APICS is the premier professional association for supply chain management.
Insights, analysis and ideas to advance your supply chain.

Thinking Supply Chain

Insights, analysis and ideas to advance your supply chain.
Join the conversation.

What is “push” and “pull” distribution?” – Part 7

Thursday January 21, 2016

push pull pt7
David F. Ross, PhD, CFPIM, CSCP

In this edition of the on-going blog on “push” and “pull,” we will start a detailed examination of the push replenishment system. The previous blog defined the push system in distribution as “a system for replenishing field warehouse inventories where replenishment decision making is centralized, usually at the manufacturing site or central supply facility (APICS Dictionary).” As illustrated in the below figure, in the push system central planners begin by assembling the demand histories from satellite entities in the distribution channel. This data is then manipulated by forecasting techniques to arrive at an estimate of aggregate channel demand. Often without regard for current channel stocked quantities, central inventory planning then uses predetermined algorithms to allocate (push) a share of the inventory at the parent facility to each child facility on the next echelon level. If inventory quantities are insufficient to meet planned allocations, planners may choose to divide the available inventory among the child facilities using predetermined rules, such as fair shares allocation or market-based judgment. Finally, shortages at the central facility trigger resupply orders placed with upstream manufacturers or product distributors.


The cornerstone of the push system is determining channel demand. How channel planners assemble demand, however, is a complex affair and varies by industry, and sometimes by product. Take for example a poultry farmer that sells to various downstream distributors and processing plants. Since inventory is “grown”, planning poses several challenges. Yields may differ and then there are always external threats, such as avian flu, that make channel replenishment difficult to predict. Normally, producers will aggregate the expected demand and forecast how many chickens should be grown. Farmers will then determine shipments (done on a short or even daily interval) to channel distributors and processing plants based on a flexible algorithm.

In a very different example, I consulted at one of the world’s largest and exclusive women’s apparel companies. Their product was governed by a seasonal cycle. There was a “black box” forecasting system (I was prohibited from seeing) that planned what styles, sizes, and quantities of garments were to be produced (in China). Percentages of garment styles were then shipped to global distribution centers (DCs). At the DCs, planners used a sophisticated algorithm to determine how the garments were to be allocated and pushed into the distribution channel ending at the store level. Downstream buyers fought for their share of the seasonal allocation. There were no backorders: once a style or size stocked out, the garment was no longer available.

While relating to two very different business environments, the examples point to some general principles about demand planning in a push system.

  1. The products and quantities produced are usually not derived from actual channel demand but rather are estimated from past historical usage and massaged by some form of proven forecasting model.
  2. Usually the producer (or DC) will aggregate demand from child channel echelons to arrive at replenishment lot sizes that promote the best utilization and lowest cost.
  3. The system depends on a historical rhythm of sales and supply. Patterns governing how much is pushed to customers arise through time and change as historical demand or marketplace conditions change.
  4. Channel shipments are usually not driven by resupply order requests from child locations.
  5. The actual quantities produced are often variable: sellers will simply ship (within a tolerance) quantities to meet customers’ historical usage, contract, and delivery quotas.
  6. Sellers will motivate buyers to purchase overproduction. Buyers, in turn, will go to open market spot-buys (if possible) when there are shortages.
  7. Allocations of inventories to the distribution channel are determined by proven algorithms and decided upon by central planning.
  8. The supply channel assumes a steady flow of inventory shipments often without reference to on-hand quantities. Product keeps flowing-in (often to contract) regardless of how much stock is currently available.
  9. When there are shortages, central planners will decide through algorithms which child facilities get what products and portion of available quantities.
  10. In many industries there are no backorders.
  11. Channels will look to inventory tools such as risk pooling and sales volume partners to decrease the number and quantities of stocked inventories.
  12. Financial settlement is based on what was actually delivered, governed by contract or marketplace prices.

In the next blog will be visiting the push allocation process in greater detail.

All comments will be published pending approval. Read the APICS Comment Policy.


  1. RadEditor - HTML WYSIWYG Editor. MS Word-like content editing experience thanks to a rich set of formatting tools, dropdowns, dialogs, system modules and built-in spell-check.
    RadEditor's components - toolbar, content area, modes and modules
    Toolbar's wrapper 
    Content area wrapper
    RadEditor's bottom area: Design, Html and Preview modes, Statistics module and resize handle.
    It contains RadEditor's Modes/views (HTML, Design and Preview), Statistics and Resizer
    Editor Mode buttonsStatistics moduleEditor resizer
    RadEditor's Modules - special tools used to provide extra information such as Tag Inspector, Real Time HTML Viewer, Tag Properties and other.