Time disaggregation in demand planning

[Former Member]

I have a planning area that has a storage bucket profile defined in weeks and month. I have getting some incorrect disaggregation in time for one month when I look at the weeks (all the forecast is on the first week). I see in the instructions the following comment

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.

I do not have a time stream definied in my storage bucket profile. Why would all periods but one work correctly

[Former Member]

Thank you for the ideas

I don't think the issue is initalization. The period in trouble in not the last. The period after the problem period works just fine.

I am using P. And it is possible (not likely) that someone could have but a value in the first week. I will check out that idea.

I am interested int the idea in E I had not considered that because I don't want eaqul amounts to the technical periods. I want a disaggregation proportional to the number of days in the technical period. I will go back an look at this option.

Thanks

George

[Former Member]

as an example:

two months both with a forecast of 100 entered at the monthly level

The month that is ok has weekly forecasts of wk1=25, wk2=25, wk3=25, wk 4=25

The month that is not correct has weekly forecasts of wk1=100, wk2=0, wk3=0,wk4=0

George

[Former Member]

Hello George,

I could think of two things

1. Is the KF initialized for WK 2, 3 and 4?

2. What is the time based disaggregation for the KF that is having this problem?

Regards,

Sanju

[Former Member]

George,

Couple of things..

1) If this is the last month, then it is possible that the weeks 2,3,4 are not initialized as mentioned here by Sanju.

2) Is it possible that someone could have manually entered the value at the weekly level in the proportion of 100,0,0,0 if that is the case whatever we enter at the monthly level will continue to be disaggregate as 100%,0%,0%,0%. To check this, try entering 0 in all the weeks for that month. After this enter 0 at monthly level and save it. Then enter 100 at monthly level and save. Now you should be able to see proper disaggregation at the weekly level.

Food for thought, if you are looking for equal time based disaggregation, use the type E - Equal distribution. Based on your explanation above, I get the feeling you are using type P - Proportional Distribution.

Hope this helps.

Thanks

Mani

[Former Member]

Hello George,

Can you pls elucidate your scenario with an example. So that it will help us in helping you better.

Thanks

Mani