Common Forecasting Myths Debunked – Part 2 – Overcoming Lumpy Demand
This is the second post in a series of common myths around forecasting. Here we explore how forecasts can be created for even the most volatile demand patterns.
Myth: You cannot create an accurate forecast for products whose demand histories contain lumpy or intermittent demand patterns.
Demand planners may think creating a unique forecast for products whose demand histories contain lumpy or intermittent demand is a lost cause. However, there are techniques that permit zero demand to reside in the history and can still forecast demand.
Reality: There are advanced techniques that can improve forecast accuracy for items with lumpy and intermittent demand.
A common challenge for many planners is to create an accurate forecast for products with intermittent or lumpy demand. For many companies, 80% of their revenue comes from 20% of their products, and the other 80% of products will have fairly low demand that is intermittent or lumpy. Many statistical forecasting methods cannot handle history that contains periods of zero demand. In addition, most statistical methods do not do a good job when demand history is lumpy, meaning a non-repeatable pattern of high and low demand periods. So, how can demand planners forecast for a product with all of these existing barriers? One method is to use the Modified Croston technique to create forecasts for items with periods of zero or lumpy demand.
A Modified Croston method handles low and lumpy demand that exhibits either a patterned variation or no pattern at all. The patterned variation looks at available history and analyzes each demand element relative to those around it. It classifies the periods into peaks, valleys, plains, plateaus, up-slopes and down-slopes. It measures the duration of plateaus and plains, as well as the severity of peaks and valleys and then conducts pattern-fitting analysis to find regularity over time, attempting to fit the pattern to the history and averaging for low and high points. If no pattern is present, the unpatterned variation method attempts to use averaged highs and lows to create a step-change forecast for future demand. This results in a forecast for your product with lumpy and intermittent demand.
What other forecast myths have you identified? If you missed part 1, read it here and learn how to develop a forecast that is unique to your business.
So, the promise of using statistical algorithms, forecasting and predictive analytics is now added to the list of a company’s number one priorities. T
In the world of commerce, every business ecosystem has a type of supply chain that is critical to corporate operations.These supply chains rely on a n