Is There a Need for a "Min-Max" Option?

Charlie Bodenstab

Occasionally distributors, and particularly retailers, suggest that despite the sophistication of the forecasting and reorder logic of an inventory control system, there should be an override, whereby the user can stipulate that certain items be pegged with a min-max method of reordering. Their rational is that certain items have such a sporadic demand pattern that they are best served by allowing a manual pegging of minimums (below which the item will be reordered) and a maximum (the amount to which the reorder will bring the inventory balance).

In my opinion, this desire is motivated by a recognition that conventional inventory management systems handle these type of items very poorly, leading to either serious overstock conditions (due to a single aberrant surge in one month), or out of stock situations (due to the failure of the logic to place sufficient safety stock to accommodate the sporadic of demand).

The reorder logic of MARS-IW, however, is designed to deal with the entire range of product characteristics--from the low demand, highly sporadic items across the range to the relatively stable, high demand items. Accordingly, the need for a special ability to set some items up on a min-max basis is eliminated.

Let’s test this assertion, however, by taking a series of items with various degrees of erraticness and seeing how MARS deals with them. We will ignore seasonality for the sake of simplicity, but actually MARS would comprehend seasonality, even for slow moving erratic items whereas a simple manually set min-max would not. Additionally I have assumed a lead-time of two weeks and an order frequency of two weeks and the on-hand balance is zero.

Let’s discuss each series of data separately.

Series 1 – This is a classic slow moving item. MARS is basically trying to keep no more than one item on hand at any one time, which seems reasonable.

Series 2 – This is a better moving item, but again very sporadic. MARS is trying to keep 2 on-hand, which looks good for the entire year except for the 4 that moved in March and again in August. Since the order frequency is two weeks, even that demand could have been met if the sales fell properly. Nevertheless, we are accepting a 95% fill and it is actually not prudent to meet every peek demand. (If the reorder frequency had been four weeks, MARS would have kept 4 on-hand.)

Series 3 – This item is still stronger with higher movement in the months that demand occurs. MARS is attempting to keep 6 on-hand, which seems eminently sensible and covers the whole year except for the demand of 8 in August. Even that month may have been satisfied if the demand fell conveniently.

Series 4 - This is a very slow moving item, but one which had a big surge of 15 in December. Note that MARS did not "go crazy" when the 15 hit, since it was treated as an aberration and filtered down to a more nominal level. MARS did, however, order in 2 after the surge, whereas it would normally kept only one in inventory. The slight increase is in recognition that something did happen, and a slight increase is justified. If the pattern repeated itself, MARS would then start to respond very quickly to the new level of demand. Additionally, this item would have been highlighted as an exception for manual intervention if necessary.

Series 5 – This item has virtually no demand, but we can assume that you kept it as an "authorized stock item" for marketing reasons. MARS would have kept one on-hand, which again seems reasonable.

If we can accept that MARS is handling these sporadic items effectively, then there are very compelling reasons to utilize MARS to trigger the reorder quantities.

• MARS’s approach is automatic and therefore eliminates the special handling needed in setting up min-max levels. The key is to eliminate special manual intervention by users in areas where the system can do a better job, thereby freeing up the human to perform other tasks that are uniquely suited to humans.
• If the demand for the item changes, MARS will respond accordingly, whereas the manual approach needs to be overtly dealt with. The reality is that it will not be dealt with, and will only come to someone’s attention after a crisis develops.
• All the other features of MARS will be retained, such as EOQ, lot sizing, shelf minimums, etc.
• Seasonality characteristics of the product, despite its slow moving nature will be comprehended.
• Lead times and order frequency considerations will be automatically recognized, particularly if they are changed during the year.

This entire issue has an analogy in the modern automobile. For the longest time we had to manually adjust the spark advance and also choke the engine when it was cold. Eventually, the auto-makers developed technology to automatically perform all these functions–not only automatically but more effectively. The early users relinquished their prior options reluctantly, but with hindsight it is clear we are well ahead having the old options eliminated.