Intelligent Part Numbers – Not so clever

Fundamental to any manufacturing organization, and by extension its PLM system, is the part number. This is a unique identifier for any end component used in a manufacturing process.

Anyone who has dealt with part numbers is familiar with the so called "intelligent part number". To an extent, this is a carryover from the days before sophisticated database systems and fast computers. It involves baking multiple pieces of information into single part number.

For example, consider the fictitious case of the omnipresent widget and assume that it has a Part Number built up as follows:

Initially, the structure of the number seems perfectly logical and in keeping with our concept of what part numbers should look like. However, a more detailed analysis will diagnose some (but not all) potential problems:

1. Clearly, a transition into the 21^st Century is a problem for the first two digits of the number
2. The 1001^th part will not be able to fit
3. The enterprise expands its manufacturing base offshore and wants to introduce country codes

If the intelligent part number is the key attribute in the database, the problem rapidly compounds. This is because it is extremely difficult to change the key attribute in a database once objects have been created and included in multiple relationships.
 

So what is the alternative to intelligent part numbers? The prudent alternative is to set up non-key attributes in the database that contain the information previously baked into the part number. Once this is done, the part number for objects can be a randomly allocated sequence of digits.
 

This may sound counter intuitive but, in the long run, will save a lot of heartache.