Aggregation Level of Forecast

Probably this question came into your mind

What is the best way of aggregating the forecast?

There are 3 levels to the forecast aggregation:

1- Product.

2-Sales channels.

3-Location.

Each level of those could have a long tree under it.

Product Hierarchy:

The lowest level starts with SKU then moves higher.

(SKU/Format or pack/ Family or group /Brand/Category).

Sales channels:

The lowest level is the retail store.

(Retail store/Customer name/ Region/ Country).

Location

Don't be confused between the location aggregation and the sales channels.

Location aggregation is more related to the internal supply chain network organization.

The main factor that determines the location aggregation is the mission of the organization.

Have you asked yourself why there are multiple distribution centers and why there is a CDC and small DC's?

What is the logic behind it and how it was built?

There is a significant factor that determines and optimize the supply chain network:

Translating the organizational strategy whether its (Product driven/ customer-driven/forecast-driven/demand-driven) into a supply chain strategy.

If the organizational strategy is customer-driven, then the supply chain network will be established based on the right frame that will fulfill that organizational strategy not just optimizing the supply chain cost.

For example, an organization wants to lift up the service level for a customer where his stores are scattered across a huge geographical area, then instead of consolidating one mega CDC (central distribution center), the organization has to split this mega CDC into two CDC.

One is in the north and the other in the south, to reduce the ordering cycle time for this customer in addition holding safety stock that could cover demand uncertainty as per the below.

The last question that can come into your mind

What is the best way to aggregate the forecast at the location level?

Is it better to forecast the total national echelon?

Is it better to forecast at the DC level?

There are many factors for that decision and depends on whether it's (Retail or distribution model).

These factors are:

A) Organizational strategy which should be translated into organizational structure especially the sales department to stress the focus on.

B) Supply chain strategy which translates the distribution network design from the sourcing plant till the last point of Echelon.

C) S&OP time spent based on an agreed aggregation of "Product and sales channels" Vs. the Output.

The output here is the aggregated error either at the total national level or by store.

For instance, if the time spent on S&OP is too much and the Forecast accuracy is still low, then it's needless to spend this time on planning and stepping up on aggregation is not an option to lessen the time, hence recalculate the forecast accuracy again.

D) Re-distribution factor is a big part of aggregating the forecast based on the total national Echelon level.

For example, you can't aggregate the total forecast on 7 stores however you can't re-shuffle distribution accordingly as this will lead to some inventory health challenges in addition to some service-level issues due to aggregating the stocks and forecast.

Aggregating the forecast decision could be a game-changer for organizational thriving.

Here is the link for my Udemy courses:

https://www.udemy.com/user/ahmed-khaled-912/