HOW TO IMPROVE DEMAND FORECAST ACCURACY

A well established fact is that improved forecast accuracy leads to many downstream

improvements not only in operations, but in a variety of business areas such as customer

service and asset management. This white paper concerns itself with the improvement of

forecasts that support the management of a company’s supply chain.

In some cases even the product-level forecast is too random to be useful, which leads

to an aggregated level such as product sub group, group or category level. Even though

forecasting at these summarized levels produces improved statistical results, the challenge

of disaggregation to the prime procurement or manufacturing level (product) exists.

Methods of breaking down aggregated forecasts will be covered in subsequent white papers,

along with the consideration as to the use of units or currency.

The forecast accuracy improvement problem can be approached by examining the underlying

elements that must be considered before the analysis and adjustment process is complete

and a routine production forecasting process is operational.

There are some FORECAST IMPROVEMENT METHODS:

1. Create at least 2 additional demand data measures

2. Filter out each customer’s product sales from DM1 history where not purchases were

made in the last 12 months.

3. Produce a Baseline Forecast using the product elimination DM1 adjustments

4. Create a Stratum Planner Workbench view for analysis