Overview | Moving Average | Exponential Smoothing | Sales Profile | VRS
In this four part series, I will show you the most popular forecasting methods for retail. Each method builds upon itself until we reach the most widely used method, which I like to refer as variable response smoothing, or VRS. It goes by many names, and is almost always modified to reflect the true nature of how your business functions. Whatever the name may be, this method is definitely the secret recipe.
VRS is an adaptation of the two parameters linear exponential smoothing method taken from Charles C. Holt’s 1957 research “Forecasting Seasonal and Trends by Exponentially Weighted Moving Averages.” As you will see, the simplicity and versatility of VRS make it a widely used method in just about every retail forecasting solution that exists. JDA (INFOREM, E3), Oracle’s RDF, SAP, and Logility are just a few. You can easily model VRS in Excel as well.
The first lesson in learning Variable Response Smoothing is to become familiar with the basic forecasting methods it uses. I’ll walk you through these first, and towards the end I’ll apply VRS to some real examples. (You can skip directly to the sales profile if you’re already familiar with moving averages and exponential smoothing)
Part I: Moving Averages
Part II: Exponential Smoothing
Part III: Sales Profiles
Part IV: Variable Response Smoothing
Having some familiarity with basic statistics will help, but it’s not necessary. If you understand the concept of a standard deviation, you’ll be way ahead.