Coping with the Bullwhip Effect Reducing uncertainty reduce

Coping with the Bullwhip Effect • Reducing uncertainty: reduce uncertainty by centralizing demand information (providing each stage of the chain with information on the customer demand) • Reducing variability: reduce variability in the customer demand process (consider the extreme case of deterministic demand---echelon inventories both experience the same demand rate D/unit time)

• Lead time reduction: Lead-time=order lead time (time to manufacture and ship) + information lead time (time it takes to process the order) • Strategic Partnerships – VMI (manufacturer manages the inventory at the retailer)

Barilla Case Continued: • Initially distributors are very skeptical for the JITD (just -in-time distribution) system • Pilot program with Barilla-owned warehouses (September 1990) • Results for Florence warehouse: – Inventory: decreased from 10. 1 days to 3. 6 days of stock – Service levels: increased from 98. 8% to 99. 8%

• An implementation (sort of) begins at Cortese DC – Barilla logistics and sales management, a consultant (with former ties with Cortese) meets with Cortese managers – Set-up to simulate the system for six months (using Cortese data on sales and inventories in a virtual environment) – At Cortese’s Marchese DC after six month test period • Inventory: 50% reduction • Stockout: from typical 2 -5 % (and occasional 10 %) drops to 0. 25 % – Several other Cortese DC’s with similar results

• By the end of 1993 – – – Barilla’s JITD customers are connected electronically Sent information to Barilla through EDI Inventory levels previous day’s sale data Stock-outs for the previous day advanced order for promotions • JITS Sales 75 billion lire in 1991, 200 billion lire in 1993

• Fill trucks in priority order (1) item expected to stock out (2) items farthest below order-up-to levels (3) advance shipment of promotional items (4) items least above target stock levels (1)Order-up-to levels are determined using average and standard deviation of demand from exponential smoothing forecasts.

Types of Shared Information • Inventory Levels • Sales: – – Instead of orders actual sales data! Avoids double, triple forecasting One way to avoid the bullwhip effect Sales data sharing is only meaningful when inventory levels are also shared – P&G and Wal-Mart (POS) – Batch orders disguise the real demand – Promotions of the supplier can distort the ordering behaviour

• Order Tracking – Independent units of the supply chain share order information. – Customer-oriented service. Customer can learn its order status from a central point. – Customer satisfaction as well as faster turnover in payments.

• Sales Forecasts: Forecasts – VMI requires sharing of sales forecasts and avoids the bullwhip effect – In some cases, the supplier knows the end customer better than the retailer – Wal-Mart developed a Continuous Forecasting and Replenishment (CFAR) system with Pharmaceutical companies – HP, Sun, and TI shares their sales forecasts with their suppliers via Flexible Supply Contracts.

• Production/Distribution Schedules: – The producing partner can utilize the production plan of its suppliers to enhance its own production plan – US Automotive industry manages the production and distribution of its own steel orders. – To better plan the distribution access to production plans is crucial – Motorola plans its own chip production using the production plans of computer producers.

• Other types of information: – It may not be possible or even useful to share the sensitive cost information. – Capacity information could be shared to come up with a “virtual supply chain”. – Performance Measures : Daimler-Chrysler, broadcasts all quality and on-time delivery data of its customers. – Performance measures can be shared to understand how efficient the chain is operating.

Difficulties • One of the major hurdles is “trust”. • To manage trust along the supply chain is a major element of the whole process. • It is always possible to use the information in unintended and unlawful ways. • In some cases information sharing may face antimonopoly laws.

Models of Information Sharing A Inf. System Third Party B Data Transfer A B 3 PL A B Inf. System Data Center

The Beer Game:

• Each component in the supply chain has unlimited storage capacity, and there is a fixed supply lead time and order delay time between each component. • Each week, each component in the supply chain tries to meet the demand of the downstream component. • Any orders which can not be met are recorded as backorders, and met as soon as possible. No orders will be ignored, and all orders must eventually be met. • At each period, each component in the supply chain is charged a $1. 00 shortage cost per backordered item.

• At each period, each component owns the inventory at that facility. In addition, the wholesaler owns inventory in transit to the retailer, and the distributor owns inventory in transit to the wholesaler, and the factory owns both items being manufactured and items in transit to the distributor. • Each location. is charged $. 50 inventory holding cost per inventory item that it owns. Also, each supply chain member orders some amount from its upstream supplier. It takes one week for this order to arrive at the supplier. Once the order arrives, the supplier attempts to fill it with available inventory, and there is an additional two week transportation delay before the material being shipped by the supplier arrives at the customer who placed the order.

• Order of Events: Each week, for the retailer, the wholesaler, the distributor, and the factory, the following events occur, in order – 1. Contents of Delay 2 moves to Delay 1, Contents of Delay 1 moves to inventory. Delay 2 becomes 0. – 2. Orders from upstream are fulfilled as much as possible. – 3. Backorder and inventory holding costs are calculated and added to the accumulating costs. – New orders are placed.