Time Profile Based Disaggregation

[Former Member]

We maintain a Key Figure at Monthly level but need to disaggregate this value to weekly. The time profile is set up the way that it might have 4 days in a week(like 3 full weeks and one 4 days week). We want the system to disaggregate the way so that it allocates less value to the last week that has less days. Is this possible?

[andrew_craven]

Hi Anna,

We had the same problem. Unfortunately there is no way to do this except by defining the week buckets as 'technical weeks'. The constraint is that different periodicities cannot overlap: thus eg month period December 2014 must end 31st December 2014 and cannot be a part of both year period 2014 and 2015 etc. This is required to facilitate simple aggregation disaggregation.

An example of tech weeks and months::

week 43      2014: 20/10      -> 26/10 (Mon-> Sun)

week 44T1 2014:  27/10     ->31/10 (Mon -> Fri)

week 44T2 2014:  01/11     ->02/11 (Sat -> Sun)

week 45      2014: 03/11      -> 09/11 (Mon-> Sun)

This obviously has various shortcomings from a planning/ user perspective!

Regards

Andrew

[Former Member]

Hi Anna,

We have a time profile on weekly bucket and in such a case of split weeks, we are accounting that week to the month on which Monday falls. Therefore we have a 7 day calendar for each week except first and last week of the year.

In your case, you will need to have day level in your time profile allocated to appropriate weeks to meet your requirement.