Skip To The Main Content

What is “push” and “pull” distribution?” – Part 15

Senior Manager, Professional Development

Wednesday June 1, 2016

In the previous blog we examined closely the mechanics of the reorder point (ROP) pull system. Although reorder points do provide an effective technique for multiechelon inventory replenishment, they have several important drawbacks. To begin with, the ROP technique illustrated in the previous blog provides a very poor window into demand moving up the supply channel. A supplying facility is unaware of satellite facility demand until it actually occurs. In this sense ROP pull systems are “event” versus “planning” driven. Nothing happens until an item’s reorder point is tripped, and when it occurs, supplying facilities must respond quickly to avoid satellite channel stock out. Also note that each occasion of resupply is unconnected. Requirements appear at supplying facilities randomly.

For the most part, this weakness in the system is not serious. When it is remembered that channel replenishment lead times are short (for the most part a matter of days, sometimes hours, to transport stock from one location to another) ROP logic is the logical choice to effectively manage channel resupply. In fact, the use of buffering strategies at supplying facilities (such as safety stocks, safety lead times, and risk pooling) easily solves the problem of differences in replenishment lead times between resupply dyads.

While an effective technique for replenishment in an environment where finished goods are acquired in short lead times from outside suppliers (such as in pure distribution), ROP has a serious deficiency when the source of finished goods is an internal manufacturing plant. The problem is purely one of lead time. Simply, products can be moved through the distribution channel much faster than they can be built at the manufacturing plant. This condition is illustrated in the below echelon flow chart. Note the differences in delivery lead times between the distribution facilities and the lead time needed to produce product at the plant. While the Central DC expects replenishment in 5 days, the cumulative production lead time at the plant is 40 days!


Clearly there is a serious mismatch between the fast delivery time (literally the time it takes to prepare, ship, and receive product) between channel echelons and production lead time at the plant. When it is considered that the ROP system will generate a single distribution order (DO) at the plant at some unknown time, the plant cannot possibly do effective planning for future production. To cope with this lack of delivery lead time synchronization, the plant would have to carry significant amounts of finished goods inventories as insurance against Central DC random demand.

As we will discuss in the next blog, the problem is solved with distribution requirements planning (DRP).

Live Web Chat