Charlie's Paper 11

Dealing with the Unique Problems of "Cut Goods"

Charles J. Bodenstab

Remnants, Short Lots, and Non-Standard Lengths

There are a number of businesses that share a very common characteristic when it comes to inventory management, although on the surface, the business would appear to be totally unrelated. These are the businesses that deal in “cut goods” and “run lots”. In these situations, part of the inventory may be in nonstandard lengths, or a lot quantity that is less than normally needed for a complete job. Floor covering distributors, steel service centers, fabric distributors, wall paper distributors etc., all have this problem in common. In each of these cases the company may have a significant amount of product in inventory for a given stock number, but they cannot rely on the product being in an adequate length or lot quantity to meet customers’ demands if it is made up of non-standard lengths or short lot quantities.

The fundamental solution is to have two parameters that classify product into three categories. For purposes of discussion let's coin the following definitions:

Good Stock - This is stock that exists in standard lengths or batch quantities that have a high probability of meeting customers’ needs.

Remnants - This is product that has been cut to meet a prior order but the “remnant” is still large enough to satisfy some orders. While you will wish to keep this stock on hand it should not be counted as the good stock since it will be misleading to the product manager.

Scrap - This is product that is so small that the likelihood of meeting a customers order is very low, and unless it is disposed of, it will eventually clog the warehouse.

For MARS (or any inventory management system) to operate in this situation, the host system must classify product into the above three categories. Then, only the quantity of “good stock” available should be passed on to MARS as the on-hand inventory. The logic of this action is that while we may fill an order with remnant product, there is no assurance that we will, and therefore our objective of providing 95 or 98% fill is undermined if we count remnants as on-hand product.

There are two methods of setting the parameter that classifies product as good vs. remnant.

The first is to manually set a factor in the host system that indicates at what point the product should be reclassified as remnant. This can be a percent that is applied against the good product standard size. (80% would mean that once the product was less than 80% of a standard length it is a remnant.) In other cases the percent could even be zero, meaning that as soon as product is less than the good size, then it cannot be trusted to meet customer needs. In still other cases it may be fairly liberal (or low percent), since experience indicates that even a small percent of the standard size will take care of most orders.

When product falls below the factor it is then set aside in the system and not downloaded to the inventory system. It is still retained in the computer, however, and is checked by the order entry clerks for possible application. (When the remnant falls below the scrap percent, it should be actually scrapped.)

Another, more sophisticated approach to set the factor at which the product becomes a remnant is to have the host system automatically, and statistically determine the factor. This can be accomplished by logging the actual customer order sizes by SKU, and creating a frequency distribution of the individual order sizes. It would then be possible to statistically determine what factor will yield what percent of probable fill. For example, you could indicate that you wanted the remnant factor set so that there was an 80% chance that an incoming order would be satisfied out of the mix of product being retained as “good product.” (Actual fill rates would be considerably higher, since “good product” would exist to supplement the remnant quantities.) The frequency distribution would then automatically set the factor accordingly.

This technique could also set the scrap point. By saying that you wanted the scrap factor set so that there was only a 10% chance of filling an order out of the remnant quantity, the factor would be set accordingly and as product fell below the factor, you could be sure that its chance of filling an order was minimal.

This technique of having factors that classify product, and then keeping the product separate creates demands on the host system that many systems cannot meet. MARS can facilitate converting a conventional system to handle these needs. The technique is as follows:

For every bona fide warehouse create a “pseudo” warehouse (e.g., for warehouse 01, create a warehouse 11 that actually does not physically exist). Then arrange to shift product into the pseudo warehouse when it falls below the remnant factor. You now have the product split and the order entry people can check both classes of stock against incoming orders.