Disaggregation of data in DP

[former_member847114]

Hi all,

My client has a requirement of Disaggregating the forecast(not statistical but will be a manual input from the marketing team) at all levels starting from the highest level of product customer and loc to the lowest level of material/customer/loc and from periodicity of 1 year to months/weeks/split weeks . The disaggregation should be both time and structural and should be based on the previous history. The requirement is to see the proportions in both percentages and values for all the levels and the planner should be able to make changes to the proportions and the total should add up to 100%.

For example

System suggested proportions

Level 1 - 100%

Level 2 - 50% 20% 30%

Planner changes the proportions for the  to the following

Level 2 - 30% 30%

Will the percentage for the third product be automatically calculated as 40%? Or should the planner fix the 100% or the value on the level 1?

Please suggest any best practices and the detailed process of how to configure this in DP. Are any macros required to perform this function?

Sunny

[Former Member · Accepted solution]

It's all in the KF settings for the planning area. Just enter the dis-aggregation KF - APODPDANT or the sales history in both types of dis-aggregations. I believe you can display system calculated proportional factors in percentages.

[former_member847114 · Accepted solution]

Thank you for the response

How is the result going to be different if I  base the forecast KF on  APOPDANT and on History KF.  Which KF works best for time and structural disaggregation?  Also if the planner wishes to change the values, how will the system re calculate to get the total percentage 100%.

For example

System suggested proportions

Level 1 - 100%

Level 2 - 50% 20% 30%

Planner changes the proportions for the  to the following

Level 2 - 30% 30%

Will the percentage for the third product be automatically calculated as 40%? Or should the planner fix the 100% or the value on the level 1?

Please suggest any best practices and the detailed process of how to configure this in DP. Are any macros required to perform this function?

Any inputs are greatly appreciated

[Former Member · Accepted solution]

The proportional factors are just numbers - they are not percentages. The system adds all numbers and then divides each factor with the total internally for dis-aggregation.

For example:

Level 1 product group A total demand = 200

Level  2 Product x - demand                = 75

Level  2 Product y - demand                = 25

Level  2 Product x - demand                = 100

The APODPDANT is based on history should look like 75, 25, 100.

As you can see these are not percentages.

Next, the planners adjusts the factors to , say, 60, 50, 40. 

Then the dis-aggregation will be:

200 x 60

----------------

(60+50+40)

200 x 50

----------------

(60+50+40)

200 x 40

----------------

(60+50+40)

totaling the split demand to 100.

As usual SAP terms are very confusing. The prop factors are just number used to derive the ratios.

They are not the ratios or the percentages themselves.

The same applies to time based dis-aggregation.

Does this help?

[former_member187488 · Accepted solution]

Hello Sunny,

If you want to base your forecast KF on the history KF, you must have the data populated in history KF for the future buckets.

As long as you have the data in the KF to behave as proportional, there's no difference when you use any KF.

There's a functionality in the planning book to swtich between absolute display and percentage display.

As you said, if you want to keep the value at aggregated value unchanged, you'll need to fix it, or write a macro to apply your own logic.

Best Regards,

Ada

[former_member847114 · Accepted solution]

Thank you for your reply.

As I understand APOPDANT is only based on volumes and not percentages. Can E/E icon be used to switch between % and absolute values for APOPDANT?

In case the percentages can be seen then in APOPDANT the percentages look like.

Level 1 product group A total demand = 330

Level  2 Product x - demand                = 25% = 82.5

Level  2 Product y - demand                = 25% = 82.5

Level  2 Product z - demand                = 50% = 165

Planners would like to change only the Product X % to = 40% and do not touch the rest. Will the system generate proportions automatically for Product y and z as 20% and 40% respectively in the ratio of 1:2? Expected results would be

Level  2 Product x - demand                = 40% = 132

Level  2 Product y - demand                = 20% = 66

Level  2 Product z - demand                = 40% = 132

Or is a complex macro required for this?

I am not sure about this but the APOPDANT works for only structural and not for time. For time disaggregation in KF setting the Forecast KF should be based directly on History KF  instead of APOPDANT and in this case the manual maintenance of time disaggregation factors is not possible.Is this right?

Here my struggle is how to design the system so that the planners can see what system suggest with regards to both structural and time and also maintain their own proportions and save the data.

One idea that I have is to have a separate planning book with 2 KF.

System suggested proportions KF(APOPDANT) -

Manual maintenance of proportions KF -

I will copy the values from APOPDANT to manual maintenance KF. In KF settings The APOPDANT will be based on History KF and Forecast KF will be based on Manual Maintenance of proportions KF. Do I have to maintain different PB for Level 1 to Level 2 and Level 2 to Level 3 and Level 3 to Level 4 of the product hierachy or Can I just have one PB and manage all the proportions depending on the selection ID? DO you have a better simple solution? Please reply!!

Sunny

[Former Member · Accepted solution]

Sunny,

This is just a quick reply to pass some hints on to you and I will simulate and give you a complete answer shortly.

When  you switch to percentages the system should auto adjust the remaining to keep the total to 100%. If you work with absolute PFs , then the total at the aggregate level needs to be fixed in order to force the adjustments of the disaggregated vlaues instead of increasing / decreasing the tolal.

If you want the system can use the APODPDANT for time-based disggregation.

If you stop and think carefully about it, the structural and time based disagg should be based on ONE KF so as to keep the data integrity and ease of understanding.

I think what you are trying to do is supported out of the box, if I understand your requirements.

Do you want me to set up a test case for you and possibly discuss try different scenarios out via a web-ex?

[former_member847114 · Accepted solution]

Hi,

Thank you very much for your concern and tyring to understand the problem. Can you please set up a test case and let me know. I will also test some scenarios and we can discuss.

Regards,

sunny