Beer Game Report
The bullwhip effect 28/09/2011 Tiphaine Ribetto – st Perrine Trullemans – st112855 Background: “Beer Game” is a simple simulation of a Make-to-Stock Supply chain. The goal of this game is to minimize cost of capital employed in stock while avoiding out-of-stock situations. The issue here is how to forecast demand accurately. Tiphaine and I assume the roles of beer factory in the production department. As our work does not involve any decision in the order flow, our discussion will focus on our experience as manufacturer.
Results analysis: * What happened to the order quantity as we move backwards, up the supply chain from retailer to manufacturer? And why? What happened during the first rounds (weeks 1-10) is that our team places small orders in an attempt to get rid of some of the inventory. As the incoming orders size were getting lower and lower, we interpret this reduction as a signal of declining demand. Our production batches were consequently reduced and sometimes equal to 0 (weeks 5 – 6 -7).
Only $13.90 / page
Our supply chain was adjusted to a low demand scenario.
When a pic in demand occurred in week 10, we were not ready to fulfill the total order. Thanks to a ‘safety stock’, we achieved to limiting the number of backlogs. Nevertheless, this jump in demand entailed consecutively an increase in our orders. Up-streamed participants ordered extra in an attempt to fill the pipeline – forgetting that orders served two rounds before would be eventually filled. This leaded us to produce even more than the demand in the market (round 11 -17: order size oscillated from 8 to 14).
In the end, we lose track of what we ordered to answer of the real incoming orders and ordered way too much. Such a behavior shows that any part of the supply chain tends to overreact in front of periods of rising/ falling demand. One overreaction in the up streamed part is enough to break the perfect equilibrium and make downstream participants ever more overreact. The order amount increased or decreased with every stage in the supply chain. * What are the causes of the phenomenal that you observe?
In theory this phenomenal (called bullwhip effect) should not occur if all orders exactly meet the demand of each period. The main issue in this game is that information about consumer demand is only passed up the supply chain through the orders that are placed. Information is systematically lost and result to high buffer stocks. The causes of this phenomenal can be divided into behavioral and operational causes: * What makes you change the order quantity each time? Why? appendix 1 ) For most of the game we had a low average order size, which averaged to 4 units per week. However, the order size was not constant and oscillated from 0 to 14. Our team did his best to adjust product ordering according to the signals transmitted by up-streamed parts through customer orders. Often times we would order slightly more than what we actually thought we needed. This behavior could be justified in many ways. First, we did this to account for different occasions such as graduation.
Next, we ordered more to fill inventory down the supply chain. Lastly, we occasionally ordered more when we had a “just in case” mentality. To an increase of customer order, we answered by placing more large orders. Such pathological incentives lead definitely to produce even more or less than the market demand and breaks definitely with the initial optimal market balance. If all supply actors act in such way, information about consumer demand is systematically lost from up to down streamed participants. The order amount increases with every stage in the supply.
And even if each part does it best to act ‘optimically’ individually, the result is less than optimal for the whole supply chain. * What are the impacts of such phenomenal to the supply chain performance in terms of cost and responsiveness? (a faire) * What would have helped mitigate the impacts of the phenomenal? (a faire) countermeasures: * Planning * reduce lead time of information (orders, demand and capacity forecast, point of sales data for the whole supply chain) * reduce lead time of material (just in time, postponement)