Business Forecasting

Practical Problems and Solutions

Michael Gilliland, Len Tashman, Udo Sglavo

• Publisher: Wiley
• Year: 2015
• ISBN:
  9781119228271
• Edition: 1
• Topic: Business
• Subtopic: Previsioni
• Index: Business

A comprehensive collection of the field's most provocative, influential new work

Business Forecasting compiles some of the field's important and influential literature into a single, comprehensive reference for forecast modeling and process improvement. It is packed with provocative ideas from forecasting researchers and practitioners, on topics including accuracy metrics, benchmarking, modeling of problem data, and overcoming dysfunctional behaviors. Its coverage includes often-overlooked issues at the forefront of research, such as uncertainty, randomness, and forecastability, as well as emerging areas like data mining for forecasting.

The articles present critical analysis of current practices and consideration of new ideas. With a mix of formal, rigorous pieces and brief introductory chapters, the book provides practitioners with a comprehensive examination of the current state of the business forecasting field.

Forecasting performance is ultimately limited by the 'forecastability' of the data. Yet failing to recognize this, many organizations continue to squander resources pursuing unachievable levels of accuracy. This book provides a wealth of ideas for improving all aspects of the process, including the avoidance of wasted efforts that fail to improve (or even harm) forecast accuracy.

• Analyzes the most prominent issues in business forecasting
• Investigates emerging approaches and new methods of analysis
• Combines forecasts to improve accuracy
• Utilizes Forecast Value Added to identify process inefficiency

The business environment is evolving, and forecasting methods must evolve alongside it. This compilation delivers an array of new tools and research that can enable more efficient processes and more accurate results. Business Forecasting provides an expert's-eye view of the field's latest developments to help you achieve your desired business outcomes.

CHAPTER 1
Fundamental Considerations in Business Forecasting

Challenges in business forecasting, such as increasing accuracy and reducing bias, are best met through effective management of the forecasting process. Effective management, we believe, requires an understanding of the realities, limitations, and principles fundamental to the process. When management lacks a grasp of basic concepts like randomness, variation, uncertainty, and forecastability, the organization is apt to squander time and resources on expensive and unsuccessful fixes: There are few other endeavors where so much money has been spent, with so little payback.

This chapter provides general guidance on important considerations in the practice of business forecasting. The authors deal with:

• Recognition of uncertainty and the need for probabilistic forecasts
• The essential elements of a useful forecast
• Measurement of forecastability and bounds of forecast accuracy
• Establishing appropriate benchmarks of forecast accuracy
• The importance of precisely defining demand when making demand forecasts
• Guidelines for improving forecast accuracy and managing the forecasting function

■ ■ ■

Although we were unable to secure rights to include it in this book, Makridakis and Taleb’s “Living in a World of Low Levels of Predictability” from the International Journal of Forecasting is an important piece worth mentioning in any consideration of fundamental issues.

Spyros Makridakis is very well recognized as lead author of the standard forecasting text, Forecasting: Methods and Applications, and of the M-series forecasting competitions. Through his books, Fooled by Randomness and The Black Swan, Nassim Nicholas Taleb has drawn popular attention to the issue of unforecastability of complex systems, and made “black swan” a part of the vernacular. Their article, published in the International Journal of Forecasting (2009), speaks to the sometimes disastrous consequences of our illusion of control—believing that accurate forecasting is possible.

While referring to the (mostly unforeseen) global financial collapse of 2008 as a prime example of the serious limits of predictability, this brief and nontechnical article summarizes the empirical findings for why accurate forecasting is often not possible, and provides several practical approaches for dealing with this uncertainty. For example, you can’t predict when your house is going to burn down. But you can still manage under the uncertainty by buying fire insurance.

So why are the editors of a forecasting book so adamant about mentioning an article telling us the world is largely unforecastable? Because Makridakis and Taleb are correct. We should not have high expectations for forecast accuracy, and we should not expend heroic efforts trying to achieve unrealistic levels of accuracy.

Instead, by accepting the reality that forecast accuracy is ultimately limited by the nature of what we are trying to forecast, we can instead focus on the efficiency of our forecasting processes, and seek alternative (nonforecasting) solutions to our underlying business problems. The method of forecast value added (FVA) analysis (discussed in several articles in Chapter 4) can be used to identify and eliminate forecasting process activities that do not improve the forecast (or may even be making it worse). And in many situations, large-scale automated software can now deliver forecasts about as accurate and unbiased as anyone can reasonably expect. Plus, automated software can do this at relatively low cost, without elaborate processes or significant management intervention.

For business forecasting, the objective should be:

To generate forecasts as accurate and unbiased as can reasonably be expected—and to do this as efficiently as possible.

The goal is not 100% accurate forecasts—that is wildly impossible. The goal is to try to get your forecast in the ballpark, good enough to help you make better decisions. You can then plan and manage your organization effectively, and not squander resources doing it.

1.1 Getting Real about Uncertainty¹

Paul Goodwin

---

Business forecasters tend to rely on the familiar “point” forecast—a single number representing the best estimate of the result. But point forecasts provide no indication of the uncertainty in the number, and uncertainty is an important consideration in decision making. For example, a forecast of 100 ± 10 units may lead to a much different planning decision than a forecast of 100 ± 100 units.

In this opening article, Paul Goodwin explores the types of “probabilistic” forecasts, the academic research behind them, and the numerical and graphical displays afforded through prediction intervals, fan charts, and probability density charts. Providing uncertainty information, he explains, can result in better decisions; however, probabilistic forecasts may be subject to misinterpretation and may be difficult to sell to managers. There is also an unfortunate tendency we have to seriously underestimate the uncertainty we face and hence overstate our forecast accuracy.

Goodwin’s article provides practical recommendations and additional sources of guidance on how to estimate and convey the uncertainty in forecasts.

---

Avoiding Jail

In October 2012, the scientific world was shocked when seven people (engineers, scientists, and a civil servant) were jailed in Italy following an earthquake in the city of L’Aquila in which 309 people died. They had been involved in a meeting of the National Commission for Forecasting and Preventing Major Risks following a seismic swarm in the region. At their trial, it was alleged that they had failed in their duty by not properly assessing and communicating the risk that an earthquake in the area was imminent. Their mistake had been that they had simply conveyed the most likely outcome—no earthquake—rather than a probabilistic forecast that might have alerted people to the small chance of a strong earthquake (Mazzotti, 2013).

Point versus Probabilistic Forecasts

This case dramatically highlights the problem with forecasts that are presented in the form of a single event or a single number (the latter are called point forecasts). They give no information on how much uncertainty is associated with the forecast. As such, they provide no guidance on what contingency plans you should make to cope with errors in the forecasts. Is the risk of an earthquake sufficient to evacuate a...