Sales Forecasting

This process is used to prepare and make calculations for forecasting expected values. The values can be specified for separate accumulator fields in datasets. This is for periodic forecasting, most suitably at the beginning of a new period when the actual results from the previous period are completed.

Transaction type 34 (Sales forecast) is available in a dataset for reporting purposes after using this process.

In M3 Sales forecasting there are two ways of forecasting: standard forecasting and advanced forecasting. The difference between these two types is that the following issues are available in advanced forecasting:

  • Forecast formula 4
  • Forecast with trend
  • Extrapolation with trend using trend quantity
  • Seasonal influence method 2, 3 and 4
  • MAD (mean absolute deviation) method
  • ME (mean error) method
  • Forecast alarm
  • Forecast error logic
  • Competing forecasts
  • Stop forecast calculation.

Another difference is that in standard forecasting, forecast on a field, e.g invoiced quantity (UCIVQT), is based on the historical data of the same field (UCIVQT). The demand used in advanced forecasting is instead based on other fields. This is further described below.

Before you start

  • Transaction type 34 (Sales forecast) is allowed in the datasets used.
  • Forecast methods are entered for the datasets containing the forecasted accumulator fields. The selected forecast formula for each forecast method determines the automatic calculation alternative used. Datasets are maintained in 'Dataset. Open' (OSS401).
  • Historical statistics are created and stored for each key field included in the automatic forecast calculations. Transaction Type - Statistics for these statistics are defined in 'Sales Forecast Method. Open' (OSS430).
  • In advanced forecasting, the field Demand (UCDEMA) is an accumulator in the dataset. If you want to use advanced forecasting, you will have to enter this field in the dataset.

Follow these steps

  1. Checking forecast method and parameters

    To start the process, the forecast methods and parameters for the dataset to be forecasted must be checked and changed, if necessary. This is done in 'Sales Forecast Method. Open' (OSS430).

    When forecasts are generated using forecast simulations, the simulation forecasts methods used must also be checked and/or changed. This is done in 'Sales Forecast Method. Connect Competing' (OSS431), which is called using option 11 = Forecast sim from 'Sales Forecast Method. Open' (OSS430).

    The dataset should also be checked to ensure the correct forecast method and forecasted accumulator fields are selected and matched.

  2. Calculating forecast

    The forecast calculation is done automatically for every dataset selected using the selected forecast methods. Forecast calculations for separate datasets can also be done in 'Sales Forecast. Generate in Dataset' (OSS440).