Business Forecasting - Practical Problems and Solutions

Business Forecasting

Contents

Foreword

Preface

Chapter 1 Fundamental Considerations in Business Forecasting

1.1 Getting Real about Uncertainty (Paul Goodwin)

Avoiding Jail

Point versus Probabilistic Forecasts

Is it Worth Communicating Uncertainty?

Limitations of Probabilistic Forecasts

What Is the Best Way to Convey Uncertainty?

Estimating Uncertainty

Conclusions

REFERENCES

1.2 What Demand Planners Can Learn from the Stock Market (Charles K. Re Corr)

Why Forecast the Future Market

What Makes Any Forecast Useful?

Some Errors Are More Forgiving than Others

1.3 Toward a More Precise Definition of Forecastability (John Boylan)

Stability versus Forecastability

Defining Forecastability in Terms of Forecast Error

Upper Bound of a Forecasting Error Metric

Lower Bound of a Forecasting Error Measure

Finding More Forecastable Series

Conclusions

REFERENCES

1.4 Forecastability: A New Method for Benchmarking and Driving Improvement (Sean Schubert)

Introduction: Establishing Comparability

What Is Forecastability?

Forecastability DNA

Building a Model of Forecastability

The Forecastability Model in Action

Conclusions

REFERENCES

1.5 Forecast Errors and Their Avoidability (Steve Morlidge)

Beginnings

Defining Success in Forecasting

Creating a Metric

What the Experts Say

Avoidability

The Way Forward: A Conjecture

The Unavoidability Ratio

The Empirical Evidence

The Next Step

REFERENCES

1.6 The Perils of Benchmarking (Michael Gilliland)

Danger, Danger

1.7 Can We Obtain Valid Benchmarks from Published Surveys of Forecast Accuracy? (Stephan Kolassa)

Introduction

Published Surveys of Forecast Accuracy

What Is a Benchmark?

Problems with Forecast Accuracy Surveys

External vs. Internal Benchmarks

REFERENCES

1.8 Defining “Demand” for Demand Forecasting (Michael Gilliland)

Introduction: Unconstrained vs. Constrained Demand

Orders vs. True Demand

Shipments and Sales vs. True Demand

Seeking an Operational Definition of True Demand

True vs. Constrained Forecasts

Assessing Forecast Accuracy and Making Financial Projections

REFERENCES

1.9 Using Forecasting to Steer the Business: Six Principles (Steve Morlidge)

Economic Forecasting Is Broken

The Narrow Focus of the Forecasting Profession

Prescription for Change

Forecasting to Steer the Business: Six Principles

Conclusions

REFERENCES

1.10 The Beauty of Forecasting (David Orrell)

Introduction

Perfect Model

Economy at Risk

Lessons from Business Forecasting

REFERENCES

Chapter 2 Methods of Statistical Forecasting

2.1 Confessions of a Pragmatic Forecaster (Chris Chatfield)

Introduction

Some History

Forecasting Methods

Which Method Is Best?

Implementation of Forecasting Methods

Publication Bias

Experience in Consulting

REFERENCES

2.2 New Evidence on the Value of Combining Forecasts (Paul Goodwin)

Forecast Combination and the Bank of England’s Suite of Statistical Forecasting Models

Why Did Combining Work?

Trimmed Means

Conclusions

REFERENCES

2.3 How to Forecast Data Containing Outliers (Eric Stellwagen)

Option #1: Outlier Correction

Option #2: Separate the Demand Streams

Option #3: Use a Forecasting Method that Models the Outliers

Summary

2.4 Selecting Your Statistical Forecasting Level (Eric Stellwagen)

Can It Be Simpler?

Do You Have Enough Structure?

Are the Relationships Between Levels Changing in Time?

