As a full inventory is time-consuming and often not feasible (e.g. inventory of an automated high-bay warehouse on a specific date), a random sample inventory is often carried out. A representative quantity of the total is recorded and extrapolated to the entire inventory using mathematical-statistical methods.
In many cases, higher-value items are subject to a full inventory, while lower-value items are subject to an S. See also Inventory.

Source: logipedia / Fraunhofer IML