Create and save forecast preferences.

Forecast preferences are like templates, where you can save the forecast parameters you want to apply while creating a forecast. 

1. To create a forecast template, click on the preferences as shown below.

2. Clicking on the preferences will open a page like the one below. This page displays all the available forecast templates.

3. To create a template, click on the Create a new template button.

4. The following page will be displayed after clicking the Create a Forecast template button.

5. The mandatory fields must be filled in to create a forecast template.

   1. Template name: The name of the forecast template must be unique across the forecast management.

   2. Description(optional): Providing a description is optional. However, it will help readers understand the purpose of creating the forecast template.

   3. Hierarchy level: The appropriate hierarchy level is crucial, as the forecast results will be generated based on the chosen level. The hierarchy defines how data is structured and analyzed in the forecasting process.

      1. Example: If you want to create a forecast template with a combination of product and region, select the 'Product' hierarchy from the Product drop-down, choose the 'Region' hierarchy from the Customer drop-down, and select location-all and source-all from the Location and Source drop-downs.

4. Time horizon: Selecting the time horizon is essential for generating statistical forecast values based on historical data. It defines the period over which the forecast will be generated.

   1. Historical period: Users can select all or specific historical data based on the requirement.

   2. Future period: The forecast engine generates statistical forecast values based on the selected future horizon. For example, Selecting a 12-month future period generates forecasts for the next 12 months based on historical data patterns.

5. What to forecast: Select the input and output measure for the forecast to generate values.

   1. Output measure: By default, the forecast selects Statistical forecast units and revenue as its output measures.

   2. Input measure: The input measures will appear in the dropdown based on the selected output measure type. For example, If Statistical Forecast Units is the output measure, only unit-based input measures (e.g., shipped units, booked units) will be available. If Statistical Forecast Revenue is selected, revenue-based input measures (e.g., invoiced revenue, booked revenue) will appear.

   3. Both output and input measures can be configured from the admin application. 

6. Influencing factors: Influencing factors help improve forecast accuracy by considering external and internal variables that impact demand. These factors adjust the forecast based on real-world conditions, ensuring more precise predictions. For example, oil/ gas prices, retail sales, and other factors affect upcoming demand and sales. Selecting the influencing factors is optional.

7. Forecast methods: Firstshift.AI leverages AI and advanced statistical models to generate your business's most accurate and reliable technical forecasts. Our system intelligently selects the best forecasting technique based on your data patterns, ensuring optimal results.

   1. The best model (let the forecast engine choose): If this option is selected, the system automatically determines the best forecasting model based on historical data patterns. AI-driven algorithms analyze trends, seasonality, and demand fluctuations to pick the most suitable model.

   2. Let me choose: If this option is selected, you can manually select a forecasting method from a dropdown menu. This allows experienced users to apply specific models, such as ARIMA, Holt-Winters, or Moving Averages, based on their preferences or business needs.

8. Remove outliers: Removing outliers from the data helps create more accurate forecasts by eliminating extreme values that could distort trends. However, it can also dampen the accuracy of statistical patterns, primarily if those outliers represent recurring business trends.

   1. Selecting ‘Yes’ will remove the outliers caused by temporary events that are unlikely to recur.

   2. Selecting ‘No’ will keep the outliers representing important business events (e.g., seasonal spikes or significant product launches).

   3. We recommend enabling outlier removal, which improves forecast reliability by focusing on core demand patterns rather than temporary anomalies.

9. Confidence level: You should select the confidence level for your forecast carefully, as it affects the reliability of your predictions.

• Higher confidence levels (e.g., 97.5%) provide more certainty but result in wider prediction intervals, meaning a broader range of possible outcomes.

• Lower confidence levels (e.g., 80%) produce narrower prediction intervals but with a greater chance of missing actual values.

• In general, choose higher confidence levels when you trust your historical data and need conservative estimates. Uncertainty increases for extended forecast periods, so balancing confidence and practical accuracy is key.

10. Disaggregation rules: “Consider the historical data of individual lower-level items” means analyzing past performance data for specific products at a granular level.

    1. Instead of relying only on aggregated data, this approach examines the sales, demand patterns, and fluctuations of individual items within a hierarchy (e.g., specific SKUs within a product family).

    2. It helps improve forecast accuracy, identify trends, and capture seasonality or anomalies that might be lost in higher-level data aggregation.

6. Save template: Once all the mandatory fields are filled, click the save template button to create a new template.

7. If any mandatory fields are left blank, the application will indicate them with a red mark, as shown below. This shows that the template cannot be created without filling in all the mandatory fields.

8. Make sure all the fields are filled correctly before saving the template.

9. All the created templates will appear in the preferences table with all the settings selected for the template.

10. The exact details can be seen in the card view as well. Click on the card view, as shown in the image below.

11. Once clicked, the following page will appear, showing all the settings.

Edit and delete the template.

The created forecast template can be edited and deleted based on the requirements.

• Editing: The forecast template can be updated accordingly if adjustments are needed, such as modifying the hierarchy level, time horizon, or forecast method.

• Deleting: If a forecast template is no longer needed, it can be removed to keep the system organized and relevant.

To access the Edit and Delete options for forecast templates, scroll the table horizontally.

Clicking on edit will take users to the edit page, where all the settings can be edited.

Clicking on Delete will open a popup like below. Click ‘Yes’ to delete the template permanently or ‘No’ to terminate the action.

Once you click Yes, the template will be permanently removed from the application.