Understanding the Elements of a Forecast
A demand forecast is a prediction of the quantity of a specific product that will be sold or used in an upcoming time period (usually a month). The accuracy of your forecast is a determining factor in whether or not you will achieve effective inventory management.
Most organizations base their forecast on an average of past usage, usually the usage recorded over the past three to six months. The idea is that what you sold or used in the past is a good indication of what you will sell in the future. This works well with an item that has consistent usage. Consider usage of the following item:
Averaging usage over the previous six months’ results in a forecast for February of 38 pieces [(42+32+40+36+44+34) ÷ 6]. This doesn’t seem to be a bad estimate. However, the accuracy of the forecast rapidly decreases if usage of the product is increasing or decreasing over time. Consider an item with the following usage:
February’s forecast, based on the average usage over the previous six months, would be 53 pieces [(40+45+52+55+62+64) ÷ 6]; well behind the increasing trend. In order to obtain an accurate forecast, we must both decrease the number of months whose usage is averaged in the calculation and include a trend factor. Average usage over the previous three months is 60 pieces [(55+62+64) ÷ 3]. Usage has increased an average of 8% per month over the past three months. If we add 8% to the 60 pieces from our calculation, the resulting forecast for February is 65 pieces. This is a much better forecast.
Next month, we will continue our exploration of “best practice” forecasting and determine how to best forecast demand for products with different patterns of usage (including seasonal products). For the time being, please understand that:
- One forecast formula will not work for all your stocked items
- You must look for increasing or decreasing patterns of usage over time
- When you detect continual increasing or decreasing usage, you must only consider the usage recorded in the last several months and apply a trend factor to reflect the change in demand over the last several months.