This page enables you to identify outliers in the historic forecast error, that is forecast vs actual demand, and determine ways to correct the data. It also enables you to review the impact of error correction in unit terms resulting from both system and manual correction at a location and product level.

Before you start

Use the context selectors at the top-right to select a location and product combination. Additionally, you can manually enter the required values using the search function.

Use the Product Filter field to filter on a specific product.

Forecast Error Correction Charts and Tables

The chart and table helps you focus on variances by exception and correct as required. The parameters are used to propose forecast error correction through an outlier detection process. The chart let's you see how correction affects your data. 

You can view outlier parameters in the Outliers Table.

ParametersValues
ABS Error (U)The absolute Forecast – Actual error.
Period by PeriodIf Period by Period correction is enabled, the table becomes editable to select which periods to correct. 

You can enter comments in the Comments box.

The Forecast Error Correction spike chart displays the impact of error correction, with a clear view of before and after correction.

ParametersValues
ABS Error (U)The absolute Forecast – Actual error.
Mean ABS Error (U)The mean value of the absolute error. 
Upper LimitThe user-defined parameter set in Outlier Magnitude. It specifies the number of standard deviations from the mean that should be used to determine an outlier. Any values which exceed this upper limit are auto-corrected by the system.  
Corrected ABS Error (U)The value once the forecast error correction is applied. 

The Forecast Error Correction table displays:

ParametersValues
ABS Error (U)The absolute Forecast – Actual error.
Period by PeriodIf Period by Period correction is enabled, then outlier periods can be manually corrected as per the Correct To parameter. 
System Corrected Error (U)The value for the selected product-customer combination once the history correction has been applied.
Manual Correction (U)Enter a value to manually override the System Corrected History (U) value. 
Corrected ABS Error (U)Displays the System Corrected Error (U) value, unless a value is specified in the Manual Correction (U) field. 

Parameters

You can override the parameter values locally, for a product and customer combination, by setting values in the Edit Parameters card:

ParametersValues
Correct From

Specifies the baseline for correction. 

  • Off: Prevents the application from calculating outliers for correction.
  • From Mean: Sets the application to create outliers based on values being greater than x standard deviations from the mean.
    In the case of From Mean, the values are calculated using periodic data starting from the product’s effective date and ending with the last effective period.
Outlier MagnitudeSpecifies the number of standard deviations from the mean (or standard errors from the trend line) that determine an outlier. A value of 2 highlights any data points that are more than 2 standard deviations from the mean (or standard errors from the trendline) as outliers. 
Correct To

Specify how an outlier is corrected:

  • To Limit: Corrects outliers to the upper or lower limit.  
  • To Mean: Corrects outliers to the Mean.  
Period by Period

Select to identify all outliers but enable  you to manually select which to correct.

  • If set to No, all outliers are automatically corrected as per the Correct To parameter.
  • When set to Yes, each outlier period can be reviewed on its own, and a decision taken as to whether to correct.