In IFP, statistical forecasting methods are available for the following types of analysis:
- Fitting of mathematical trend models to a selected period of historical values
- Estimation of average seasonal patterns in any selected period of historical values
- Calculation of forecast values using projected trend values adjusted according to estimated seasonal factors
This approach to forecasting is most useful when reasonably stable patterns exist in actual values over a period of two or more years.
- Forecasts may be created using any level of data, e.g. SKU (item code) details or brand totals
- Forecasts created at group level are automatically split into corresponding item code details on a pro-rata basis
- The base for these pro-rata allocations may be any set of base data, e.g. previously created forecasts or prior year actual data
- Up to 10 years of forecasts may be created for each item
- You may also specify the parameters of trend models using your own judgement and knowledge of market conditions
A good choice of forecasting model for any item can only be made through consideration of past and likely future market conditions as well as any patterns in historical values for the item. See Selecting the Best Statistical Forecasting Model for details, and Trend Models in Statistical Forecasting for full details on each type of Trend Model available in IFP.
The accuracy of any forecasts generated here depends critically on the selection of appropriate trend models and seasonal patterns for appropriate periods of historical data.
Any mathematical analysis of trends and seasonality will be of limited value if the historical data being used has not been previously 'cleaned' to remove the effects of known unusual events. See Data Preparation and Cleaning for details.
Clearly, the use of statistical forecasting models requires a suitable set of historical demand data for each item. Hence, these models should normally be applied to in-line items only (not new or discontinued items).
Also, these models are normally best applied to sales items (not samples or free goods). For example, it is normally best practice to forecast free goods as a specified percentage of corresponding sales units.
It is important to remember that Statistical Forecasting models on their own can only suggest future levels of monthly demand assuming that past trends and seasonality continue into the future. Hence, final demand forecasts must include adjustments to reflect any special factors that may arise in future periods which have not been present in historical demand. See Special Factor Adjustments in Statistical Forecasting for details.
Creating a Statistical Forecasting Model
- From IFP Home select Forecasting > Statistical Forecasting.
- Select the demand and forecast files to use.
- Choose the period to use and Data to be Amended and Entered.
- Choose the items that you wish to include.
- Select the Level of Detail, press Next and Finish.
- The statistical forecasting dashboard will open (an example is shown below).
- You should now Process the Item Models.