Yesterday for class, we played the Beer Distribution Game, which is a game developed by the Systems Dynamic Group at MIT back in the early 1960’s. This game simulates what can happen in a traditional supply chain and exposes some interesting dynamics that happen in real-world supply chains.

The players of this game take on the role of Retailer, Wholesaler, Distributor, and Factory. The retailer sells barrels of beer to a consumer and orders barrels of beer from the wholesaler, the wholesaler sells barrels of beer to the retailer and orders barrels of beer from the distributor and the distributor sells barrels of beer to the wholesaler and orders beer from the factory (brewery). The factory brews the beer. The beer supply chain is shown below:

Each player is directly linked, and beer cannot skip the adjacent position. For example, the Wholesaler orders beer from the Distributor, and ships beer to the Retailer. The goal of the game is to minimize team total costs — each barrel of beer has a cost, which is calculated at the end of the game. Information Flow proceeds from Retailer to Factory; Material Flow proceeds from Factory to Retailer.

Definitions

1. Orders received: This is the demand vor the current period at this position. For the Retailer the demand is determined by the Computer itself. For all positions, this demand reflects an order placed by the downstream position in the supply chain during the previous period.
2. Backlog: This is the demand that has not been met to date at this position. When a position does not meet demand by shipping barrels of beer, the backlog amount is increased. At no time should there be inventory items and a backlog simultaneously.
3. Current Costs: This is the total cumulative costs for the position (barrels * $).
4. Order: This is the field to insert the order for the actual round for each position.

The order quantities of the retailer was the same after week 4. From week 1 - 4, the order quantities was 4 barrels. From week 4 until the end of the game, the order quantity was 8 barrels. But, the players didn’t know this fact. This is what the order quantity chart looked like:

Despite the contant order quantities for the entire game, the results are very surprising — this is what demonstrates the Bullwhip Effect:

Observations

1. Variation os Stocks and Orders increases up the supply chain from customer to supplier.
2. The longer the lead time of information and material, the more exaggerated the bullwhip effect is.
3. The system is to blame: (a) If customer demand sinks or levels, then supplier needs to empty pipeline in order to minimize costs; (b) if customer demand increases, then supplier pipeline needs to be filled in order to avoid a backlog. The phenomena produces a feast/famine scenario in the supply chain.
4. Procurement or purchasing in batches adds variability.
5. Changing forecasts lead to a change in safety stocks. This creates variability in the system.
6. Promotions impact variability of demand.
7. In times of a shortage, tier x suppliers tend to order more than actual demand in order to avoid a backlog.
8. Related to (7), customers tend to order more in times of shortage; when shortage is over, cancellations occur adding stock variability to the system.

How to Manage the Bullwhip Effect

I’ve experienced the Bullwhip Effect, and I can tell you that I have not learned to effectively manage it or avoid it. But, below are some true and tried principles for how to manage the Bullwhip Effect:

1. Reduce lead time of information (orders, demand and capacity forecasts, point-of-sale data for the entire supply chain).
2. Reduce lead time of material.
3. Reduce variability with effective use of the Heijunka and one-piece flow.
4. Cooperation and good relationships with your supply chain partners.

Hansei, Reflection

Yesterday was a fascinating study into human behavior. As I walked around the room, I heard phrases such as "let’s assume orders will always be between 5 and 15 barrells" or "let’s make more now so that we won’t run out so fast" or "let’s order double each time so that we will always have inventory." Without knowing it, the players yesterday over and under forecasted; they over-compensated on inventory; they tried to "rush" production and didn’t follow pace — system and human behavior that is so classic in real Operations and Supply Chains.

For articles on queueing theory, time-traps, operations, lean and six sigma, please visit the links below:

• Managing Constraints Under Peak Volumes

• Series on Queueing Theory

• Order Pipeline of Events

• Variation for Dummies

• On Game Theory

• Queueing Theory 1

• Queueing Theory 2

• Queueing Theory 3

• Queueing Theory 4

• On Time-Traps and Waste

• Applied Regression Analysis

• Team Sizing

• 5S

• Click-to-Ship Processes

• Root Cause Analysis

• Lean at Krispy Kreme

• Theory of Constraints and Camping

• Lean for Software Development

• Don’t Waste the Customer’s Time

• Featuritis and the Focus on the Customer