Forecast Management

The seasonal indices hold the greatest potential for increased senior manager involvement and for data reliability in the fore­cast. There are four fundamental reasons for adjusting the math­ematically generated, seasonal indices and the resulting forecast.

1. Principle #4 of forecasting stated that forecasts are always more accurate for nearer periods of time. Or, the nearer the product due date, the more accurate the estimate for total demand. This implies the forecast that was originally made far enough in advance to plan production, order materials, etc. can be improved continually. However, this does not occur in practice since human nature resists changing something that is already in place.

2. Outside influences on the demand for products are not included in the mathematical model. These include general swings in the economy and competitive forces. Senior managers consider these factors and make collective changes to overall projections. These are often broad judgments expressed in qualitative terms ("sales will be strong next quarter") or in broadbrush adjustments ("Christmas sales will be up 10% this year"). These very valid considerations are rarely reflected in the mathemati­cal models.

3. There are also activities within a company's control that will influence the demand for products that are unacknowl­edged in a purely mathematical model. Product promotions will have a most interesting effect on demand that market­ing can predict with some degree of accuracy. The solid line in Figure 2 shows a typical promotional profile for a product that experiences growth during the promotion and continues with less-than-peak but higher-than-normal sales after the promotion. Exponential smoothing will lag this

Time

Other profiles exist for the introduction of new products, the phasing out of old products, changes in competition or competitive strategy, other results of promotions, etc. These demand patterns are far more predictable than any mathematical method and must be included in forecasts. They are patterns that will not be repeated until a sub­sequent promotion is held and should have their affect removed from any seasonal indices.

4. The least practiced reason for adjusting the mathematical model is to keep the data base of demand as valid as possible. For example, a company might experience much higher than normal demand for a product due to some unforeseen calamity that has hit its competition. The math­ematical model alone will interpret this as an overall increase in sales and one that will happen again the same time next year. This could be corrected in the data base by changing the past index for that period and correcting the

index for the subsequent year's period.

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