13-1: Demonstrate the importance of forecasting for business operations.

Forecasting involves generating estimates of future values for some variable or a scenario that corresponds to some future event. Demand forecasting is a critical and fundamental process for any business and is essential to the strategic, tactical, and operational planning for a company and its supply chain. Effective forecasts need to be accurate, consistent, timely, simple, and efficient.

13-2: Illustrate and distinguish between qualitative and quantitative types of forecasting methods, including their strengths and weaknesses.

Forecasting methods can broadly be classified into two categories: qualitative and quantitative methods. Qualitative methods are typically used for forecasting for the long term or if no measurable, reliable, historical, or statistical data are available. Some common qualitative forecasting methods are expert opinion, Delphi method, sales force opinions, market research, and historical life-cycle analogy.

Quantitative forecasting methods are typically used for the short-term to medium-term time frame and if measurable, historical data are available and if evidence that past behavior of demand is indicative of its future behavior. Techniques such as moving averages, exponential smoothing, and regression methods are some of the more commonly used quantitative forecasting methods.

13-3: Recognize the characteristics of good forecasts.

Good forecasts have several characteristics that demonstrate their quality. For a forecast to be useful to its organization, it must be (a) accurate— any forecast errors should be small; (b) consistent—they should demonstrate reliability in tracking actual demand; (c) timely—to be useful, forecasts should be available within some minimal time period to give decision makers opportunities to adjust their options; (d) simple—forecasts should be easy to interpret; and (e) efficient— the costs of preparing the forecast should not outweigh its benefits. 

13-4: Use the four forecast error measures to track forecast accuracy.

The best forecasting method is the one that consistently minimizes forecasting errors. Several forecast error measures are used to track the performance of forecasting methods. The mean absolute deviation (MAD) is one of the easiest forecasting error measures to compute. The MAD is the average of the sum of the absolute differences between the actual and the forecasted demand values. The cumulative sum error (CSE) is the sum of the differences between the actual and the forecasted demand values. The mean squared error (MSE) is the average of the sum of the squared differences between the actual and the forecasted demand values. The mean absolute percentage error (MAPE)

13-5: Employ the methods used to monitor and control forecasts.

Companies typically use two methods for monitoring and controlling forecasts: tracking signals and control charts. The use of tracking signals involves establishing an upper and a lower control limit to determine whether the forecasting errors related to a method are within these limits. A tracking signal value that goes outside of these control limits is an indicator that the forecasting method being used should be modified or changed. The second tool that can be used in monitoring and controlling forecasts is a control chart. Constructing a control chart is similar to calculating a tracking signal except we use the standard deviation of forecast errors to construct the control limits and plot the computed forecasting errors on the control chart. 

13-6: Identify the steps involved in forecasting for supply chains.

Forecasting important information within the broader context of the organizational supply chain is a challenging and complicated undertaking. The effectiveness of a demand planning system depends on its ability to generate forecasts, not only at the individual product level but also at product group, customer group, and regional levels, as well as for different planning horizons. The relevant steps for forecasting for supply chains include (a) determining the purpose of the forecast, (b) collecting and cleaning historical data, (c) selecting an appropriate forecasting technique and generating the demand forecasts, (d) adjusting the forecast with judgmental inputs, (e) adjusting the baseline forecast for marketing promotions, (f) sharing the forecast with suppliers and downstream customers, (g) generating a single number forecast, and (h) monitoring the forecast.

13-7: Illustrate the role ethics and ethical decision-making can play in selecting and using forecasting models.

Forecasts can contribute to ethical decision-making in two ways: We use ethics to make our forecasts, and ethics affect the results of our forecasting efforts. The ethical challenge that lies at the heart of building such a tool is deciding which predictors, or fact variables, are acceptable to use. Ethics influence decisions companies make after they developed and used these prediction models. Their challenge lies in determining what to do with the data they have generated, particularly when the information may contain bad news or place the firm in a morally ambiguous situation.

› Chapter Outline