The principle behind inventory sampling is that the count results of the randomly selected stock management units are extrapolated to all stock management units of the sampling area. Moreover, the count results of the complete-count area are taken into account in the total calculation. The system determines an estimated value for the stock population of the inventory sampling from both results.

There are two types of extrapolation:

• A provisional extrapolation exists if differences are posted only for part of the stock management units to be counted. You can create as many provisional extrapolations as you want. Only the last one you create is saved.

• A final extrapolation exists when all differences have been posted, that is to say, all count results are included in the calculation, and no subsequent random selection is necessary. A final extrapolation can only be performed once because no new results can be added.

Various possible mathematical procedures can be used for the extrapolation. The SAP System currently supports the layered mean-value estimation procedure. The following calculations are performed during layered mean-value estimation, based on the values of the posted sample elements:

• The estimated average actual value for each stratum
• The estimated actual value of the total