Inventory Planning > Inventory Forecasting > Processing > Demand History Maintenance > IO Adjust History for Outliers

IO Adjust History for Outliers

You use this program to generate adjustment entries for periods with demand that differs from the mean by more than the maximum allowed.

Periods of exceptional demand (either high or low) are refered to as outliers. Outliers are caused by a number of factors ranging from product promotions to stock-outs, hot weather, cold weather, big sporting events and religious holidays.

You should be wary of using this function for products with seasonal demand, where periods of exceptional demand are not outliers. Another example of being cautious when using this option is where your company lands a large contract which is reflected in an increase in sales for the most recent period of history. However, after adjusting for outliers this increase would be reduced and forecasts would not take the full effect of the contract into account.

The facility to adjust for outliers can be performed from Batch Forecasting and Demand History Maintenance programs. When called by the Demand History Maintenance program it processes the stock code/rev/rel/warehouse currently being maintained. When called from the Batch Forecasting program, it processes all the items that are being included in the batch forecasting run.

Adjustment Process

When you select to adjust history for outliers, the system performs the following for each stock code/rev/rel/warehouse:

  • Calculate the forecasting periods’ start/end dates according to the calendar associated with the item.

  • Check for and delete any existing outlier adjustment entries for the item.

  • Read the sales history transactions (sales, sales returns and sales adjustments) into the forecasting periods.

    [Note]

    If the item being processed uses a proxy, the item is ignored. Similarly, if the item is used as a proxy it is ignored (i.e. if an item's sales are to be reduced as a result of being used as a proxy, they will not be reduced when checking for outliers).

  • If at least two years of history are available, check the item for a seasonal demand pattern.

    If the item is found to be seasonal and the Adjust for Outliers option is run from batch forecasting, the item is ignored.

    If the item is found to be seasonal and the Adjust for Outliers option is run from Demand History Maintenance, a warning is displayed that the item has seasonal demand. You are given the opportunity to either cancel or continue with the process.

  • Calculate the average sales for all the periods since the first period with a non-zero sales quantity.

  • Calculate the standard deviation for all the periods since the first period with a non-zero sales quantity.

  • Calculate the minimum sales quantity allowable as the average minus the ‘number of standard deviations to define outliers’ defined against the stock code/rev/rel/warehouse or the Inventory Optimization Setup (if it has not been defined at the item level)

    i.e. min sales = avg sales – (std dev * no std dev’s to define outliers).

  • Calculate the maximum sales quantity allowable as the average plus the ‘number of standard deviations to define outliers’ as defined against the stock code/rev/rel/warehouse or the Inventory Optimization Setup (if it has not been defined at the item level)

    i.e. max sales = avg sales + (std dev * no std dev’s to define outliers).

  • Check each period’s sales since the first period with non-zero sales.

    If a period has a sales figure less than the minimum allowable, generate an outlier adjustment entry to bring the sales figure up to the minimum.

    If a period has a sales figure more than the maximum allowable, generate an outlier adjustment entry to bring the sales figure down to the maximum.