Physical stocktaking does not always require the counting of every single item. In practice, alternative labour-saving methods, such as weighing and estimating, are often employed. These techniques, or combinations thereof, are acceptable where specific item types cannot reasonably be counted (e.g. coal stockpiles), or where a full physical count (e.g. due to large quantities) would lead to disproportionate effort or cost.
The prerequisite for using such techniques is the uniformity of the item positions, either in terms of the product itself or its standardised packaging. For instance, where an item is stored in uniformly filled containers of identical contents, it is sufficient to determine the content of one container and extrapolate the total quantity based on the total number of containers. In the case of homogeneous items stored in stacks (e.g. rails, metal sheets, sleepers), a sample can be taken, the average weight determined, and the total volume calculated accordingly. Similarly, for small components (e.g. screws, nails, plugs), counting scales are often used to facilitate quantity determination.
Since such procedures almost invariably result in minor variances, it is considered acceptable not to treat them as errors, provided a predefined tolerance limit (typically 2–3%) is not exceeded. In such cases, the extrapolated figure may be taken as the actual quantity, and no adjustment is made that could distort the valuation outcome.
This handling of tolerable variances is also applied in inventory sampling procedures, particularly when conducting the homogeneous sequential test. This statistical test, based on a binary pass/fail result, would not be applicable if even the slightest variance rendered a stock position unacceptable. In fact, a tolerance margin of up to 5% is commonly accepted in this context.