Build a solid foundation in forecasting.
Click on a chapter below for details.
In the Introduction, we review the scope of this course, and provide some perspectives on sales forecasting. Specific topics include: why we don’t use forecasting formulas in this course; the need for comfort with numbers and Excel; the forecasting learning curve; the best attitude for forecasting; recent forecasting disasters; and why forecasting is so important today.
The naive forecast is your forecast error benchmark. All forecasts are compared to the naive forecast. If you can beat the naive forecast, you are adding value as a forecaster. We start with a definition and example of the naive forecast. Then we show you how to calculate MANE (mean absolute naive error) and MASE (mean absolute scaled error). These two calculations provide the best way to measure forecast error.
Learn how to test your forecasts by dividing historical data into in- and out-samples. Prepare a linear extension forecast and see its inaccuracy compared to the naive forecast. Learn how to calculate MASE (mean absolute scaled error); this is by far the best way to measure forecast accuracy.
Find out why exponential smoothing is so popular. Gain an understanding of trend, seasonality, and recency. Explore a free exponential smoothing software package: PeerForecaster. Apply it to a data set. Divide the data into in- and out-samples. Find out why smoothing is far more accurate than linear extensions.
Examine data distortion and inaccuracy. Discover the typical causes of data problems. Learn when to deleting “old world” data and improve forecast accuracy with cleaner data.
Multiple forecasts result from different: forecasting methods, software packages, software settings, data sets. Learn what to track for every forecast you make. Create strategies for combining forecasts to create a single forecast into the future. Learn what to avoid when combining multiple forecasts into a single future forecast.
Learn the minimum steps of forecasting: plot, divide, test, assess, apply. Everyone involved in forecasting must understand these steps. Apply them to a real-world situation.
Take your forecasting to the next level.
Click on a chapter below for details.
In Chapter 1 we review the main lessons from the Forecasting Fundamentals course, with an emphasis on the 5-step forecasting process.
The ARIMA forecast method is more complicated than exponential smoothing, but sometimes delivers more accuracy. It also leads the way to ARMAX, one of the most powerful forecasting methods. This lecture gives demonstrates the power of ARIMA and shows one method for finding accurate ARIMA forecasts.
In this chapter we provide step-by-step instruction on how to use STATA to create accurate ARIMA forecasts.
How to design field tests to greatly enhance your forecasting data.
How to convert data from marketing programs into dummy variables, create an ARMAX forecast, and greatly enhance forecast accuracy. Detailed instruction on how to use STATA to generate ARMAX forecasts.
Challenges with using economic variables in forecasts. How to determine whether an economic forecast will help or hinder your forecast accuracy.
How to calculate confidence ranges and apply them to forecasts.
Pitfalls to avoid: taking shortcuts or skipping the fundamentals, trying to forecast the un-forecastable, creating an illusion of certainty, making things too complex, and letting “gut” instinct override the facts.
Online Learning Advantage
Much cheaper than classroom training
Track learner progress and quiz grades
Option: add to Learning Management System (LMS)
Multiple Learning Modes
Real-world application exercises
Set your own pace
Available 24 x 7
Retake the course as often as you like