Calculate Forecasts

Overview

In GoldFinch, Sales Forecasts can be manually entered, imported, or calculated based on historical sales. These forecasts help predict future demand and are essential for accurate supply planning. Before running the Calculate Forecast function, you need to follow several preparation steps to ensure your forecasts are accurate.

Preparation Steps Before Running Calculate Forecast

1. Run Supply Plan Daily Calculation Routine:
   Ensure that the Supply Plan Daily Calculation Routine has been executed, either manually or through the scheduled nightly job. This routine updates SKUs, Item Usage records for historical sales, and identifies the best-fit forecast model for each SKU based on the prior period.

2. Enter Item Usage Adjustments:
   Input any Item Usage Adjustments to account for abnormal sales activities or exclusions.

3. Review Demand Planning Setup:
   Navigate to Demand Planning Setup and review key fields like Periods to Forecast and Inventory Period Type (e.g., by Month or by Week).

Running the Calculate Forecast Function

Once the setup and preparations are complete, follow these steps:

• Go to the Sales Forecast List:
  Click the Calculate Forecast button. This will delete existing sales forecasts and use the Best Fit Forecast Model for each SKU to calculate forecast quantities, broken down by warehouse.

Schedule Calculate Forecast Routine:

• Go to Setup → Apex Class → Schedule Apex.

• Set the job name to Calculate Forecast.

• Select the Apex class SalesForecastBatchDelete.

• Schedule the job with a weekly frequency, ensuring it runs after the SupplyPlanBatchDailyCalculations routine (e.g., 12 PM every day).

When SalesForecastBatchDelete finishes, the Apex class SalesForecastBatchCalculate will run to calculate the forecasts. You can monitor these jobs from Apex Jobs.

Note: The SalesForecastBatchCalculate process is CPU-intensive. If you encounter a "CPU time limit exceeded" error, adjust the SalesForecastBatchCalculateBatchSize parameter in Setup → Custom Settings → System Settings to a lower number (e.g., 25 instead of the default 50).

Forecast Calculation Example with Sample Calculations

Let’s assume the following setup in Update Demand Planning Setup for today's date of 3/25/20.

Setup Information:

• Inventory Period Type: By Month

• Periods to Forecast: 12

With this setup, GoldFinch will create forecasts starting from 3/1/20 (the current period's starting date) through 2/1/21.

Forecast Calculation Steps:

1. Retrieve Item Usage Records:

   • GoldFinch retrieves all Item Usage records with a starting date of 2/1/20 (the prior period). These records will have the Best Fit Forecast Model assigned for each SKU, as determined during the Supply Plan Daily Calculation Routine.

2. Forecast for Future 12 Months:

   • GoldFinch will use the Best Fit Forecast Model for each SKU to generate forecasts for the next 12 months, from 3/1/20 to 2/1/21.

3. Forecast Behavior Based on "Calculate First Period Forecast Only":

   • The forecast calculation differs depending on whether the Calculate First Period Forecast Only option is selected.

Scenario 1: Using the Average of the Last 12 Months for Future Forecasts

• If you want to use the average of the last 12 months to calculate the forecast for the next 12 months, you would configure the forecast model with the Calculate First Period Forecast Only option selected.

Example Forecast Model Lines:

• Forecast Calculation:
  GoldFinch will calculate the forecast for 3/1/20 using the average Actual Quantity from 3/1/19 to 2/1/20.

• Copying to Future Periods:
  The calculated forecast will then be copied to all future periods, from 4/1/20 to 2/1/21. This results in the same forecast quantity for all periods.

Scenario 2: Without "Calculate First Period Forecast Only"

If you do not check the Calculate First Period Forecast Only option, GoldFinch will:

1. Calculate Each Period's Forecast Individually:

   • The forecast for 3/1/20 will be calculated using actual quantities from 3/1/19 to 2/1/20.

   • The forecast for 4/1/20 will be calculated using actual quantities from 4/1/19 to 2/1/20, and the calculated forecasted quantity for 3/1/20.

2. Dynamic Forecast for Each Period:
   Each future period's forecast (from 5/1/20 to 2/1/21) will be calculated based on actual data from the corresponding period of the previous year, resulting in varying forecast quantities for each period.

Key Differences Between Both Scenarios:

1. Consistent Forecast:

   • When the Calculate First Period Forecast Only option is selected, the same forecast quantity is applied to all future periods. This approach provides a stable and predictable forecast across multiple periods.

2. Dynamic Forecast:

   • Without the Calculate First Period Forecast Only option, each period's forecast is calculated based on actual data from the previous year, which can lead to varied forecasts that reflect changing trends, seasonality, or other patterns in the historical data.

This flexibility in forecasting allows you to choose between a steady, fixed forecast or a dynamic approach that adjusts to historical trends, ensuring your supply planning meets the specific needs of your business.