Question

What are the levels of aggregation in forecasting for a manufacturing organization? How this hierarchy of forecasts should be linked and used?

Answer 1

In forecasting, the data is either aggregated or disaggregated to transform data to get additional information for the current trend series. This helps in obtaining forecasting accuracy. The level of aggregate is largely dependent on the data collected and the objective of the forecasting. Hierarchies should be effectively linked with the different levels of hierarchy to make accurate forecasts. The common hierarchies in forecasting in a manufacturing firm are-

- Items – those that might have the highest errors

- Product group –has moderate errors

- Country/region – has lower error rates.

Each of the above mentioned hierarchies is an attribute which has different levels of aggregation. These are just examples; however the hierarchy is dependent on the requirements of the manufacturing firm. For example, the price per unit of the item used in the manufacturing can be used for different level of aggregation. For each item input, the outcome or output can be aggregated at each change in the unit. There are two types of levels of aggregation, namely top-down and bottom-up approach. In this top-down approach, a summary forecast is prepared based on the highest level of aggregate data. This summary forecast is then associated with the individual items used in manufacturing that is directly related to the aggregate. Thus, it is pertinent to say that hierarchies have a direct influence on the levels of aggregation and they help in better forecasting.

Note- If you liked the answer, please give an up-vote, this will be quite encouraging for me, Thank you.