APO DP Planning Area Constraints and Sizing

[former_member618787]

Dear Gurus

We need your help to guide us about the Best Practices, and which are the constraints regarding sizing, characteristics combinations amount, and key figures amount, to design an APO Demand Planning area.

Thanks in advance.

[riyazahmed_ca2 · Accepted solution]

Hi Freddy,

SAP Implementation is more of a to-be process setup in the organization. Since we are talking here about APO specific to Demand Planning, I'm sharing you my views with my past experiences.

The major performance bottleneck is the number of CVCs and the initialized time periods. They covers the 70% of performance of APO-DP. Covering this up, I'm putting up my points below

1) Is it Demand Planning or S&OP? If it is S&OP, then we might need to go for Customer Level data. But if it is DP alone, then we can get the aggregated do the planning in terms of Sales Group, Sales Office, Customer Group, Sales Area and so on..

So your CVC need not be on the Customer Level, but on more aggregated level.

2) Attributes like Product Family, Product Group, Customer Country need not be a characteristics as such in the DP. They should be defined as Navigational Attributes. So whenever the things changes in them, it will be automatically changed. If we define them as characteristics, then it will increase the length of CVC and hence the count will increase appropriately.

3) Keyfigures and their aggregations should have its own focus. Unfortunately there is no way to control the number of Keyfigures as they will keep increasing according to the need. But the purpose of the keyfigure has to be well understood and should be well defined for its purpose only.

4) Time Periods and Horizons. Are you sure that you are going to plan for the 5 years in future or you need the historical data of past 5 years? If not then dont build your horizon accordingly. To be frank, with the increased competitions and variability, Product life in Market will not last more than 3 years. So 2 years in the past and 2 years in the future is optimal average. But the industry and domain also plays its own role. Reduced horizon will have appropriate improvement in macro run durations

5) Define separate planning areas for different sets of Industries involved. The Characteristics of one industry need not be same as of another. so defining different planning area for the different scope of the business is optimal

6) Data Views - You can define dataviews as much as you want. But the purpose of Dataview should be appropriately visible in the design. Eg: Lets say we have ''sales overview'' dataview. We can show only Sales History, Cleansed Sales History, Open Customer Order alone. There is no point of showing the forecast or stock keyfigure in them.

7) Defining Default, Start and Exit Macros. They have their own impact in performance and so take atmost care in assigning the macros to either of these types.

😎 Managing Version and Model is another criteria to be checked.

9) Administration Jobs such as Livecache Consistency Checks, Initialization, Reorg should be scheduled periodically to keep the Database in sync with Livecache.

10 No of Work Processes and Parallel Processing profiles has to be given the special focus if you want the increased speed in the work.

Best Regards,

Riyaz

[former_member618787 · Accepted solution]

Hello Riyaz! Thank you very much for all your views. Below I give you details and my comments about your questions.

We use Demand Planning. Indeed we currently have APO DP, SNP, PP/DS and TP/VS already implemented with a time horizon and history below two years. The reason of my inquiry is because of a business requirement, we will have a huge amount of CVC and we fear the impact on LC and overall performance, e.g. when running SDP94, when calculate Split, etc., etc...

May I ask you please what do you think about the pros and cons of design a new area using aggregates at the POS and about using navigational attributes? Bottom line we also would like to know which are the Best Practices, and which would be the best option in terms of performance and LiveCache impact of using our current planning areas, and/or design a new planning área.

Looking forward to hear from you soon.

Thanks in advance.

Freddy

[riyazahmed_ca2 · Accepted solution]

Hello Freedy,

Thats indeed a question for yourself to decide how to proceed further with the Implementation. Minimizing the number of Characteristics is at most need for the performance as said before.

Lets say, you have a scenario where planners like to plan at Product Family Level. Just because of it, You aren't supposed to make Product Family as a characteristics. This will lead to the growth of the CVC length in the table. Better procedure is to make the Product Family as navigational attribute of Product to minimize the CVC length, coz navigational attributes are not accounted in the length of CVC. This is pure technical, but indeed a major player in the performance.

Also as explained earlier, why does Demand Planning needs to be done at Customer Level? No one dares or bothers to go that detailed level. Better procedure is to aggregate the customer level data to a higher structure such as Sales Group or Sales Office. This will reduce the no of CVCs at Product Location level.

Also for SNP, we anyway need the data at Product-Location Level. Customer or Customer Grouping is just an helper in this case to have at most detailed demand.

I have shared most of the details in my previous post. For further best practices, you need to check in SAP support site. There are lot of documents available in terms of DP Best Practices.

All the best..

Regards,

Riyaz

[former_member618787 · Accepted solution]

Thanks a lot!