When a company uses a perpetual inventory system or a periodic physical count, it will find some variances between the quantity found in stock and the amount listed in the inventory database. These variances will occur in the best of companies and are caused by a myriad of problems, the most frequent of which is parts being physically added to or removed from the inventory without a corresponding adjustment to the underlying records. When these variances occur, one should follow the series of reconciliation steps I noted in this post.


Each of the following steps is a filter that blocks out further action at the next step, thereby continually reducing the amount of items to review as one progresses to the next reconciliation step. The steps are as follows:

Step-1. Accept variances with small dollar values. The bulk of all inaccuracies will be for large quantities of small and inexpensive items, such as fitting and fasteners. These are not worth the trouble of a further review, especially when there is a minimal change in the inventory cost, no matter what the outcome of a recount may be.

Step-2. Recount items with large dollar variances. The obvious next step is to recheck the count to see if there was a counting error. If this does not resolve the problem, it is sometimes useful to recount the items in adjoining inventory locations in case there is a problem with a part having been incorrectly stored or counted in an adjacent space. The recount can also be extended to similar products to determine whether an item was mistaken for another part that looks the same.

Step-3. Check the identification. Checking the part number that the counter marked down against the part number in the database for that location sometimes reveals the problem. This is because the part number on the physical part is missing, is mislabeled, or the code is smudged enough to alter its meaning.

Step-4. Check the ownership. A company may have expensive parts in stock that are actually there on consignment and should not be valued. If these items were counted, there will be no corresponding record in the inventory database. One can then ignore the count, because the company does not own the item.

Step-5. Check receiving records. If everyone thinks a part count is low, the answer may simply be that it was never received. Purchasing records may show that a part was due for receipt, but the supplier never sent it. If so, one can go back through earlier listings of the inventory to see when a part was listed as having been received and then compare the first date on which it appeared in the inventory database to the receiving records in that time period to see if there was a corresponding receipt.

Step-6. Review job cost records. It is common for a part to be missing because it was used on product work but was never logged out. For this problem, the first place to look is the job cost records for any jobs that were open during the period when a part was recorded as missing. If the job cost records indicate an unusually high profit, it is likely that a part was not charged to it.

Step-7. Accept the variance. When all else fails, one must conclude that there was either an earlier counting problem that created an initial inaccuracy in the inventory database or that a part is missing because of shrinkage. At this point, it is necessary to record the variance. However, one should keep track of part numbers for which there are unexplained variances on a continuing basis, to see if a pattern emerges that explains the problem.


The preceding investigation process is designed to reduce the inventory reconciliation work to a minimum while still ensuring an accurate inventory valuation. The first few steps either accept inventory counts or call for a quick review, which resolves the bulk of the variance analysis work. Subsequent steps narrow down the range of problems, so that by the time one is reduced to checking on the purchasing and job cost records for a missing part, there are few items for which this much work must be done. Thus, this system results in accurate inventory records while spending the smallest amount of time on inventory variance reconciliation.