Inventory control is not a science. More nearly, it is a set of methods for figuring out how much stock to order and when, and how to receive it. Sometimes it is an actual set of recorded regulations that stock controllers must obey. Oftentimes it is a set of guidelines in the mind of a stock controller or purchasing agent.

Good inventory control holds inventory levels low. This means that inventory control involves the finesse of ordering as little stock as possible - but not too little. One basic principle of inventory flow is that perturbations near the start of the supply chain have a greater effect than perturbations near the end. Imagine holding your finger in the stream of water coming out of a faucet. If you place your finger near the faucet, a lot of water will splash around everywhere. But if you place your finger near the bottom of the stream, the stream will be almost unaffected. In the same way, there is nothing in inventory control that has a greater effect than purchasing. A small purchasing mistake can be more disastrous than a big manufacturing mistake. Most inventory difficulties, in fact, come about not due to large and obvious mistakes, but because a small inefficiency early on in the inventory chain remains unfixed for a long time. Companies lose much more profit by ordering parts that could be got cheaper, for example, than they lose by ordering the wrong parts.

Company management can leave purchasing decisions entirely up to experience of its purchasing agents, or can set out firm rules for purchasing, or can combine the two approaches. Purchasing agents may have a hard time seeing the big picture and may depend overly on intuition and tradition, but guidelines can leave purchasing inflexible, rigid. And each mistake made at purchasing ripples down the entire inventory chain. So how much to purchase and how? Is there a method? The answer to this question is one vital subject matter of inventory control theory, a discipline that includes elements of mathematics, engineering, and management in equal doses.

The movement of stock through the locations of a large company is complex, but it is not nearly as complex as, for example, the life that moves through a single water droplet taken out of a pond, and is not impossible to analyze. There is a best solution for each situation. To find it, a company needs to keep excellent inventory records. Given some stability in company activities, a set of optimal guidelines can be worked out. Complicated mathematical calculations are not necessary, and there is no need to fear getting lost in abstruse considerations. Inventory control theory does not require a Ph.D. to grasp. Actually, the complexity of an inventory control algorithm (sequence of steps to take) says nothing whatsoever about its effectiveness.

The simplest kind of inventory control system of rules is based on MAX/MIN, or Maximum and Minimum. This just means that every inventory item (kind of unit in stock) has a MAX and a MIN value. Every time stock drops below MIN, Purchasing orders enough of the item to get stock levels up to MAX. Usually by stock is meant the quantity of items on hand plus the quantity on order, minus the quantity of items on customer backorder (item ordered by a client or customer, but not currently in stock). Say the MAX for widgets is 100 and the MIN is 50. There are 20 widgets in stock, 40 on order, and 5 on backorder. 20 + 40 - 5 = 55 widgets considered to be in stock. Now say 10 customer backorders come in. Recalculating, we get 45 in stock. Since the level of widgets has dropped below the MIN of 50, more widgets must be ordered. In this case, MAX=100 minus STOcK=45 equals 55 widgets that must be purchased. That is a simple (and fairly tight) inventory control system.

We said before that company management can leave purchasing decisions entirely up to experience of its purchasing agents, or can set out firm rules for purchasing, or can combine the two approaches. Most experienced inventory controllers believe that for the best results, a system of rules should be combined with human ingenuity. A race car going 500 on a track is a combination of a precisely engineered machine and an experienced driver. Neither can win a race without the other. But how to combine algorithms with humans?

One excellent approach is to use algorithmic (automatic, planned, systematic...) ordering, and to overrule the automatic ordering when there is a reason to believe that the ordered stock, or whatever is to be made from the stock, will not sell. This means not overruling the algorithmic ordering on a hunch, but only for a good reason that can be well defined and clearly stated. In this way, bad hunches and incorrect assumptions will do the least harm possible to purchasing, and the human ability to recognize a pattern, an ability which computer programs are toddlers at, will come into play whenever logically justifiable.

In a future article we will take a look at some algorithms.