Predictive Sales Forecasting: Getting Started

Our AI Sales Forecasting module provides monthly sales predictions to help distributors make data-driven decisions around inventory, sales strategy, and marketing. This module is automatically enabled for your organization; you can access pre-built dashboards and reports that display the forecasted data, without needing to manage any AI models directly.

What You’ll See

1. Executive Review Dashboard (Full-Invoice Forecast)
   • Located in Catalog → Dashboards → Executive Dashboards (Library).
   • Shows overall, high-level sales predictions vs Actual Sales comparison by Quarter from the FC.SALES cube.
     
     
     
2. AI Sales Forecasts:
   • Located in Dashboards → Executive Dashboards (Library).
   • Shows overall, high-level sales predictions from the FC.SALES cube for different time dimensions
     
     
     
     
3. Multi-Dimension Forecast Reports
   • Located in Reports → (AI Forecasts).
   • Separate reports for:
     • Location (FC.SALES.LOC)
       
     • Product Group (FC.SALES.PT)
       
       
     • Sales Rep (FC.SALES.REP
       
       
   • Each report independently forecasts sales at that specific dimension. 

Key Benefits

Plan Inventory & Logistics

• Identify which regions, product groups, or sales reps will see higher sales in the upcoming month.

Allocate Resources Effectively

• Shift marketing budgets or additional staff where the forecasts indicate the most opportunity.

Company-Wide Visibility

• Quickly review your overall sales trajectory in the Full-Invoice Forecast dashboard.

Manage Sales Rep Performance by Setting Realistic Targets

• Use Forecasted Sales by Sales Rep to set realistic monthly goals and identify coaching opportunities.
• A Sales Manager can even embed their reps’ forecasted data directly on their CRM home page, allowing quick, at-a-glance updates every day.

Important Details & FAQs

1. When Does the Model Run?
   • The AI engine refreshes forecasts at the beginning of each month, using complete data from the previous month.
2. Why Don’t All the Totals Match?
   • Each forecast (Location, Product Group, Sales Rep, or Full-Invoice) is generated independently and uses different levels of detail. It’s normal for the total numbers to vary.
3. Can I Drill Across Different Dimensions?
   • Not in this version. The forecasts for Location, Product Group, and Sales Rep are standalone; you can’t drill from one dimension into another.
4. Who Can See the Forecasts?
   • By default, admins control who can view these AI cubes. If you see the dashboards/reports, you already have access to the forecasted data.
5. How Do I Customize?
   • Feel free to copy any dashboard object or report into your own library/dashboards.
6. Why don’t I see any forecasts, or why are forecasts missing for certain categories/ids?
   • The forecasting model requires two key conditions to generate forecasts:
     1. At least 5 years of sales data must be available from your ERP. Without this, the model cannot generate forecasts.
     2. The model applies a minimum accuracy threshold of 75% for each forecast id (e.g., Location, Product Group, Sales Rep). If the accuracy for a specific dimension id falls below this threshold, the forecast for that category will not be displayed

7. How is accuracy calculated?

   • Accuracy is calculated during the model validation phase, after training is complete. The model is tested on a holdout (test) dataset that was not used during training. The accuracy represents how close the forecasted values are to actual sales in this test set, using statistical error metrics (such as MAPE or RMSE) converted into a percentage. Only forecasts that meet the minimum accuracy threshold of 75% are displayed to ensure reliability.

8. How does the model know what data to use?
   • The forecast is based off sales that match the filtering specified within your existing SALES cube.