Forecasting Approach
The SMS approach is a fast and automatic way to forecast stock usage by analyzing past demand patterns.  Since each item has a unique pattern, applying a variety of models will cover all possible patterns and objectively select the best model.

Other software approaches are often limited to only a pre-defined model choice. Some "forecasting" software lacks comprehensive replenishment planning capability.  Other "forecasting" software requires seasonal factors, manual trends controls and other manual controls for each item that may be forecasted.  Further, customization of these models is cumbersome since they have to be manually addressed by the user individually.

SMS provides 15 statistical models which are precise, yet easy to customize.  There are five (5) steps in the SMS process:
  1. Forecast SimulationApply all models to analyze existing demand patterns.
  2. Error QualificationCheck simulated forecasts and remove bias or volatile forecasts..
  3. Model SelectionReview remaining models and select model with lowest error ("best fit").
  4. Forecast Calculation - Apply "best fit" model to forecast the coming year.
  5. Forecast Adjustment - User adjusts forecast (if necessary) to reflect current conditions.

SMS includes the following statistical models:
  • Moving Average ( 3 month / 6 month / 12 month )
  • Trend Average ( 3 month / 6 month / 12 month )
  • Seasonal (No trend this year = last) (With trend - 3, 6, 12 moth)
  • Single Exponential Smoothing ("alpha" .08/ .25/ .50)
  • Linear Regression (No adjustment / with adjustment)
  • Custom Exponential Smoothing ("alpha" = user value)