Entering content frame

Function documentation Forecasting with Like Profiles Locate the document in its SAP Library structure

Use

Like modeling uses the historical data from one or more characteristic value combinations to create a forecast for another characteristic value combination.

This topic describes the use of like modeling when running forecasts either in interactive planning or background processing. It does not apply to aggregated lifecycle planning with like profiles. This differs in details. For more information, see Aggregated Lifecycle Planning with Like Modeling.

Integration

Like Modeling is fully integrated with product interchangeability. For more details, see Structure linkProduct Interchangeability in Demand Planning.

Prerequisites

In the forecast profile that you have assigned to the selection you have set the Lifecycle Planning Active indicator.

Activities

You create like profiles and assign them to characteristic values combinations as described in Creating "Like" Profiles and Assignment of Lifecycle Profiles.

When starting a forecast you do not have to undertake any additional steps to start like modeling.

Finding Like Profiles

When executing a forecast the system checks for each characteristic value combination if a like profile has been assigned. A like profile is only taken into account, if exactly the values for the specified characteristic value combination are found in the current selection.

You can use wildcards when assigning like profiles to characteristic combinations. For more details, see Assignment of Lifecycle Profiles and Like-Profiles.

Characteristic Value Combinations

If the system finds a like profile (see above) the characteristic value in the selection is replaced by the one(s) in the like profile.

The system now checks whether the new characteristic value combination exists. If it does not, the system issues an error message and forecasting is stopped.

Calculation of the Historical Data

The system then uses the information (characteristic value(s), factors, action, and possibly weighting profile and lag) to calculate the historical data. See Creating "Like" Profiles and "Like" Profile Examples.

Example

Finding Like Profiles

The assignment of the like profile ‘Like1’ to the combination ‘Product’ = 321, ‘Location’ = 0002 will be considerd in the following cases:

Selection 1:    ‘Product’ = 321

‘Location’ = 0002

Selection 2: ‘Product’ = 321

Location’ = 0002

‘Brand’ = A

But in following cases the like profile will not be considered:

Selection 3: ‘Product’ = 321

‘Product’ = 333

‘Location’ = 0002

Selection 4: ‘Product’ = 321

‘Location’ = 0001

‘Location’ = 0002

Selection 5: ‘Product’ = 321

 ‘Brand’ = A

Selection 5: ‘Product’ = 321

‘Location’ = *

Selection 6: ‘Location’ = 0002

Selection 7: ‘Product’ = 321

 

 

Characteristic Value Combinations

Product P1 has been assigned like profile Like2. This profile refers to product P2 with action S-Sum; the factor is 100%. The following characteristic value combinations exist that include P1

No.

Product

Brand

Location

Customer

1

P1

Super

0001

Gonzalez

2

P1

Super

0002

Smith

3

P1

Premium

0002

Smith

 

The following exist that include P2

No.

Product

Brand

Location

Customer

4

P2

Super

0001

Gonzalez

5

P2

Super

0002

Jospin

6

P2

Premium

0002

Smith

 

If your original selection includes combination no. 1 (P1, Super, 0001, Gonzalez), like profile Like2 replaces P1 with P2. This corresponds to combination no. 4 (P2, Super, 0001, Gonzalez). The system then uses the historical data from this combination, multiplied by the factor 100%. If however your original selection includes combination no. 2 (P1, Super, 0002 Smith), when the system replaces P1 with P2 it cannot find a characteristic value combination P2. Super, 0002, Smith. As a result it issues an error message and stops the forecasting run.

 

 

Leaving content frame