Skip to main content
All CollectionsPivot Forecasting
Pivot Forecasting - Admin Overview
Pivot Forecasting - Admin Overview
Ruvisha Pillay avatar
Written by Ruvisha Pillay
Updated over a year ago

If you are a Power User, this article is for you. In this article, we discuss Model Administration in Pivot Forecasting.

Model Administration includes:

  1. Setting and changing model options

  2. Importing data into models

  3. Backing up models

  4. Restoring backups to models

Let’s take a closer look at each of these functions.

Setting Model Options

Model Options define the settings of a model, such as the model’s name, period type, how many historic and future periods it has and how many decimals it should display. In Pivot Forecasting, some of the model options settings come from your inventory app and should never be changed. Other options can be changed as needed.

The Model Options settings can be found by:

  • Opening a model (File → Open Model)

  • Clicking the Admin drop down on the top of the screen

  • Selecting Model, then Options

Once “Options” has been selected, the below screen will appear.

We will take a closer look at some of these fields.

The Model Name is set in the inventory app via the Configuration screen → Maintenance tab during the initial setup.

The Period Type can be set up in months or weeks. This is also established in the initial set up in the inventory app. Please note: this field should never be changed in Model Options.

The Number of Historic Periods and Number of Future Periods is determined by the number of historic periods and number of future periods in the inventory app. These settings should never be changed in Model Options for Pivot Forecasting.

Models where demand forecasts are created should have Log Transactions turned on by clicking the checkbox. The system will log every change made interactively via Finalize, Copy Measure, Clear Measure, Import Forecast Spreadsheet, Adjusted History, Promotions, and Top Down Forecasts. The Log is helpful in determining who, how, and when a forecast changed.

Log Retention Days sets how long transactions remain in the log. A value must exist for this setting, and the default of 60 days is usually sufficient.

The Log can be turned on at any time and the Retention Days can be reset as needed.

The Base Item Description is the measure that is used to show item descriptions in

the Navigation Tree as well as certain panels that show descriptions for base-level

Items. This can be set/changed by an administrator. Selecting the drop-down will give you options which can be chosen to populate this field.

The Forecast Comparison Measure determines which measure is used (along with Actual History) to calculate Forecast Deviation, Percent Deviation, and WMAPE which are all used to determine forecast accuracy.

The Number of Recent History Periods is used with the standard filters to define how many periods are “recent”. The applies to:

  • Recent Outliers

  • Recent Promotions

  • Forecast No Recent History

  • No Recent History

Any changes to the Forecast Comparison Measure and Number of Recent History Periods fields must be followed by a manual Recalculate from the File menu to see the change.

References are also important when defining Model Options. With Pivot Forecasting, the first four Reference Measures are set automatically and should never be changed.

The remaining Reference Measures five to ten are left blank by default and available to users as needed. Within references, you have the following options:

  • Add a reference measure to the model by entering a Label.

  • You can check and uncheck the “Convert?” box. It is checked by default. Convert ensures that values will be converted to price, cost or other alternate UOMs.

  • You can check the “Editable?” box to make a measure editable by a user with “write access” to that measure (set in routings). Measures that are populated via automation, such as forecast archives, should never be made editable.

NOTE: The below Model Options are set by the inventory app configuration ONLY. Users should never change these:

  • History Periods and Future Periods

  • Period Type/Periodicity

  • The first four Reference Measures

  • Start Day of Week in a weekly model

Importing Model Data

Once a model is created and the Model Options have been set, the model needs to be populated with data. It is important to note the following:

  • Data in Pivot Forecasting models is imported nightly from the inventory app.

  • Pivot Forecasting does not allow for imports from other sources.

Backup Model

With Pivot Forecasting, the model database is backed up automatically each night. Manual backups can also be made ad hoc, such as saving a snapshot of the model database at a point in time during the day.

To create a backup file:

  • Open the model to be backed up

  • Select Model → Backup from the Admin menu

  • Enter a backup file name

  • Either leave the Overwrite box unchecked or you can check it to overwrite an existing backup

  • Click Process

Restore Backups

Administrators can restore models from backup files to recover lost or corrupted data. Restoring a backup completely overwrites the existing model database to a previous point in time, and there are no incremental backups.

To restore a model:

  • Open the model to be restored

  • Select the menu item Admin → Model → Restore

  • Select the backup file to restore by name and date

  • Never restore a backup file with a name that starts with “configuration”

  • Click Restore.

Be sure to double check that you are restoring the correct model.

Did this answer your question?