Manufacturing Part Numbers 

The objective of this article is to guide production and inven­tory management practitioners to a reasoned and logical under­standing of the issues surrounding part numbering systems. The benefits and problems with significant and partially significant part numbers, as well as use of alphanumeric characters versus numbers only, short versus long part numbers, and methods of assigning part numbers will be covered. A recommendation for short, numeric-only, nonsignificant part numbers assigned sequentially will be given and justified. Finally, the interaction between part numbers and commodity coding structures will be discussed with attention to today's relational database capabilities.

First, let's get some background (and a story) out of the way. The APICS Dictionary (7th edition) defines an item number as "a number that serves to uniquely identify an item." This simple definition doesn't convey the importance of the concept. Almost every business today runs on information, and the part number is very often the piece of data that ties all the other pieces together and makes them information rather than just discrete, individual "datums."

Greene's Production and Inventory Control Handbook* defines three kinds of part numbering systems (extracts mine):

• Significant: "All characters... represent a dimension, type of material, color, or some other factor."

• Semisignificant: "Some of the characters ... are significant; the balance are not and do not represent a group or other factors."

• Nonsignificant: "None of the numbers assigned has any sig­nificance."

Significant part numbers are also frequently referred to as intel­ligent or descriptive part numbers. The idea is relatively simple and seems logical and harmless enough: Instead of relying on the description field in our item master, why not encode a partial or full description of the item's characteristics into the part number itself? Then, anyone who knows the code will know all the important parameters of the part, whether or not he or she has access to the MRP system or a written document.

This argument is especially attractive to people whose everyday work is somewhat detail-oriented, such as engineers, computer programmers, finance professionals, anyone who has to use MS-DOS, etc. I found the idea eminently sensible when, as a young engineer, I was confronted with my first "designer" part number, consisting of twelve digits that indicated everything from the business unit using the part to the part type to the size. The other engineers at my facility almost took it as a matter of faith that the part number should be intelligent, and thus we added more significance to the system, bringing the number of digits used up to the maximum of fifteen allowed by our system (which was viewed as a real limitation), specifying the value of individual capacitors and resistors, the technology of individual parts, and resulting in a series of spectacular disagreements between per­fectly well-meaning people who felt their system was the best.

Our component engineering manager designed a computer pro­gram to automatically assign the correct number for a part, given all its critical parameters. All told, I am sure that we spent a man-year on the system itself, using the talents of our best and brightest to ensure that, among other things, we never had enough letters in a row for the system to spell out any four-letter words.

Of course, once you have significant part numbers, you must maintain them. Thus, we found ourselves changing part numbers, or creating new bills of materials, when we shifted production of an item from one plant to another, or when the business group responsible for it changed, or when we found that the initial description of a part was not quite correct and therefore it required a different part number, even though the part itself hadn't changed. And so we found out the first great truth about intelligent part numbering systems: they require a great deal of effort to create and to maintain.

Still, the benefits we perceived from the system seemed to far outweigh the effort required to maintain it. We thought our system was robust in an engineering sense because it made it easy to detect errors. We were unaware of any of the problems associated with intelligent part numbers other than the effort required to maintain them. Only our poor materials personnel and the general manager (who had a Ph.D. in physics rather than an engineering degree) had a different perspective, but somehow the engineers' ideas always seemed more cogent. That is why it's important for us to understand the real issues and arguments for and against intelligent part numbers.

We'll see that there are some valid reasons for using an intelligent part numbering system. Benefits can and do result from the use of significant part numbers. However, many of the same benefits can be achieved by other means, making significance unneces­sary. More important, significant part numbers bring some sig­nificant disadvantages with them, including higher costs and cycle times, lower quality, increased errors, and unnecessary work.

To be Continued

Lean Six Sigma Consulting   World Class Manufacturing   
Balanced Scorecards  Strategic Tactical Planning  
Supply Chain Inventory Management  Principles of Total Quality Management
  Manufacturing Process Improvement