Beer Game Report - The Bullwhip Effect

THE BULLWHIP EFFECT IN SUPPLY CHAINS SIMULATION, CAUSES & SOLUTIONS

GANESH RAMKUMAR PATRICK TRAIN MATTHEW VIDOTTO CHRISTOPHER WATSON Prof. MICHAEL KIRK SCM 512 MARCH 11th 2014

THE BEER GAME SIMULATION Were you surprised at the behaviour of the supply chain? The Retailer’s Perspective | Ganesh I was not surprised at the behaviour itself, but rather at the amplitude of the behaviour. I did not expect variability to „bullwhip‟ along the supply chain to such an extent. The game started off with zero demand at my (retail) echelon. Since my lead time from the wholesaler was just a week, I placed no order until there was demand. It appeared that everyone did the same, and that cause a huge lag between when goods were needed to when they were actually received. As a result, we were stocked out and racking up backorder costs while demand was high, but when demand began decreasing we began receiving all those previous orders and were now racking up holding costs. To add to the problem, I placed orders in anticipation of a trend of growing demand by factoring in safety stock. Unfortunately, everyone along the supply chain did the same and when demand started to decline, we had already placed large orders that included significant safety stock quantities.

The Wholesaler’s Perspective | Matthew I was surprised by the erratic behavior of the supply chain. When comparing demand after the game, I was also surprised by the relatively small demand fluctuation at the retailer level in comparison to the fluctuation at my (wholesaler) echelon. When demand increased, we placed orders accordingly until a point was reached where every level of the supply chain had too much stock; by this time, demand had started declining. As the game came to an end, everyone along the supply chain were no longer buying inventory and had high inventory carrying costs and were no longer ordering more inventory but rather trying to get rid of it. From the perspective of the wholesaler, it was especially surprising how much demand differed between me and the retailer.

Group 6 | SCM 512 | March 11th 2014

Page | 2

The Distributor’s Perspective | Christopher I was surprised by sudden unanticipated fluctuations throughout the game. We have had inclass about the risks of variability in demand and supply and I found it interesting how variability can occur so rapidly and on such a large scale. It was also interesting how internal supply chain processes can impact variability too. In our game, the lack of supply chain integration or data sharing lead to colossal changes in the weekly expectations. In my role as a distributor with an opening order of 4 units, my first thought was that demand would be low generally throughout the game; I did not place orders and build stock ahead of time, since I did not anticipate demand fluctuations. However, orders quickly got much higher as the bull whip effect started to influence decision making. On week 10 order quantities from the wholesaler began to rise. I was surprised by this and at that point, I could not meet the needs as I had not planned for this. Also there was a spike in wholesaler demand of 200 units in week 15, this was also unexpected and came just when I presumed the factory would be able to fulfil the requirement on back order.

The Manufacturer’s Perspective | Patrick As the manufacturer, fluctuations in demand throughout the supply chain were aggregated up towards me because each echelon also included safety stock into their order quantities to protect against the lead time of information and physical goods transaction. This in turn led to high variability in order quantities from week to week; on certain weeks I would receive no orders and on other weeks, orders would be placed for over 250 units creating large back order costs and backing up the supply chain. This behaviour did not surprise me since I knew each member of the chain would attempt to mitigate their costs as much as possible by holding minimal stock when demand is low, and chasing demand fluctuations with appropriate orders to try and minimize backorder costs. There was also the issue of latency between raw materials acquisition, beer production and order shipment; I could not satisfy large orders in a just-in-time manner and the negative impact of this caught me by surprise.


Page | 3

How did you feel about making decisions with partial information? Not having any information from upstream or downstream partners was a major challenge to all of us; it made it virtually impossible matching supply with demand and keeping costs down. Without information order quantities and safety stock were a gamble and my order was lower than demand again, creating a deeper back order and tremendous costs. Back orders swiftly increased as without knowing the status and lead times of the factory node it was hard to plan how to deal with the backlog. This combination of demand uncertainty and demand variability resulted in significant inventory fluctuations that out of sync with demand.

Without any

indications as to when demand would drop I ordered more.

Additionally, the lead times between each echelon made it quite difficult to satisfy demand in a just-in-time manner. Without the necessary information (i.e. sales forecasts shared along the supply chain), our order sizes failed to compensate for the lack of responsiveness. We were unable to anticipate and plan inventory based on future demand; instead, decisions made were reactionary and we were operating at high risk. Without visibility and information transparency across the supply chain we felt like decisions were made arbitrarily. It was a dangerous guessing game from beginning to end, and this highlighted the importance of relevant and reliable information in optimizing supply chains.

MAJOR CAUSES & PRACTICAL SOLUTIONS Demand Forecast Updating In this situation, each node maintains their forecasts independently, regardless of supply chain influencers from other nodes. As a result, true demand at the end-customer level is lost; it becomes progressively distorted as it moves upstream. Long lead times, moving forecasts, aggregated forecast errors extend this problem and increase demand variability.


Page | 4

Solution 1 | Visibility of POS Data Nodes may benefit from using the same raw sales data as a basis for calculating forecasts. Point-of-sales data from the farthest downstream customer is the most true representation of market demand. This data can be shared across the supply chain via EDI or integrated systems. If all echelons used the same historical sales data, forecasting and planning would shift from a purely autonomous process to a more harmonized process.

Solution 2 | Centralized Forecasting In some supply chain, it may be possible to centralize the forecasting function by implementing an administered Relational Collaborative Agreement (RCA). The RCA leader (usually the node directly upstream from the end-customer) would be responsible for generating forecasts from POS data, factoring in inter-node lead times and forecast error/variability, and relaying final sales forecasts upstream to each node. Ideally this would be the optimal solution; however, it requires the presence of a common goal, the establishment of trust and long-term compliance across the chain.

