FORESIGHT, Issue 7

Description

Summer 2007 Issue

Special Feature: Cost of Forecast Error: New Perspectives Issues

• Assessing The Cost Of Forecast Error A Practical Example Peter Maurice Catt
  Peter provides a detailed tutorial on the calculation of the costs associated with forecast errors. His procedure considers inventory costs, including safety stock, as well as the costs of lost sales attributable to poor service (out-of-stock). He shows how the cost of forecast error (CFE) can be used to determine appropriate safety stock levels.
• Key Assumptions In Calculating The Cost Of Forecast Error John Boylan
  The significance of the Cost of Forecast Error (CFE) concept lies in the recognition that measures such as the Mean Absolute Error are not attuned to the criteria of decision makers. In a supply chain context, service and cost are generally the dominant factors. Of course, statistical measures still have their place, particularly for diagnosis of forecast errors.
  • Commentaries
    • Use Of The Normal Distribution In Calculating The Cost Of Forecast Error Thomas R. Willemain
    • Supply Risk And Costing Challenges Michael E. Smith
    • Lost Sales And Customer Service Scott Roy
• Reply To Cost Of Forecast Error Commentaries Peter M. Catt

Articles

• 1. Assessing The Cost Of Forecast Error A Practical Example by Peter Maurice Catt
     Peter provides a detailed tutorial on the calculation of the costs associated with forecast errors. His procedure considers inventory costs, including safety stock, as well as the costs of lost sales attributable to poor service (out-of-stock). He shows how the cost of forecast error (CFE) can be used to determine appropriate safety stock levels.
  2. An Expanded Prediction-Realization Diagram For Assessing Forecast Errors by Roy Pearson
     Forty years ago, Henri Theil created a prediction realization diagram to compare forecasts with the actual changes that were realized. The diagram emphasizes an element of accuracy that is not accounted for in traditional metrics – the accuracy with which you forecast the correct direction of change. Roy expands upon the original diagram to incorporate evaluation of whether the forecasts are improvements upon the standard benchmark of naïve (no-change) forecasts.
  3. Use Scaled Errors Instead Of Percentage Errors In Forecast Evaluations by Lauge Valentin
     Lauge Valentin presents a case for abandoning percentage errors when evaluating forecasts and replacing them by scaled errors. He describes how the shift from percentage errors to scaled errors was motivated by his company’s need for an accuracy statistic that would lend itself to benchmarking across product groups. Lauge shows how scaled error measures are used in the LEGO Group for evaluating forecasting performance.
  4. S&OP, Forecasting, And The Knowledge-Creating Company by John Mello And Terry Esper
     Organizational knowledge can serve as a competitive advantage for companies. John and Terry draw examples from audits of the forecasting process at many companies to demonstrate how firms can use sales forecasting and sales and operations planning to gather and refine information about changing business environments, create organizational knowledge, and transform that knowledge into actionable plans.
  5. Decision-Directed Forecasting For Major Disruptions The Impact Of 9/11 On Las Vegas Gaming Revenues by Stephen Custer And Don Miller
     When a major disruption such as 9/11 occurs, managers don’t know what the future holds and may have to put off important decisions while awaiting future data. In many cases, post-disruption conditions are unprecedented, and neither management’s prior experience nor traditional extrapolation methods are of much value. Steve and Don propose a new procedure, decision-directed forecasting, that provides a rational basis for evaluating decision options as new data become available. The decision maker using it is less prone to making a premature decision and better able to recognize when the postdisruption data support a decision option. They use the dramatic fall in Las Vegas gaming revenue after 9/11 to illustrate their approach.
  6. How To Get Good Forecasts From Bad Data by Ellen Bonnell
     The fundamental challenge of forecasting is that regardless of the quality of the inputs, a forecast must be created . No matter what the cause, less-than-perfect data is all you’ve got to work with. My experience has taught me that there are fundamental principles or guidelines toward better forecasts.
  7. The Forecaster As Leader Of The Forecasting Process by James Borneman
     From his vantage point as Senior Director of U.S. Forecasting for a major pharmaceutical firm, Jim Borneman analyzes what it takes for the pharmaceutical forecaster to successfully manage the forecasting process. He writes that the requisite skills far surpass the technical aspects of forecasting; indeed, the key attribute is leadership. Forecasters within and outside the pharmaceutical industry will recognize in this discussion the diverse demands and difficult challenges imposed on the business forecaster.
  8. Forecasting Software: A Progress Report For The First Seven Years Of The 21st Century by Jim Hoover
     This article won’t provide an evaluation of today’s new software against the scoring criteria. What it will do is describe the major changes in the computing environment and consider how these changes have facilitated implementation of best practices. I’ll also provide my personal assessment of whether forecasting software has taken advantage of these changes to upgrade its implementation of best practices.
  9. Supermarket Forecasting Check Out Three New Approaches by Paul Goodwin
     Supermarkets have particularly challenging forecasting problems. They need to make daily forecasts of demand for tens of thousands of products, and many of these products exhibit volatile demand patterns and low average-sales volumes. Forecasting tasks this large require a high level of automation.