Tuesday, January 28, 2014

Multi-Echelon Inventory Optimization

For a large enterprise such as Nike and Oracle, managing inventory can be a challenging task with thousands of products located in thousands of locations all over the world. The challenge magnifies when locations are placed in different tiers or echelons of the enterprise’s distribution channel. According to Forrester Research, the key differentiator these days between a highly successful company (e.g. Wal-Mart) and a company that has sub-optimal performance (e.g. Kmart) is an ability to increase the inventory turnover.

Broadly, there are two types of inventory systems: - the single-echelon (or, single-tier) inventory system and the multi-echelon (or, multi-tier) inventory system. We will look at them briefly here.

Single-Echelon Inventory System:
A single-echelon inventory system is one wherein a single Distribution Center (DC) acts as a central repository between the supplier of the inventory and the customer-facing outlets.
In a single-echelon network, an individual material-location combination is not affected by any other material or location. If a business was selling products from a single location, then it would be categorized as a single-echelon system. The DC is under the control of a single enterprise.

Multi-Echelon Inventory System:

A multi-echelon inventory system is one that relies heavily on layers of suppliers distributed across multiple distribution centers and that is based on outsourced manufacturing. In such a system, new inventory shipments are first stored at a central or regional distribution center (RDC). These central facilities are the internal suppliers to the customer-facing outlets, also called forward distribution centers (DCs). For example, Nike’s distribution network consists of 7 RDCs and more than 300,000 DCs; and these DCs serve end customers. Here, the DC and RDC both are under the control of a single enterprise – Nike, Inc.

Multi-Echelon Inventory Optimization:
Multi-echelon inventory optimization (MEIO) right-sizes safety stock buffers across the entire supply chain, taking into account the complex interdependencies between stages, as well as variables that cause chronic excess inventory, such as long lead times, demand uncertainty, and supply volatility.

However, there are some significant issues in optimizing a multi-echelon inventory system:- 

  • Demand variation measure for the RDC.
  • Demand measure for the RDC, and how to forecast this demand.
  • Defining optimal service level goals between the RDC and its “customers” - the DCs.
  • Allocation of inventory down to the DCs when faced with a limited supply situation at the RDC.

Managing Inventory in Multi-Echelon Networks:
The objective of multi-echelon inventory management is to deliver the desired end customer service levels at minimum network inventory, with the inventory divided among the various echelons. With the primary focus on inventory, transportation and warehouse operations expenses also are taken care of, because their cost factors are part of the overall optimization.

The inventory drivers, denoted in red in Figure 1, are
  • Replenishment review frequency
  • Order supply strategy
  • Service level goal
With a multi-echelon approach, the decisions regarding the inventory drivers are made at the enterprise level in a single optimization exercise rather than in a sequence of sub-exercises for each echelon.

A multi-echelon approach optimizes the networks inventory on various counts:-
  • Avoid multiple independent forecast updates per echelon: The forecasts in all echelons are dependent on the primary customer demand signal at the DCs. A multi-echelon approach, however, is independent of demands from the immediate downstream customer.
  • Account for all lead times and its variations: In each echelon, the replenishment decisions account for lead times and its variations of all upstream suppliers, not just the immediate suppliers.
  • Monitor and manage the bullwhip effect: The enterprise measures the demand distortion and determines the respective root cause in order to take corrective measures.
  • Enable visibility up and down the demand chain: Each echelon takes advantage of visibility into the other echelon’s inventory positions—what is on hand, on order, committed and back ordered.
  • Synchronizing order strategies: Synchronizing the ordering cycles at the DCs with RDC operations reduces lead times and lead time variation between the RDC and the DCs. Multi-echelon models can evaluate the impact on both echelons of different synchronization strategies.
  • Offering differentiated service levels, etc.: The RDC can provide different service levels (for the same product) to different DCs. A multi-echelon approach makes this possible, because the enterprise controls how and when a product enters and leaves the RDC.

Bullwhip Effect in Multi-Echelon Networks:
The bullwhip effect is an observed phenomenon in forecast-driven distribution channels. It refers to a trend of larger and larger swings in inventory in response to changes in customer demand, as one looks at firm’s further back in the supply chain for a product.

In multi-echelon networks, the enterprise must consider and manage the bullwhip effect. The bullwhip effect is caused by independent rational decisions in demand signal processing, order batching, reactions to price variations and shortage gaming. Order batching in a lower echelon leads to excess demand fluctuations between echelons. Also, the lack of visibility up and down the demand chain can cause inventory stocks to pile up. A multi-echelon network can tackle the bullwhip effect by offering proper measurement, by identifying its root cause and by reducing or eliminating its impact on demand chain performance.

Two Paths to Multi-Echelon Performance in an Organization

In the end, some key questions to think about are:-
  • Are we contemplating the optimal approach?
  • Is this the best we can do?
  • Are we ready for a corporate wide program? If not-why not?

1 comment:

  1. The bullwhip effect can only be eliminated by a Automatic Replenish System. The pull is only based on End Customers


Note: Only a member of this blog may post a comment.