Once you have created a new project, the next step is to generate statistical forecasts. In this article we will describe:

• how to generate the forecasts,
• the settings impacting them, and
• the viewing capabilities of the software.

To generate the forecasts, do the following:

1. Set the forecast Horizon to the number of time periods to forecast.
2. Set the starting period to either a complete or incomplete period in the As of control.
3. Click the Forecast button in the program toolbar.

The As of control selects the period for which you want to have forecasts and inventory plans. If you open the drop-down list, Streamline shows you the periods for which forecasts can be generated. It also indicates the periods where forecasts were previously calculated, displaying a little blue circle before these periods in the list. In this way, Streamline allows you to go back to previous periods and see what the forecasts and plans were at that time, or to forecast past periods, (where we have actual data), and consequently, evaluate them.

The As of parameter allows you to choose one of two options telling Streamline how the sales data for the last period should be treated when forecasting. If the last period has the entire month’s historical data, we recommend using the complete period forecasting. In the above figure, this is the Dec 2016 option.

Otherwise, use the incomplete period forecasting (the ‘Incomplete Dec 2016’ in the figure above). This option makes Streamline ignore the sales data in the last period and build the forecast as if the last period is the previous one. It is designed to ensure Streamline does not underestimate the forecast for last period if it is forecasted prior to having the full month’s data available.

Streamline has a lot of forecasting settings. All of them were specially designed to be of help in particular situations. In this section, we will introduce them and point out the cases they can be applied in.

All of the parameters are located on the Forecasting tab of the Demand forecasting and can be applied on an item, category or location level, generally, at every level of the Tree view tree.

Most of them have the Inherit option selected by default. It means that the option of its parent node will be applied to the currently selected node. Information in brackets shows the choice inherited from the parent.

The table below shows the cases, settings, and the appropriate options.

Case	Setting	Option
New product forecasting	Forecast approach	Top-down
	Use model from	Item code of the borrowed model
Accounting for lost sales	Zeros are lost sales	Yes
Accounting for calendar events	Holidays	Calendar of your country
Accounting for recent changes in the sales history level	Ignore data before	The data will be ignored prior to the selected Date
Seasonal forecasting with recent sales history level changes	Use ignored data for seasonality	Enabled
Adjusting the model type	Model	The model type
Increasing/Decreasing the model output	Multiplier	The coefficient
Disabling exponential smoothing for seasonal coefficients	Seasonality weighting	1

To put your settings into effect, re-forecast the project by clicking the Forecast button.

Streamline represents the data you have imported as a hierarchy tree in the Tree view on the Demand forecasting tab. The Tree view allows navigating through the item codes (products), their categories and locations, to view their properties. The Table, Plot and the Properties panel show the information regarding the tree node currently selected in the Tree view.

When you select a node in the Tree view its forecast is immediately shown in the Statistical Forecast row of the Table view and depicted by the red line on the Plot view. If the selected node is a tree leaf, the Table view and the Plot view also show the forecasting model fit in blue color.

Forecasting models are only built for the tree leaves. Forecasts at higher levels of the hierarchy are an aggregation of the forecast of lower ones. The forecasting settings and the resulting model for the selected tree node are shown in the Forecasting and Model tabs of the Properties panel.

To view the forecasts in detail graphically, the Plot View supports such manipulations as zooming a particular area and panning the plot in the zoomed scale.

The Tree view allows viewing the forecasts and other information by item or location.

The By item option creates a hierarchy where item codes and their categories reside on top of locations. This is the default view.

If the By location option is active, the locations and their categories are on top of the item codes allowing viewing and managing the forecasts by locations.

The By channel option is available if you import channels instead of locations. It allows you to view and plan your future demand by channel.

Use the Expand and Collapse links of the Tree View to show or hide children of the currently selected tree node.

Streamline allows viewing forecasts for all the planning items at once by using the forecast reports. There are three forecast reports in Streamline: Final forecast report, Forecasts adjustments report, and the Forecast accuracy report. To access the reports, go to the Reports tab. To select a report, use the Select report control.

Each report can be exported to Excel by clicking the Export report button. These can then be edited and saved in Excel and imported back into the application using the Import changes button.

Final Forecast Report

This report is shown by default. The report depicts the statistical forecasts and any overrides for all of the planning items. Any overrides are shown in blue.

Forecast Adjustments Report

This report shows only the forecast adjustments made manually in the Forecast adjustments row of the Table view on the Demand forecasting tab. It is useful when you need to ensure that all the adjustments you intended to do were done.

Next: Adjusting and Approving the Forecasts

Download PDF