During dinner at a recent APICS chapter professional development meeting, I was asked, "What has been your most perplexing inventory problem?" How could I single out just one? There are those issues associated with planning hazardous co-products and the never-ending intricacies of scrap reporting. There's synchronizing inventory to a run-out date dictated by an engineering change order. Or what about planning parts to be supplied to our supplier? I was thinking about all the possible answers when the correct one popped into my head: determining how to manage parts without part numbers.
My tablemates stared at me wondering how such a thing was possible. There was a time when I, too, thought it could not be done without an unreasonable amount of effort and risk. Twenty-plus years ago, I consulted for a company that used lumber as its raw material. Purchasing employees bought lumber in a small variety of standard dimensions. Each standard size of lumber was assigned a part number, as were the parts fabricated from them. But whenever parts were cut from the standard pieces, there was offal left over.
It was this offal that was perplexing because the factory saved most of it to use when fabricating smaller parts. That was the practice in the factory, but the planning system did not recognize nonstandard pieces--all parts were planned for standard lumber. Using offal in the factory and not incorporating it into the planning for raw lumber meant that standard lumber usually was overstocked. The creation of offal was constant, so the volume would grow until undisciplined storage became hazardous. Eventually, all the offal would be scrapped for safety reasons, and the cycle would repeat.
The ideal strategy was to use offal where possible and reduce the purchase of standard lumber accordingly. The controller suggested assigning part numbers to each size of offal and changing certain bills of material to specify offal as the required raw material. This was rejected when I explained the amount of additional inventory transactions and discipline that would be required to enter offal into stock--not to mention the issues surrounding substituting standard lumber when appropriate offal was unavailable.
Even more discouraging was the fact that--when using offal as raw material--new, smaller offal often was generated that also would require a part number. In addition, workers were not equally skilled, so offal from the same fabricated part were not always dimensionally consistent. From every perspective, assigning part numbers to the offal seemed the wrong solution.
I determined that four steps were required to solve the problem:
- First, we reduced the volume of offal. To accomplish this, we determined that just 40 of the more than 100 parts were small enough to be made from offal sizes typically available. We scrapped offal with smaller dimensions and reduced the storage space for offal by 50 percent.
- We removed the standard lumber from material requirements planning control and began managing it with a visual replenishment method. This synchronized the replenishment of standard lumber with the usage of offal.
- When any of the 40 small parts needed to be fabricated, a note automatically printed on the shop order reminding the supervisor to use offal if it was available. If it wasn't, he substituted standard lumber.
- When fabricating parts from standard lumber, a note automatically printed on the shop order reminding the supervisor to save offal if the dimensions were equal to or greater than the predetermined usable size.
In the end, we had no more overstocked standard lumber and no more dangerous piles of offal. And we accomplished that without creating new part numbers or adding inventory transactions. Occasionally, doing less can yield more success. Randall Schaefer, CPIM, is an industrial philosopher and independent consultant. He is editor of the "Lessons Learned" department in
APICS magazine. He may be contacted at firstname.lastname@example.org.