Proportional factor usage and rounding

[Former Member]

Hi,

If I have dis-aggregate "Corrected Forecast" KF and then aggregate proportionally at all levels after planning , should I use proportional factors using APODPDANT KF or Pro-rata distribution ?

Which dis-aggregation rule minimizes effect due to rounding ?

thanks in advance,

subram

[Former Member · Accepted solution]

Subram,

Pro-rata give the least problems with rounding, but that is irrelevant.

You MUST use the type of disaggregation that makes business sense. If you enter your KF at the highest level (e.g. Total), and you use pro-rata, then the total will be disaggregated equally amongst all the sublevels, and likewise from sublevels to sub-sub levels. This is almost never appropriate for a sales forecast. Normally you distribute the total to the lower levels according to some business logic (such as using the proportions found in sales history). Otherwise, the lower levels will be meaningless.

Best Regards,

DB49

[Former Member · Accepted solution]

thanks. Few clarifications required

1. If i dis-aggregate forecast KF from highest level with Pro-rata, does it aggregate same way when plan changes at lowest level.

2. Can I reduce rounding effects by defining no of decimal places in planning book and using Pro-rata

Documentation says APODPDANT KF for proportional factors is used only for authorization purpose.So in this case, i can just use pro-rata.

regards,

subram

[Former Member · Accepted solution]

Subram,

If i dis-aggregate forecast KF from highest level with Pro-rata, does it aggregate same way when plan changes at lowest level.

Generally yes. When you enter data at an aggregated level, the data is immediately disaggregated to the detail, and is stored only at the detail level (there are a few exceptions, but this is the rule). Any time you edit the planning area, at any level, the data is disaggregated and aggregated on the fly to all levels.

Can I reduce rounding effects by defining no of decimal places in planning book and using Pro-rata

It depends upon what you mean by rounding effects. We are not talking about accounting here (e.g., money) we are talking about forecasts, which are inherently inaccurate to begin with. There is little value in making a forecasted number more precise, that effort is better spent on making it more accurate (a better match to reality). Only a fool would spend a lot of effort trying to make every detail entry, when summed up, to equal EXACTLY the total at the subtotal and total levels.

The biggest problems with rounding are usually related to people creating and using Macros that are not suitable for extremely small values.

Best Regards,

DB49