Why should I forecast my demand, when forecasts are never right?   This is the question we are continually asked when the subject of forecasting arises.  If you are a make-to-order manufacturer, and you can efficiently keep up with demand then don’t bother.  But if you have issues in any of these areas read on:

Insufficient capacity to meet demand

Customers that demand shorter lead time

chk3Longer raw materials lead times from your vendors

chk4Innefficiency of your manufacturing process due to a high number of changeovers

chk5High finished goods inventory levels to support demand

chk6High inventory of obsolete finished goods

The primary purpose of forecasting is to feed your planning process.  In the planning process you will take what you think you need to make(your forecast) and determine what you would like to make(your plan).   Forecasting provides a starting point for the planning process, as well as providing the basis for evaluating spikes and drops in demand to assist in planning for the right level of finished goods inventory.

There are many good forecasting systems on the market, and we have worked with many of them.  What we have found is that most of these systems do an excellent job of applying statistical algorithms to your historical demand, but that is only part of the solution.   In order to forecast in a way that will truly be of value there are four critical factors:

  1. The forecasting PROCESS is the key
  2. Forecasting systems have a role, but cannot be the only step.
  3. Forecasting is an iterative process that must communicate exceptions to those who can provide input.
  4. The process must be automated, incorporating checks and balances or it will fail.