Measure Demand Plan accuracy¶

Demand Accuracy compares actual sales with the Demand Plan that was official for each period. When a Demand Planning workflow is connected to the Released Demand, the same analysis shows whether each collaboration stage improved or worsened the plan.

Path in the platform¶

Demand Planning > Analytics > Demand Accuracy

Before you start¶

Confirm that:

• the Released Demand contains a non-overlapping official range for every period to be analyzed;
• every connected Demand Plan has the same bucket size as the Released Demand;
• the selected sales source contains closed actuals for the period;
• the requested target unit can be converted for all materials in scope;
• the Released Demand references a workflow if stage-level comparison is required.

Required selections¶

Optional selections¶

Procedure¶

1. Select the Released Demand.
2. Select the Target Unit of Measure.
3. Enter the Start Date and End Date. The end date must not be before the start date.
4. Enable Material Detail or Location Detail only when the decision requires that grain.
5. Open Advanced Filters to restrict the analysis by saved material/location filters, individual IDs, or characteristics.
6. Select Load Accuracy.
7. Start with the consolidated view and drill into the groups that explain the largest error or bias.

Read the results¶

Sales vs Forecast¶

This chart presents actual sales, forecast, BIAS %, and absolute percentage error for the selected workflow stage. Use Workflow Stage to compare the statistical baseline, intermediate collaboration stages, and Final Demand Plan without mixing them.

Workflow Accuracy¶

For one selected period, the workflow charts treat the baseline and Final Demand Plan as totals. Intermediate bars show the incremental change in the metric caused by each stage.

• for absolute error, a negative contribution means the stage reduced error;
• for BIAS %, improvement means moving the result closer to zero, regardless of sign;
• a stage that improves the total may still damage a relevant product, customer, or location group, so always drill into the table.

Dynamic Accuracy Table¶

The table combines period, workflow stage, optional material/location detail, and returned planning characteristics. Measures include Sales, Forecast, Residual, BIAS %, and WAPE %. Use the pivot controls to compare the same metric across stages and periods.

Metric interpretation¶

The residual is calculated as:

Forecast − Sales

When sales are zero, a percentage metric has no valid denominator and is not presented as a zero-error result.

Determine whether a stage adds value¶

Evaluate a stage over several closed periods and economically relevant groups. A stage adds value when it consistently:

• reduces WAPE;
• moves absolute BIAS closer to zero;
• improves priority products, customers, channels, or locations rather than only the grand total;
• adds explainable future information that was not already present in history, model signals, or a previous stage.

A stage destroys value when it repeatedly increases error, transfers error between groups, or adds adjustments with no traceable premise. Use this evidence to redesign responsibilities, approval rules, or the collaboration workflow—not merely to rank people by one isolated period.

Troubleshooting¶

Edition availability¶

Continue¶

• Understand Released Demand
• Analyze a Demand Plan
• Configure the collaboration workflow
• Collaborate on the Demand Plan