• Section X Contemporary Issues in Logistics R. William Gardner
• Appendix A. Publications and Other Sources of Information
• Appendix B. Trade and Professional Organizations of Interest to Logistics Management Personnel
• About the Contributors
• Index