DP time bucket conversion

[Former Member]

Hello,

For statistical reason, our statistical forecast is made with monthly time bucket (Gregorian calendar). The business requirement is to have converted the results in fiscal period month. After conversion based on the DP logic we are loosing 5-7 % from our accuracy.

How we can fix that difference?

Any idea would be appreciated.

Thanks, Marius  

[Former Member · Accepted solution]

Marius,

If you are forecasting at an aggregated level (most companies do), and you have chosen the conversion method of disaggregating months down to days, and then re-aggregating back up to fiscal periods, then 5-7% difference is not unreasonable.  You might be able to achieve better precision by tweaking a few things, or by creating a custom ABAP conversion routine.

Seriously, though, how much effort do you really want to spend on this?  High levels of precision in a forecast are generally inconsequential.  Remember, it is just a forecast, which is a prediction of the future.  When you consider the most common forecast KPI, ( forecast vs actual), a world class forecasting process will only yield 80% accuracy.  In many companies, using a forecast that is 'imprecise' by 5% is generally not worth the time and money to improve.

Best regards,

DB49