Skip to Content

Interactive Forecasting Disaggregation with Phase Out APO Demand Planning

When running Stat Forecast interactively, user loads up a selection profile with data at an aggregated level.(For instance one product with many plants, customers etc below it)

User runs Statistical Forecast interactively.

When reviewing the messages tab in interactive forecasting we observe that within the selection there are some qualifying combinations that have Phase Out Profiles assigned.

User saves results of interactive forecast and returns to interactive planning screen.

When user drills down to detailed level they observe that in sporadic weekly buckets, a portion of the aggregated forecast is assigned to a combination with no sales history. (and I have verified 0% proportional factors assigned)Why would this combination be allocated a % of the forecast when Prop Factors suggest it should get any.

My further findings:

I load the results into a data view which allows direct editing of Statistical Forecast.

I display Statistical Forecast and Proportional Factors Key Figure and confirm that stat forecast is assigned tosome combinations with 0% proportional factor

In one of the weekly buckets with the incorrectly allocated forecasts, I overwrite the aggregated forecast with a manually entered number.

I press enter and then review how the system disaggregated the manually entered number.

I observe that still a portion of the manually entered forecast has been disaggregated to some CVCs whose proportional factor is 0 %.This should nbot happen.

Next I delete the manually entered aggregate forecast in that weekly bucket and press enter which in turn blanks out the forecast at lower levels.

I then manually enter a new value in stat forecast at the aggregated level.

I press enter and now I see that the newly entered value has disaggregated as I expect.It respects the proportional factor % and does not allocate any of the stat forecast to those combinations.

Now if I run the same selection and same aggregation level as a background job, I do not see this issue at all.The disaggregation happens exactly as I expect it to.

Add comment
10|10000 characters needed characters exceeded

  • Get RSS Feed

2 Answers

  • Best Answer
    avatar image
    Former Member
    Feb 28, 2017 at 09:00 AM

    Hi Adam,

    The fact that you didn't see the issue after zeroing out/cleansing the keyfigures explains why SAP recommends cleansing out of target keyfigures as best practice. Before you calculate the proportional factors or copy the keyfigures manually or using macros,key figure should be zeroed out. This is done to nullify the existing internal factors if any otherwise it might get disaggregated according to keyfigure calculation type and time based disaggregation.

    Hope this helps.

    Thanks & Regards,


    Add comment
    10|10000 characters needed characters exceeded

  • Mar 15, 2017 at 02:11 AM

    Thank You Sonali. I have never seen this behaviour before. Nor have I ever seen a recommendation from SAP that target key figures be initialized before populating them using the interactive statistical forecasting. However if you are saying that this is a best practice we can certainly develop a process to do this. Are you aware of a note that specifes this or is it just accepted knowledge?


    Add comment
    10|10000 characters needed characters exceeded