Push, pull and production

Imagine that you are a blacksmith in the olden days, capable of turning out five swords per day. Now imagine two cases. In the first case, you turn out five swords every day and hope that all five are bought, today or eventually. Sometimes not many people come in to buy swords - there is a shortage of demand, in econospeak. Then you start, naturally, to turn out only three swords per day, or two, or one - but a nagging feeling sometimes tells you that no, you should still turn out five swords per day even though no one is buying them, since a day may come when demand shoots up to ten swords per day, and if that happens you will be wishing that you had continued to stockpile swords in inventory. Lacking any good data on whether or not such a day will come, you will probably do what your father, who taught you the art, did in similar situations, and what his father did.

In the second case, you do something radically different: you train yourself to be able to make a sword as quickly as possible, then only make a sword when somebody wants a sword. This way, you never have any stockpile of swords, and if demand rises past your ability to fulfill it you will lose out. Also, if your supply of the raw materials that you need to make swords is terminated, then even if demand is strong you will have no stockpile to get you through the tough times until this supply is restored. At the same time, though you will be taking risks, you will never waste any effort on making swords that no one will buy.

This is a simple example of the difference between the Push and Pull principles, or philosophies, of inventory management. Push places the center of gravity for inventory management decisions in the manufacturer. Pull places the center of gravity for inventory management decisions in the purchaser. Push forecasts what demand will be like in the future and tunes manufacturing to that. Pull prepares to fulfill demand as quickly and efficiently as it can, then actually waits for the future's demand to present itself, and replies to this demand when it comes.

The Push and Pull principles have always been around. One can see them even in the animal world. Some animals prepare food stocks for the winter, whereas other animals simply eat whenever they are hungry. These are two different ways of responding to the variability of the supply chain that goes from outside through their mouths and into their bodies. In nature, though, the two principles are often mixed. A cat is a sleek, lithe killer whose body seems to be designed for hunting. But even a cat will sometimes bury food.

Likewise, in the human world of logistics, one can find few companies that operate on purely Push or purely Pull principles. Even the central economic planners of the Soviet Union experimented to some extent with free markets as a supplement to their Push-based production system. Even Toyota Motors, which pioneered Just-in-Time Manufacturing, the "Toyota Production System", the peak of Pull, so to speak, holds safety inventory as a guard against the unexpected.

There are ways to combine Push and Pull. The place where Push gives way to Pull is known as the Push-Pull boundary. For example, a group of food manufacturers raise chickens in numbers that are based on projected demand, but only slaughter chickens based on actual demand. The Push-Pull boundary in this case is just before the slaughterhouse. The statistical methods used to determine an appropriate Push-Pull boundary are too complex to go into here, but it may be said that they essentially involve determining the point along the production chain where the savings that accrue from Pull stop being worth the risk associated with Pull. These calculations involve statistical facts and certainties and follow the model of all business planning.

Written by Anton Dolinsky for Almyta Systems. When author's and Almyta Systems names are mentioned, the reproduction is freely allowed.