The Lean manufacturing, PPT Plus,
"Back-to-Basics" e-Training Package 
Lean Manufacturing Solutions

Increase the effectiveness of your
Lean Manufacturing Training Program
Lean Manufacturing Simulation Game 

Concerns

This approach is troubling to many people. Letting the cycle counter choose the sample parts—that is, giving the cycle counter discretion to determine what is difficult to count—leads some to fear that the overall sample will be skewed in a way that misrepresents the overall population of parts. There is also the question of when do the difficult-to-count parts get counted.

First, the issue of counting only easy to count parts needs addressing. There is nothing in statistics that differenti­ates the value of a specific sample member by its quantity. That is, a part that has quantity of 7,637 pieces is not any more or less statistically valuable than a part that has a quantity of only 1 piece. It certainly is not 7,637 times more statistically valuable. A sample member is a sample member and has equal weight within that sample as do all other members. However, a quantity of 1 is much easier to count than 7,637. If counting efficiency is of concern, counting the parts that are easy to count is a valid consid­eration. The difficult to count parts will get counted when their quantity is reduced through usage and thus become easy to count.

Second, a cycle counter's job is to count and reconcile records. If a cycle counter is not counting, it makes no difference what method he/she uses for selecting the sample. As inventory data become more accessible via improved data process communications, a cycle counter that wants to cheat on his or her counting need not take great efforts to accomplish that objective. The notion that a cycle counter should count blind (that is, without knowledge of what should physically be on location) is childish, and implies they can't be trusted.

To a certain degree, a cycle counter is really an inspector. The author is not aware of a single company that asks an inspector to inspect a part without a print showing the characteristics to be checked. No company expects its inspectors to measure the outside diameter of a shaft and then input that data into a computer for verification. On the contrary, the inspector is given the information ahead of time and is trusted to make the comparison. This practice should be just as valid when checking part number identification, location and quantity on hand.

As for the argument that the counter will not count any parts, this is easily solved by monitoring the number of parts counted each day. If a supervisor is in doubt, a simple recount can satisfy this concern.

It is the author's position that problems will always arise with certain individuals; however, designing a cycle count­ing sample selection procedure that implies distrust for the cycle counters and imposes major inefficiencies into the job is not the proper solution.

If the concern persists, then the usual resolution is to use process control cycle counting to find and fix the causes of errors, and then use either random sample cycle counting or ABC cycle counting to measure inventory record accu­racy. We recommend an approach that uses process control cycle counting to efficiently find and fix the cause of errors, augmented with either of these two other methods to measure accuracy.

To be Continued