Randall Schaefer, CPIM | September/October 2011 | 21 | 5
A regimen for avoiding the “starve and stuff” phenomenon
Perhaps the oldest and simplest method for managing the replenishment of independent demand items is the order point. According to the APICS Dictionary, this time-tested method is “a set inventory level where, if the total stock on hand plus on order falls to or below that points, action is taken to replenish the stock.” It is most effective when the items being managed are unlikely to be obsoleted and demand during the replenishment lead times is fairly consistent.
The order point is the trigger number—meaning, when on-hand inventory is depleted to that level, a replenishment order is generated. The quantity used for the trigger represents the likely demand for the part during its replenishment lead time plus any safety stock. The quantity of the replenishment order is the order quantity, which can be determined by any reasonable lot-sizing method.
The problem is that the order point trigger quantity typically is determined from past history—and history may not suffice for determining likely demand during the replenishment lead time. In such instances, order point-generated replenishment orders can result in too much inventory when demand is contracting or too many setups when demand is growing.
There is another problem with the order point technique. Just as in material requirements planning (MRP), order points do not release work into the shop at a steady pace. MRP provides visibility as to which parts are likely to be required next so planners can level the workload by releasing some orders early. But order points offer no such visibility. Factories using order points often go several days without putting replenishment orders into the shop because no item reached its order point trigger. Then, many orders may be released the same day and compete for resources. In such instances, the order point system will starve or stuff its work centers.
However, it is possible to overcome this tendency using an internally programmed document, which I call the target inventory report. Your enterprise resources planning system knows which end items are managed with an order point, as well as each item’s specific order point. It also knows current on-hand balances. With simple programming, a target inventory report of just five columns will show
- the part numbers managed with an order point
- each part’s work center(s)
- each part’s order point
- each part’s current on-hand balance
- each part’s on-hand balance
expressed as a percentage of its order point.
The primary sort is by work center; the secondary sort is by percentage, with the smallest percentage listed first. For example, within a given work center, an item with an order point of 100 and an on-hand balance of 10 would show an on-hand balance equal to just 10 percent of
its order point. It would be listed above an item with an order point of 50 with an onhand
balance of 150. That item’s on-hand balance would equal 300 percent of its order point, probably because a replenishment order was recently completed.
If work centers begin to starve,orders for items listed near the topof the target inventory report can bereleased into the shop with the leastimpact on overall inventory becausethose parts are the ones most likelyto be generated by the order pointsystem within a few days anyway.
Generating work just to keep workcenters busy is contrary to lean philosophybut not contrary to good managementof constrained work centers. Atarget inventory report is an excellenttool for managing input to constrainedwork centers—which must never beallowed to run out of work—and for allwork centers in factories that have notyet embraced lean.
Randall Schaefer, CPIM, is an industrialphilosopher and independent consultant. Hemay be contacted at firstname.lastname@example.org.
Do you have an anecdote thatteaches, enlightens, or amuses?Consider sharing it with the readersof APICS magazine. Stories shouldbe approximately 700 words.Email submissions to “LessonsLearned” editor Randall Schaefer email@example.com.