Analyzing Advanced Demand Forecast Methods

After running the Advanced Demand Forecast Program, you can retrieve visual results using the Advanced Demand Forecast Analysis. The analysis provides the results for each selected for product or line and the selected methods.

You can use this information to study the difference between the demand calculated by the current method compared to the demand that would be used with other forecast methods and if the advanced demand forecast is a more viable or reliable solution. you can use the table display that Eclipse provides, or graph the forecast or sales options, to review the information more closely.

You can run the forecast programs and display this information based on the last forecast that was run. After running your demand program, the system alerts you that the program is complete. You can then review the latest information through the Advanced Demand Forecast Analysis window.

Note: Eclipse uses the closest Saturday before the day when the report runs. The system considers the As of Date (that Saturday) to be the Date Last Calculated from the audit window. Therefore the Date Analyzed field matches the date the report runs.

Important: If the seasonality test of a product fails without returning an ACF score or if a product has less hits than the minimum hits to test seasonality as defined in the Advanced Demand Forecast Parameterscontrol maintenance record, then the system calculates a standard demand forecast. Meaning, no advanced demand forecast calculations are made on the product. You can check this information by using the audit screen in Primary Inventory Maintenance.

To analyze your demand forecast methods:

From the Maintenance menu, select Product to display Product Maintenance.
In the Product field, enter the product or line you want to review.
From the Inventory menu, select Primary Inventory Maintenance.
From the Edit menu, select Advanced Demand Forecast Analysis.
Use the Inquiry menu to do one of the following:

Display weekly and monthly calculated values for the selected product in the selected branch.
Display the Product Sales History.

These options display in new windows.

About the Analysis Window

The Analysis window is broken into sections so you can see at-a-glance what your current forecast method looks like compared to what is determined a best fit, or other forecasting methods.

Important: Using the Exclude Negative Quantities From Item Demand Calculations control maintenance record will eliminate any negative quantities, but does not eliminate the system actually calculating a negative demand. This means if you use Holt or Holt-Winters and your sales have actually gone down over the specified period, the system will show a negative demand.

Following are the sections and options available in the Advanced Demand Forecast Analysis window:

Number / Section	Description
1	Description of the product for which you have run the Advance Demand Forecast.
2	If running for a product, the Buy Line the product is contained in, indication if you have the product marked as seasonal, the current Vendor Target, and the date the product was analyzed.
3	ACF Scores. Auto-correlation Function. Displays visually how the seasonal and non-seasonal scores were determined. For more information, see About ACF Scores below.
4	Product image, if available, as entered in the product maintenance file.
5	The Saturday date for the last full week of sales.
6	Option to exclude the product when the update forecast method selected calculates demand based on when Eclipse forces the Advanced Demand Forecast best-fit analysis.
7	For each forecast method, both current and advanced, the system displays the following values: Mean Squared Error (Annual) - The metric that compares actual sales to past forecasted demand. For more information, see How Eclipse Calculates Best Fit . Daily Demand - The current daily demand the system determined when running demand forecasting.. Monthly Demand - The current monthly demand the system determined when running the demand forecast. Best Fit - The system selects the forecast method that best fits the product sales using the Mean Squared Error. The lowest mean squared error considering the most recent year, makes it the best fit. For more information, see How Eclipse Calculates Best Fit. Current Method - Indicates the current forecast method you are using for the selected product.
7	If a row displays as blank, that forecast method was not selected to calculate demand in the Advanced Demand Forecast setup parameters.
8	Graphing options for the selected forecast methods. Select from the following: Current & Best Fit Methods - Graphs only your current forecast method and the best fit analysis. All Methods - Graphs all values as displays regardless of selected methods. Weekly / Monthly - Indicate which values you want to display the last week or the last month's forecasting. Graph It - Click to display the graph.
9	Graphing options for the selected sales. Select from the following: Weekly / Monthly - Indicate which values you want to display the last week or the last month's sales. Years to Display - Indicate the number of previous years for which you want to display information. Graph It - Click to display the graph.

About ACF Scores

The auto-correlation function (ACF) displays visually how the seasonal and non-seasonal scores were determined. Calculating your seasonal items is vital to maintaining your warehouse stock. While forecasting is never perfect, the ACF scores helps you see how Eclipse has calculated your seasonal products to you can better plan and prepare.

The ACF chart represents time spans depending on how you have the Advanced Demand Forecast Parameterscontrol maintenance record parameters set.

The graph provides a visual to see where in the seasonality spectrum your products are falling. You can then address the Advanced Demand Forecast Parameters, if you feel that the results are not accurate.

To view the ACF score:

From the Maintenance menu, select Product to display Product Maintenance.
In the Product field, enter the product or line you want to review.
From the Inventory menu, select Primary Inventory Maintenance.
From the Edit menu, select Advanced Demand Forecast Analysis.
Next to the ACF Score field, click the question mark icon to display the graph.

Audit Advanced Demand Forecasting Data

Use the Demand Calculation Audit to review primary data your system uses for Advanced Demand Forecasting. The information in the Demand Calculation Audit window is compiled from various sources in the system. You can use this information to help you verify and understand the way Advanced Demand Forecasting is handling a product.

Access Demand Calculation Audit as follows:

From the Maintenance menu, select Product to display Product Maintenance.
In the Product field, enter the product or line you want to review.
From the Inventory menu, select Primary Inventory Maintenance.
From the Edit menu, select Advanced Demand Forecast Analysis.
On the Advanced Demand Forecast Analysis window, from the Inquiry menu, select Audit.
The system displays the Demand Calculation Audit window.

On the Demand Calculation Audit window you have the following information:

Days out of Stock - The number of days the products on-hand quantity was zero or less during the forecast period.
Age Used - This is the calculated age used for demand forecasting. This is the result of a base age times the multiplier. You can view the formula by selecting the Information icon.
Date Last Calculated - This is the date Advanced Demand Forecast was last ran on your system.
Demand Period Used - The period used in the calculation for the product. It is a previously established method for determining this number. The system determines this by the length of time the product been on your system and the system settings in the Forecast Parameters For Demand Calculation control record.
Trend Factor Used - This number comes from the Buy Line Forecast Parameters Maintenance window.
Trend Cap Max Increase and Trend Cap Max Decrease - These figures are taken from the Product Advanced Trend Cap Maintenance window.
Exponential Smoothing, Holt, and Holt-Winters buttons - These open the corresponding formula windows. For example, the Holt button opens the Holt Formulas Values window. For more information about these calculations, see About Advanced Forecasting Calculations.