Demand Forecasting Configurations And Process Flow



Demand forecasting options:

-Based on historical transactions (loaded via DMF in staging table)

  • Example: New product that was maintained in separate system prior.

-Based on historical transactions in D365FO

  • Note: This option is only useable with minimum and maximum forecasted value method.

-Based on historical transactions via Azure AI Machine learning web service

  • Note: Here, a user can leverage additional methods beyond minimum and maximum forecasted value and leverage custom ML studio web services for forecasting.


Configurations and Setups:

Item Allocation Key

  • A demand forecast is calculated for an item and its dimensions only if the item is part of an item allocation key. This rule is enforced to group large numbers of items, so that demand forecasts can be created more quickly.

Demand forecasting parameters

  • Demand forecasting runs cross-company, the setup is global. (applies to all companies)

  • Demand forecasting generates the forecast in quantities. Therefore, the unit of measure that the quantity should be expressed in must be specified in the Demand forecast unit field. Note: Thus, for every unit of measure that is used for SKUs that are included in demand forecasting, make sure that there is a conversion rule for the unit of measure and the general forecasting unit of measure.

  • Demand forecasting can be used to forecast both dependent and independent demand. For example, if only the Sales order check box is selected in the parameters, and if all the items that are considered for demand forecasting are items that are sold, the system calculates independent demand.

  • However, critical subcomponents can be added to item allocation keys and included in demand forecasting. In this case, if the Production line check box is selected, a dependent forecast is calculated.

  • Forecast generation strategy: There are two methods for creating a baseline forecast. You can use forecasting models on top of historical data (Azure Machine Learning), or you can just copy over the historical data to the forecast.

  • Forecast dimensions: You can also select the set of forecast dimensions to use when the demand forecast is generated. A forecast dimension indicates the level of detail that the forecast is defined for.

When Azure machine learning is selected:

  • To generate the forecast via AML, Supply Chain Management uses a Machine Learning web service. To connect to the service, you must provide the following information if you sign in to Microsoft Azure Machine Learning Studio (classic):

  • Web service application programming interface (API) key

  • Web service endpoint URL

  • Azure storage account name

  • Azure storage account key

Forecast algorithm parameters for Azure AI can be set in parameters for all or just for specific item allocation group:

  • To view the parameters that can be configured for the demand forecasting service, go to Master Planning > Setup > Demand forecasting > Forecasting algorithm parameters. The Forecasting algorithm parameters page shows the default values for the parameters.

  • Forecast algorithm setup options for Azure AI:

  • Confidence level percentage: A confidence interval consists of a range of values that act as good estimates for the demand forecast. A 95% confidence level percentage indicates there is a 5% risk that the future demand falls outside the confidence interval range.

  • Force seasonality: Specifies whether to force the model to use a certain type of seasonality. Applies to ARIMA and ETS only. Options: AUTO(default), NONE, ADDITIVE, MULTIPLICATIVE.

  • Forecasting model: Options: ARIMA, ETS, STL, ETS+ARIMA, ETS+STL, ALL. To select best fit model, use ALL.

  • Maximum forecasted value: Specifies the maximum value to use for predictions. Format: +1E[n] or numeric constant. (Can be used for forecast method based on historical transaction.)

  • Minimum forecasted value: Specifies the minimum value to use for predictions. Format: -1E[n] or numeric constant. (Can be used for forecast method based on historical transaction.)

  • Missing value substitution: Specifies how gaps in historical data are filled. Options: numeric value, MEAN, PREVIOUS, INTERPOLATE LINEAR, INTERPOLATE POLYNOMIAL.

  • Missing value substitution scope: Specifies whether the value substitution applies only to the data range of each individual granularity attribute, or to the entire dataset. Options: GRANULARITY_ATTRIBUTE(default), GLOBAL.

  • Seasonality hint: For seasonal data, provide a hint to the forecasting model to improve forecast accuracy. Format: integer number, representing the number of buckets a demand pattern repeats itself. For example, enter "6" for data that repeats itself every 6 months.

  • Test set size percentage: Percentage of historical data to be used as a test set for forecast accuracy calculation.

Process flow - Demand forecasting:

  • Optional – Demand import: If the historical demand data isn't already existent, use the Historical external demand (ReqDemPlanHistoricalExternalDemandEntity) data entity in Dynamics 365 Supply Chain Management to import it. (Example: New product that was maintained in separate system prior.)

  • Optional - Outlier removal: Click Master planning > Setup > Demand forecasting > Outlier removal to open the Outlier removal page, where you can use a query to select the transactions to exclude.

1. Baseline: When you create a baseline forecast, you must first specify the parameters and filters that are used in the calculation.

  • To generate a demand forecast, go to Master planning > Forecasting > Demand forecasting > Generate statistical baseline forecast.

  • The forecast bucket can be selected at forecast generation time. The available values are: Day, Week, and Month. (set in forecast horizon)

  • If copying over from historical demand is selected, then the run is copying historical demand from a certain date forward, production planners can make the plan for the next quarter in two ways:

  • By copying the demand from the same quarter last year.

  • By copying the demand from the previous quarter.

  • Manual adjustments made in previous demand forecasting iterations can be automatically applied to the new baseline forecast if the ”Transfer manual adjustments to the demand forecast” check box is selected. If the check box is cleared, the manual adjustments are not added to the baseline forecast.

2. Manual Adjustments:

  • After the baseline job is ran, a user can modify the forecast on the “adjusted demand forecast” form and override the value with an adjusted forecast value. The edited cell immediately becomes bold to indicate that the forecast that it shows isn't the forecast that the demand forecasting service created.

3. Viewing details of the forecast:

  • The Demand forecast details page shows the following information in graphical and tabular formats:

  • The historical demand that the forecast predictions are based on.

  • The current forecast that is used by Master planning.

  • The new demand forecast values and the amounts they have been manually adjusted by.

  • The confidence interval for the forecasted values.

  • The forecast model that was used to generate the forecast. If you're viewing aggregated data, you will see the list of all the methods that were used for all the aggregated time series.

4. Authorize adjustments

  • On the Details tab of the Authorization page, you can view details about the forecast that was most recently generated. You can select the companies and the forecast models to authorize the forecast for use.

5. Monitor forecast accuracy:

  • Supply Chain Management calculates the following types of forecast accuracy:

  • Historical forecast accuracy, by comparing the historical forecast that Master Planning uses with the historical demand. To view the values (both absolute values and percentage values) for historical forecast accuracy, click Show accuracy on the Demand forecast details page.

  • The estimated accuracy of the forecasting model that is used to generate the predictions. You can view the accuracy percentage under Model details on the Demand forecast details page.

6. Include forecast demand in forecast master plan:

  • When you include a forecast in a forecast master plan, you can select how the forecast requirements are reduced when actual demand (sales orders) is included.

  • None

  • Percent – reduction key

  • Transactions – reduction key

  • Transactions – dynamic period

  • Note: You can include supply/demand forecast also in static/dynamic master plan.

7. Run forecast planning

  • On the forecast planning form, you can select forecast plan and filters to be used for calculating forecast requirements.

8. Firm planned order in forecast plan:

  • On the planned orders form, you can select planned forecast requirements and approve/firm those into orders.







Recent Posts

See All

Agenda: 1. Overview 2. Internal Educational Assistance Approvals via Teams 3. D365FO PO approval via Teams Approval Application 1. Overview Approvals in Microsoft Teams is an simple way to streamline