Aggregation and Disaggregation

Aggregation and Disaggregation

Use

Aggregation is the automatic function by which key figure values on the lowest level of detail are summed at run time and displayed or planned on a high level; for example, if you display the forecasted demand for a region in the interactive planning table, what you see is the forecasted demand for all of the different sales channels, product families, brands, products, and customers in that region which the system has added together.

Disaggregation is the automatic function by which a key figure value on a high level is broken down to the detailed level; for example, if you forecast demand in a particular region, the system instantly splits up this number among the different sales channels, product families, brands, products, customers, and so on, in this region

Aggregation and disaggregation ensure a consistent planning approach throughout your organization. The sum of the details on the low level always equals the total on the high level. Key figure values are always saved at the lowest level of detail. If aggregates exist, the data is also saved on the aggregate level.

Features

Aggregation and disaggregation of each key figure are defined by its calculation type and its time-based disaggregation type. You choose from the following calculation types:

S - Pro rata

Pro rata disaggregation works in two ways, depending on whether you are creating or changing the demand plan:

¡ If you create data on an aggregate level, the data is distributed to the lowest level of detail in equal proportions.

¡ If you change data on an aggregate level, the values at the detailed level change so that each one represents the same proportion of the aggregate values as before.

P - Based on another key figure

The data is distributed to the lowest level of detail in the same proportions as those that can be derived from the values of another key figure. For example, you can apply the proportional distribution of the key figure "Baseline/statistical forecast" to the key figure "Manual adjustment". Note, however, that if no values exist for the key figure "Baseline/statistical forecast", pro rata disaggregation is used.

If you enter 'P', you must also specify on this screen the key figure on which disaggregation is based. Do this in the field "Disaggregation key figure".

A - Average of key figures

The average of the key figure values at the next lower aggregation level is displayed as the result at runtime. Use this option for key figures that represent percentages (for example, the percentage difference between this year's and last year's forecast) or sales prices (for example, sales price per unit). If you enter a value on an aggregate level, the system disaggregates it by copying the value to the details on the lower level.

Aggregation with Calculation Type A

D – Average at lowest level of detail

In contrast to calculation type A, the system uses the values at the most detailed level, that is the level for which each characteristic has a single value assigned, to form the average. This can result in different numerical values in comparison to type A.. Disaggregation is exactly the same as for calculation type A.

Aggregation with Calculation Type D

If values already exist at detail level and you then change a value at aggregate level, the system disaggregates the changed value according to the existing proportions and not in the manner described above. This is valid for both type A and type D.

Since both average calculation types are generally used for descriptive purposes, for example the price per unit, you use them mainly in conjunction with the time-based aggregation type N (see below). If however you do use time-based aggregation/disaggregation, you may observe slight differences in the rounding behavior between calculation type A and calculation type B.

N - No disaggregation

Disaggregation costs performance, so set this option for text rows, auxiliary rows, and other rows that do not need disaggregating.

I - Pro rata, except for the initial disaggregation, which is based on another key figure

If you create data on an aggregate level, the data is distributed to the lowest level of detail in the same proportions as those that can be derived from the values of another key figure. For example, you can apply the proportional distribution of the key figure "Baseline/statistical forecast" to the key figure "Manual adjustment". If there is no value for the basis key figure, the values are disaggregated evenly.

If you enter 'I', you must also specify on this screen the key figure on which disaggregation is based. Do this in the field "Disaggregation key figure".

If you change data on an aggregate level, the values at the detailed level change so that each one represents the same proportion of the aggregate values as before.

The time-based disaggregation type defines how planning data is disaggregated in time. The time buckets in which data is saved are given by the storage buckets profile. For example, if you select months and weeks in the storage buckets profile, data for June and July 2000 is saved in the following storage buckets:

Example of Storage Buckets

Dates	Days	Storage bucket
June 1-4	Thursday through Sunday	4 days
June 5-11	Monday through Sunday	7 days
June 12-18	Monday through Sunday	7 days
June 19-25	Monday through Sunday	7 days
June 26-30	Monday through Friday	5 days
July 1-2	Saturday and Sunday	2 days
July 3-9	Monday through Sunday	7 days
July 10-16	Monday through Sunday	7 days
July 17-23	Monday through Sunday	7 days
July 24-30	Monday through Sunday	7 days
July 31	Monday	1 day

You choose from the following time-based aggregation types:

P - Proportional distribution

Data is distributed in time so that each key figure value in the smallest storage bucket represents the same proportion of the value in the aggregate bucket as before.

If the key figure values prior to distribution were zero, and if a time stream ID forms part of the storage buckets profile definition, the system checks to see if time-based weighting factors exist for this time stream. If so, the data is distributed according to the time-based weighting factors of the time stream. If no factors exist, the data is distributed equally to each storage bucket.

E - Equal distribution

Data is distributed equally to each storage bucket.

This option is available only for key figures that were not defined as fixable in the Admin Workbench.

N - No disaggregation in time

The value in the planning bucket is copied to the storage buckets; for example, a planning value of $100 for the month of June is copied to each of the storage buckets June 1-2, June 5-9, June 12-16, June 19-23, June 26-30. If you display the planning value for June at run time, the system shows the average value of the storage buckets.

This option is available only for key figures that were not defined as fixable in the Admin Workbench.

K- Based on another key figure

This option is basically the same as calculation type P.

The data is distributed to the lowest level of detail in the same proportions as those that can be derived from the values of another key figure. For example, you can apply the proportional distribution of the key figure "Baseline/statistical forecast" to the key figure "Manual adjustment". Note, however, that if no values exist for the key figure "Baseline/statistical forecast", no disaggregation takes place.

If you enter 'K, you must also specify on this screen the key figure on which disaggregation is based. Do this in the adjacent field "Disaggregation key figure".

We recommend that if you use option 'K' for time-based aggregation, you also use calculation type 'P'.

L- Read: Value from last period; Write: No allocation

This aggregation type is intended for time series SNP.

When you aggregate from shorter time buckets to larger, for instance weeks to months, the system take the value from the last period and copies it to the larger time bucket. In the reverse case, the value is written to all the smaller time buckets.

Activities

You set the Calculation type and the Time-based disaggregation type of a key figure in Administration for Demand Planning and Supply Network Planning. These settings apply to all planning books in which the key figure is used.

See also Example of Disaggregation and Rounding.