Summary

2.5 When Is a Flat-line Forecast Appropriate? (Eric Stellwagen)

2.6 Forecasting by Time Compression (Udo Sglavo)

Introduction

The Challenge

Approach 1: Traditional Forecasting

Approach 2: Forecasting by Time Compression

Conclusion

Acknowledgments

REFERENCES

2.7 Data Mining for Forecasting: An Introduction (Chip Wells and Tim Rey)

Introduction, Value Proposition, and Prerequisites

Big Data in Data Mining for Forecasting

NOTES

REFERENCES

2.8 Process and Methods for Data Mining for Forecasting (Chip Wells and Tim Rey)

Time-Series Data Creation

Two Phases of Data Mining for Forecasting

Three Methods for Data Mining in Time Series

NOTE

REFERENCES

2.9 Worst-Case Scenarios in Forecasting: How Bad Can Things Get? (Roy Batchelor)

Introduction

A Standard View of New Car Sales

Time-Varying Volatility in Car Sales

Concluding Remarks

Appendix: The GARCH Model

2.10 Good Patterns, Bad Patterns (Roy Batchelor)

Introduction

Good Patterns

Bad Patterns

Conclusion

REFERENCES

Chapter 3 Forecasting Performance Evaluation and Reporting

3.1 Dos and Don’ts of Forecast Accuracy Measurement: A Tutorial (Len Tashman)

The Most Basic Issue: Distinguish In-Sample Fit from Out-of-Sample Accuracy

Assessing Forecast Accuracy

Accuracy Metrics

Benchmarking and Forecastability

Costs of Forecast Error

REFERENCES

3.2 How to Track Forecast Accuracy to Guide Forecast Process Improvement (Jim Hoover)

Introduction

Obstacles to Tracking Accuracy

Multistep Tracking Process

Conclusions and Recommendations

REFERENCES

3.3 A “Softer” Approach to the Measurement of Forecast Accuracy (John Boylan)

The Complement of Mean Absolute Percent Error

Forecast Researchers and Practitioners: Different Needs and Perspectives

The Soft Systems Approach

Relevant Systems and Root Definitions

Effectiveness Measures and Accuracy Measures

Using a Structured Approach in Practice

Postscript: Advice on the Complement of MAPE

REFERENCES

3.4 Measuring Forecast Accuracy (Rob Hyndman)

Training and Test Sets

Forecast Accuracy Measures

Time-Series Cross-Validation

Conclusions

REFERENCES

3.5 Should We Define Forecast Error as e = F - A or e = A - F? (Kesten Green and Len Tashman)

The Issue

The Survey

Support for A – F

Support for F – A

3.6 Percentage Error: What Denominator? (Kesten Green and Len Tashman)

The Issue

Survey Results

REFERENCES

3.7 Percentage Errors Can Ruin Your Day (Stephan Kolassa and Roland Martin)

Introduction

Rolling Dice

Variants of the APE

Alternatives to the APE and Its Variants

Conclusion

REFERENCES

3.8 Another Look at Forecast-Accuracy Metrics for Intermittent Demand (Rob Hyndman)

Introduction: Three Ways to Generate Forecasts

An Example of What Can Go Wrong

Measurement of Forecast Errors

REFERENCES

3.9 Advantages of the MAD/Mean Ratio over the MAPE (Stephan Kolassa and Wolfgang Schütz)

The MAD, the MAPE, and the MAD/Mean

The MAD/Mean Ratio as a Weighted MAPE

The Issue of Forecast Bias

The MASE

The Case of Intermittent Series

Recap

REFERENCES

3.10 Use Scaled Errors Instead of Percentage Errors in Forecast Evaluations (Lauge Valentin)

Evaluating Forecasts in the LEGO Group

Problems with Percentage Errors

Scaled Errors

The GMASE

The Problem of Bad Forecasts

Perspectives for Intercompany Benchmarking

How to Make MASE and GMASE Management Friendly

Summary

REFERENCES

3.11 An Expanded Prediction-Realization Diagram for Assessing Forecast Errors (Roy Pearson)

Introduction

The Prediction-Realization Plot

The 17 Possible Outcomes

Predicted Changes vs. Predicted Levels

Applying the PRD to Energy Price Forecasts

Payroll Employment Forecasts: The Difficulty of Improvingon a Naïve Forecast

Summary and Recommendation

REFERENCES

3.12 Forecast Error Measures: Critical Review and Practical Recommendations (Andrey Davydenko and Robert Fildes)

1. Introduction

2. Data

3. Critical Review of Existing Measures

4. Recommended Scheme for Measuring the Accuracy of Point Forecasts across Many Series

5. Results of Empirical Evaluation

6. Conclusions

REFERENCES

3.13 Measuring the Quality of Intermittent Demand Forecasts: It’s Worse than We’ve Thought! (Steve Morlidge)

Introduction

The Problems with Intermittent Demand

The Numerator Problem

Solutions to the Numerator Problem

The Bias-Adjusted Error

Conclusion

Appendix

REFERENCES

3.14 Managing Forecasts by Exception (Eric Stellwagen)

What Is an Exception Report?

