Inventory sampling can be carried out either with an estimation procedure, or a testing procedure.
In both cases, counting becomes an easier task as a representative sample is drawn from the total of all material numbers or storage bins. Only this sample needs to be counted before the total stock value is extrapolated. The accuracy of the book value is established with the help of mathematical-statistical procedures.
In the estimation procedure, the method implemented by SampleTime, the so-called stock phenomenon is utilized: a small part of the inventory items account for a large part of the total stock value (80/20 rule). These few but high-priced items are counted in full. The software then divides the remaining items into homogeneous strata and calculates the random samples on this basis.
The application operates in a similar way to election forecasts in which the tendency to vote for a particular political candidate is extrapolated for the whole population of a country on the basis of representative polls.
Only the calculated samples are physically counted and the resulting differences are extrapolated across the entire population. This procedure significantly reduces the time, cost and effort compared with conventional counting methods and eliminates much of the human error.