Order Batching In this scenario, nodes place aggregate multiple planned orders into larger batched orders to reduce transportation costs, fixed order costs and/or other transaction costs, or to take advantage of bulk discounts or other offers. Buyers usually factor in safety stock requirements as well, in anticipation of demand variability during the now longer (batched) cycle time. As demand is passed upstream towards the manufacturer, batching and safety stock increments may occur at each echelon, resulting in significant demand fluctuations further up the chain.


Page | 5

Solution 1 | Synchronization & Order Constraints A supplier can mitigate the bullwhip effects of order batching by coordinating orders with customers and managing order fulfillment based on the delivery timeframes, lead times, order sizes and other prioritization factors. This strategy would reduce concurrent orders and large spikes in demand. If possible, suppliers may set order size constraints based on their immediate capabilities; this would limit the size of batched orders and discourage forward buying.

Solution 2 | Reducing Lot Sizes Although it may be expensive to achieve, lot size reduction (e.g. via lean/six sigma methods of process optimization) could be an effective method of mitigating the negative effects of batch ordering between certain nodes. Lower lot sizes would mean lower fixed ordering costs and smaller minimum order quantities for the customer. Customers would have greater order size flexibility, and LTL shipments (leveraging 3PL capabilities) may be used if transportation lead times to customer(s) are relatively short.

Price Fluctuations Recurring increases and/or decreases in price often leads to a poor buying patterns downstream (towards the customer). Buyers begin to purchase in bulk when prices are low, and reduce orders when prices rise. Known as forward buying, this practice causes significant demand variability upstream that are amplified by long lead times and large lot sizes.

Solution 1 | Stable Pricing Strategy Sellers can mitigate the bullwhip effects of price fluctuations by adopting a long-term pricing model like EDLP (everyday low prices). By employing a strategy that applies consistent price points regardless of demand and availability, sellers can stabilize prices and in turn reduce the demand variability caused by price changes. Periodic offers and promotional discounts can be


Page | 6

used to drive demand for a short period of time, while the base pricing model provides constancy in long-term price expectations among downstream buyers.

Solution 2 | Disintermediation A supplier may opt to circumvent the distributor entirely and utilize a direct-to-store or consignment channel to sell goods directly to retailers or even end customers. These nodes are less likely to have the buying power of a distributor or wholesaler, and demand variability would be minimized to a much smaller range. In addition, lower end-to-end lead time would make the entire chain more responsive to market demand.

Solution 3 | Systems & Incentives A seller may benefit from the use of an activity-based costing system that provides insight into the financial implications of unstable prices and poor buying patterns. It would help encourage price stability within the organization and would allow for the development of customer incentives aimed at stabilizing demand.

Rationing & Shortage Gaming “Gaming” is the name given to the common customer response to manufactures rationing strategy. Rationing is done to help allocate reduced supply by reducing availability in relation to the shortage. For example if supply was reduced by 50% then orders to customers would also be reduced by the same amount. Gaming then is when customers try to work around rationing by exaggerating their orders and/or placing multiple orders in anticipation of a shortage. This is especially a problem with suppliers that offer flexible reverse logistics; buyers order „greedily‟ and simply return excess stock later. In these circumstances, distortion of true demand can add cost, waste and inefficiencies across the supply chain.


Page | 7

Solution 1 | Information Transparency Technological advances that allow for shared information and visibility across the supply chain allows customers to no longer speculate or worry about product shortage. By sharing historical sales data or sales forecasts for example, suppliers can match anticipated demand rather than having to meet requested order quantities. Buying nodes can also plan orders at optimal times and at optimal quantities based on future demand, supplier capacity and lead times.

Solution 2 | Capacity Reservation In some supply chains, nodes may be able to discourage shortage gaming by allocating set proportions of their effective capacity individually for their key customers, or collectively for their smaller customers. These allocations, based on historical sales, would ensure that customers can order enough cycle stock to meet demand. (Rong, Synder, & Shen, 2008)

Solution 3 | Return Policy Restrictions Suppliers may review their return policies and/or develop stronger relationships with their customers; strict abuse prevention clauses and stronger relationships would help lessen shortage gaming and the substantial returns that ensue.

The Optimal Solution | CPFR Strategies Most of the solutions described above have one thing in common – to some extent, they all require varying forms of collaboration, information sharing, and supply chain visibility. CPFR is just that – the exchange of data and the alignment of supply chain activities to ensure optimal flow of goods through the supply chain. A successful implementation of CPFR would mitigate all four major causes of the bullwhip effect.


Page | 8

CITATION Rong, Y., Synder, L., & Shen, Z.-J. M. (2008, August 20). Bullwhip and Reverse Bullwhip Effects under the Rationing Game. Retrieved March 9, 2014, from Social Science Research Network: http://ssrn.com/abstract=1240173

APPENDIX A: COSTS BY ECHELON Pie Chart | Distribution of Costs

2,758

2,157

Retailer Wholesaler

4,834.50

Factory

5,443.50

Echelon

Distributor

Cost (Units)

Percentage of Total

2,157

14.20%

Wholesaler

4,834.5

31.82%

Distributor

5,443.5

35.83%

2,758

18.15%

15,193

100%

Retailer

Factory TOTAL


Page | 9

APPENDIX B: BEER GAME GRAPHS Retailer – Demand, Orders & Inventory

Wholesaler – Demand, Orders & Inventory


Page | 10

Distributor – Demand, Orders & Inventory

Factory – Demand, Orders & Inventory


Page | 11

Plot of Upstream Orders

Plot of Inventory/Backorders


Page | 12

Beer Game Report - The Bullwhip Effect

Recommend Documents