How Do I Select the Thresholds?

3.15 Using Process Behavior Charts to Improve Forecasting and Decision Making (Martin Joseph and Alec Finney)

Introduction

Data to Information to Insight

Control Limits to Distinguish Signals from Noise

Application of PBCs to Forecasting

Applying PBCs for Decision Making

Bringing PBCs into S&OP and Other Planning Activities

REFERENCES

3.16 Can Your Forecast Beat the Naïve Forecast? (Shaun Snapp)

Background

What to Expect?

How Long Should You Test the Naïve Forecast Against the Current Live Forecast?

System Implications

Conclusion

REFERENCES

Chapter 4 Process and Politics of Business Forecasting

4.1 FVA: A Reality Check on Forecasting Practices (Michael Gilliland)

Introduction

Calculating Forecast Value Added

How Organizations Are Using FVA

Which Naïve Model to Use?

A Reality Check on Forecasting Practices

REFERENCE

4.2 Where Should the Forecasting Function Reside? (Larry Lapide)

Executing an Effective Forecasting Process

Evaluation Criteria

The Pros and Cons of Departments

Conclusion

4.3 Setting Forecasting Performance Objectives (Michael Gilliland)

Five Steps for Setting Forecasting Performance Objectives

REFERENCES

4.4 Using Relative Error Metrics to Improve Forecast Quality in the Supply Chain (Steve Morlidge)

Introduction

Background

The Practical Challenge

Focus the Efforts

Devise Improvement Strategies

Setting Realistic Targets

Conclusion

REFERENCES

4.5 Why Should I Trust Your Forecasts? (M. Sinan Gönül, Dilek Önkal, and Paul Goodwin)

Introduction

Trust and Forecasting

The Determinants of Trust

Trust and Adjustments to Provided Forecasts

The Need for Open Communication Channels

Working to Earn Trust

REFERENCES

4.6 High on Complexity, Low on Evidence: Are Advanced Forecasting Methods Always as Good as They Seem? (Paul Goodwin)

The Complexity Love Affair

A Case in Point

Proper Testing of Accuracy

Inadequate Evidence

Conclusions

REFERENCES

4.7 Should the Forecasting Process Eliminate Face-to-Face Meetings? (J. Scott Armstrong)

Introduction

The Wisdom of Crowds

Face-to-Face Meetings Could Be Effective

The Case Against Face-to-Face Meetings

Alternatives to Face-to-Face Meetings: Markets, Nominal Groups, and Virtual Teams

A Prediction Case

Are Face-to-Face Meetings Useful Under Some Conditions?

Action Steps

Conclusions

REFERENCES

4.8 The Impact of Sales Forecast Game Playing on Supply Chains (John Mello)

Introduction

The Nature of Supply Chains

Games People Play

Consequences for the Supply Chain

Conditions Fostering Game Playing

How to Control Game Playing

Conclusion

REFERENCES

4.9 Role of the Sales Force in Forecasting (Michael Gilliland)

Three Assumptions About Salespeople

Gathering Sales Force Input

Can Salespeople Forecast Their Customers’ Behavior?

Can You Trust the Forecast from a Salesperson?

Compensation as an Incentive for Honesty

Does Improving Customer Level Forecasts Always Matter?

Commitments Are Not Forecasts

Conclusions

NOTES

REFERENCES

4.10 Good and Bad Judgment in Forecasting: Lessons from Four Companies (Robert Fildes and Paul Goodwin)

Introduction

Adjustments Galore

The Illusion of Control

When Do Adjustments Improve Accuracy and When Do TheyNot?

The Importance of Definitions

History Is Not Bunk

Conclusions

REFERENCES

4.11 Worst Practices in New Product Forecasting (Michael Gilliland)

Unrealistic Accuracy Expectations

Reverse Engineering the Forecast

Cherry-Picking Analogies

Insisting on a Wiggle

The Hold-and-Roll

Ignoring the Product Portfolio

Using Inappropriate Methods

REFERENCES

4.12 Sales and Operations Planning in the Retail Industry (Jack Harwell)

Sales and Operations Planning

Three Plans

Three Levels

Sales and Operations Planning Escalation Process

Other Keys to S&OP Success

Goals and Key Performance Indicators

Top-Level Support

Conclusion

4.13 Sales and Operations Planning: Where Is It Going? (Tom Wallace)

Summary

REFERENCE

About the Editors

Index

EULA