Tool Box
Issue No. 24 - August/September 2005
Predicting the future
Demand forecasting for manufacturers and retailers
by Mr Matthew Michalewicz
Although technology has moved at warp speed during the last few decades, the basics of retailing and manufacturing have remained the same: Understanding customer needs, forecasting future demand, and planning inventory and production to meet that demand, is as important today as it was 30 years ago.
But even with all the advances in technology, accurately forecasting demand is still a formidable task.
In truth, the process of predicting the future—of predicting what products will be needed and in what amounts—is one of the most difficult challenges facing manufacturers and retailers today. And because demand forecasting underpins the entire logistics process, affecting capacity plans, master production schedules, and inventory levels, its importance cannot be overstated.
Given all of this, what can manufacturers and retailers do to improve their demand forecasting capabilities?
To answer this question, let’s briefly outline three common approaches to forecasting:
The first approach is based on intuition. Also known as the “judgment method,” this approach relies on sales force estimates and executive opinions, and is most often used in small companies where historical data is not available or when demand is driven by vacillating consumer sentiment (such as fashion trends and fads).
Using intuition to forecast demand won’t produce consistently accurate results, and because the results are subject to the bias of the people creating the forecast, most companies favor methods that are more objective.
The second, and far more impersonal approach, is based on statistical methods.
Using a basic spreadsheet package that contains linear regression functions, every manufacturer and retailer can use past sales data to project future sales.
By collecting additional data on variables that impact demand (such as the consumer price index, interest rates, housing starts, and seasonality), companies ca...
