Forecast Management

There are many methods of forecasting that predict a variety of things. Extrinsic methods involve factors outside a company, such as economic conditions or the sale of other goods. Attempts are made to show some pattern or correlation linking products with these factors to forecast the demand for your products. Qualitative forecasting techniques involve judgment, intuition and informal opinions and are used mainly by senior managers. Also, while this method must be used for new products, since no historical data exist, it is not specific enough to forecast many different manufactured products for extended periods of time. Most forecasting that is done for manufacturers, however, is related to a products sales history and, typically, does not recognize any outside factors. There are some simple techniques that can be used to apply some of these outside factors to traditional methods that, most important, can improve the reliability and trust in the company's forecasts.

Forecasts have six major characteristics or principles that, when sufficiently understood, can provide insight into the need for further enhancements of a purely mathematical projection:

Forecasts are almost always wrong. With the exception of sheer luck, it is very difficult to project future unknown events exactly. We must learn to accept this first principle and not use incorrect forecasts as an excuse for other problems. Late shipments, backorders and excess inventories can be the result of problems on the shop floor. Small errors in forecasting often take the blame, which leaves the real problems unattended.

Forecasts must include some measure of error. Since they will be wrong, the second question is "by how much?" Estimates of the forecast error can be made statistically by studying past performance of actual demand to forecast. This estimate of error can be used for capacity planning and, of course, for estimating safety stocks.

Forecasts are more accurate for families or groups. The behavior of individuals within a group is random even when there is stability in the overall group. For manufacturing purposes, the group should be based on the similarity of process or equipment used. This is often difficult to do since marketing personnel will base their group designation on customer patterns and not on commonality of manufacturing process.

Forecasts are more accurate for nearer periods of time. It is far easier to predict tomorrow's weather than it is to predict the weather a month or a year away. The same holds true for forecasting the demand for a company's products.

Forecasts are more accurate for longer periods of time. The total demand for a product can be more accurately predicted over a number of periods than for individual periods. As an example the

forecast of rainfall for a given month is more accurate than the individual daily predictions. Or as another example, early sales of a product are often offset by reduced sales later on since most of the total demand for a product has already been satisfied.

Forecasts are no substitute for calculated demand. Manufacturers have a definite need to forecast finished products. However, the demand for subassemblies is entirely dependent on the demand
for finished products and can be calculated more accurately than can any attempts to forecast.

To be Continued

Lean Six Sigma Training   Thinking Out of the Box   
Balanced Scorecards  Strategic Tactical Planning  
Supply Chain Inventory Management  Total Quality Management Principles
Lean Manufacturing Implementation  Lean Manufacturing Principles