Dynamic safety stock

[Former Member]

We can calculate the dynamic stock in R/3 using the following formula:

Safety Stock = MAD * R * SQRT(D/30) + .5

Where MAD = Mean Absolute Deviation from forecasting

R = Service rate derived from the service level defined in MRP2 tab

D= Delivery time

Instead of using the MAD from R/3, I use the MAD from forecasting in APO. I use a function module and lookup the MAD in APO and pass it on to R/3.

My question is: Is this logically correct??? For the same set of data and same forecasting model and paramerts like alpha, I get a high MAD in R/3, so high sasafety stock while the MAD is never so high in APO, so I am getting a low safety stock.

Am I doing anything wrong??? is ti logically correct to use APO MAD instead of R/3 MAD?

[Former Member]

The calculation of MAD depends on the forecast model that is initialized in APO. The MAD for a constant model for eg is 0

that for a trend model = E|basic-historyi)/n

(check the help on model initialization)

check what is initialized for your user or DEFAULT_USER

If you are happy with the calculation in APO you should be ok to use it.

btw if you have the formula you can set it up to calculate the full Safety Stock in APO itself and transfer the calculated figure to R3, right?

also, if you want to do the calculation in ECC, instead of using the function module you can use a macro to calculate the MAD and send it to ECC through BW.

MAD( X1 ; ... ; Xn ) returns the mean absolute deviation of the values of the values X1 through Xn. The arguments can be numbers that you enter in Operator/function dialog boxes, as well as cell(s), row(s), column(s) or area(s) from the planning table or the auxiliary table.

[srinivas_krishnamoorthy]

It seems you are seeing difference in logic of MAD calculation in APO and R/3 for the same history data. Do you see the same forecast data in both the systems. In APO, MAD is calculated on the ex-post zone of forecast that depends on the forecast model that is selected. You should be able to back calculate on how MAD gets calculated using an excel sheet containing history and ex-post forecast. (The formula used in APO is in

http://help.sap.com/saphelp_scm50/helpdata/en/ac/216ba1337b11d398290000e8a49608/frameset.htm)

In R3

http://help.sap.com/saphelp_470/helpdata/en/7d/c27a21454011d182b40000e829fbfe/content.htm

There could be little differences in formula used. I would not use formula across systems. So use R3 MAD in R3 and APO MAD in APO.

[Former Member]

Srinivas,

Thanks for the reply. I will read through the links. By the way, the forecast generated in APO is always higher than the one in R/3. The R/3 one was about 125 and the APO generated 175 for the same history.

By "I would not use formula across systems", are you saying that I should not use mad in APO to calculate the dynamic safety stock in R/3??