The Bullwhip Effect on Supply Process: Analysis of the Effect Using the Simulated Game Results - Essay Example

THE BULLWHIP EFFECT ON SUPPLY PROCESS: ANALYSIS OF THE EFFECT USING THE SIMULATED GAME RESULTS 1.0 Introduction The bullwhip effect is the change that occurs in supply of goods process where the quantity of goods demanded from the supplier is of different volume than the goods sold to the final consumer. Lee et al (1997) suggests that thiis occurrence creates a greater variation between amount of goods demanded at the consumer level and that ordered at the supplier level. This difference changes the smoothness of the process because at every level of supply process there exists over-approximation or under-approximation of the quantity of goods demanded. Commonly supply process begins from consumer/ customer then retailer then wholesaler/ distributor and lastly ends with the manufacturer/ supplier. The flow chart below shows the supply chain. The bullwhip effect on the supply chain is commonly caused by many factors which are also referred to as, “contributors of bullwhip effect in supply chains”. The common factors include the number of steps or levels involved in the process supply chain, the time taken to fulfill or accomplish an order, the incentives that are involved in making inventory levels to decrease and availability of information about supply chain status (Forrester, 1961). Fransoo (2000) cites that attempts to decrease inventory volume by avoiding to order, problems with communication throughout the supply process and existence of backlogs are contributors of bullwhip effect. Other causes of bullwhip effect are lack of coordination through the supply chain, delays in making delivery of goods in time, consumers ordering excess of what they require during low supplies and inaccurate forecasts of demand as contributors of bullwhip effect. The example below illustrates bullwhip effect on the process of supply starting from consumer level all the way to supplier. The actual demand for a product begins with the customer, but the actual demand for a product gets distorted as one moves down the supply process. Suppose a customer demands from a retailer 10 units of product, then the retailer may then order 12 units from the distributor (supplier). The extra 2 units are to ensure that the stock does not run out. The distributor (supplier) may then order 18 units from the producer or manufacturer. The producer or manufacturer allows the distributor to buy in bulk so that they increase their stock and ensure that the retailers get the goods without delay. As a result the manufacturer will get an order from the distributer or wholesaler and then demand from the supplier in large scale (bulk), ordering 30 units to ensure economy of scale in production to meet demand.  This results to production of more than 30 units for a demand of only 8 units hence the retailer will have to increase demand by dropping prices or finding more customers by marketing and advertising. 2.0 Analysis of bullwhip effect using simulated game results According to Wouters (2000), bullwhip effect is an increase in demand along all levels of supply process. In this effect the quantity of goods demanded at consumer level is constantly low but the quantity demanded at the supplier level is so magnified. Due to this changes of demand along the supply process the exists variations in demand at every level. To put this argument into perspective, we consider the two sets of data generated from the simulated games. These sets are named as SC1 and SC2. 2.1 Analysis of the bullwhip effect using the SC1 game results Going back to our data on simulated game results, we require the totals (sums), means, standard deviations and variances for the five levels across the supply chain starting with the consumer through the retailer, the wholesaler, the distributor and lastly the manufacturer. Then analysis of the above data gives the following results: Table 1.0: Statistics of the simulated game results of SC1 Level of Order Number Sum Mean Standard Deviation Variance Consumer 12 106 8.83 1.899 3.606 Retailer 12 42 3.50 2.276 5.182 Wholesaler 12 54 4.50 3.529 12.455 Distributer 12 56 4.67 4.119 16.970 Manufacturer 12 82 6.83 6.058 36.697 Starting with the sum (total) column, we realize that down the supply chain from the consumer to the manufacturer the sum of demanded quantities irregularly changes. Initially, the demanded volume is 106 which drops to 42 then starts increasing steadily to 54 then 56 and eventually 82. The column under variance shows that as we move down from the consumer to the manufacturer the variability in demand/ order increases. This is in agreement with the explanation of Lee et al (1997) definition of the Bullwhip effect. Therefore, one conclusion that we draw from this study is that across the supply chain the variance of demand (order) increases. In the statistical analysis, the variation described by variance or standard deviation is often replaced with the measure of relative spread. If absolute spread is measured by standard deviation (𝜎) as the root mean square of the deviation from the mean, and if the average is the mean of supply chain flows (μ), then the relative dispersion normally used in statistics is the coefficient of variation (Porteus, 2002). The coefficient of variation is given as; V = 𝜎/μ The estimation is expressed as a percentage, which would also be a good indicator here. But according to Chen et al. (2000), the following approaches of the bullwhip measure will be used as suggested and we shall not change the units upstream of flows. Cachon (1999) cites that, mathematically the Bullwhip effect per level is calculated using the standardized quantities for the incoming demand and outgoing demand. The means and standard deviations of the demand values are used to standardize the quantities involved. The expression below shows how this is done: Where; and Dout is the outgoing demand quantity Din is the incoming demand quantity And are mean and standard deviation respectively. Considering the above data and analysis based on Table 1.0, the bullwhip effect can be calculated by getting the quotient of standard deviation against its respective mean. The results are indicated below: Table 1.1 Results of standard deviation, mean and bullwhip effect of SC1 Level Consumer Retailer Wholesaler Distributer Manufacturer Standard deviation (𝜎) 1.899 2.276 3.529 4.119 6.058 Mean (μ) 8.830 3.500 4.500 4.670 6.830 Bullwhip effect (𝛚) 0.215 (21.5%) 0.650 (65%) 0.784 (78.4%) 0.882 (88.2%) 0.887 (88.7%) The results show that the bullwhip effect increases from 21.5% (0.215) at consumer’s level, and then 61.5% (0.615) at retailer’s level, 78.4% (0.784) at wholesaler’s level, 88.2% (0.882) at distributer level and eventually to 88.75 (0.887) at manufacturer’s level. These results concur with our expectations that along supply process the amplitude of bullwhip increases. The second part of analysis involves looking at the inventory and shortage expenses and how their trends changes across the supply chain. The results below show the totals of inventory and shortage accompanied by the respective costs across the supply chain. Table 1.2: Results of inventory and shortage costs in SC1 supply chain Level Retailer Wholesaler Distributor Manufacturer Inventory 42 54 56 82 Shortage 6 10 16 04 Inventory cost 21 27 28 41 Shortage cost 6 10 16 4 Total cost 27 37 44 45 The total cost of running the inventory and the shortage costs increases across the supply chain starting from 27 then 37 then 44 and eventually 45. Bearing in mind that the cost of running a unit shortage is twice as expensive as the cost of running a unit inventory then conclusively, it is better to maintain inventory rather than incurring expensive shortage costs. This is one of the reasons why bullwhip effect keeps magnifying itself across the supply chain. 2.2 Analysis of the bullwhip effect using the SC2 game results The simulated SC2 game results show different results on the ordering process from those of SC1. The totals (sums), means, standard deviations and variances for the five levels across the supply chain are shown by the following table. Table 2.0: Statistics of the simulated game results of SC2 Order level N Sum(Total) Mean(μ) Standard Deviation(𝜎) Variance (𝜎2) Customer 12 106 8.83 1.899 3.606 Retailer 12 115 9.58 3.370 11.356 Wholesaler 12 137 11.42 4.981 24.811 Distributor 12 103 8.58 6.431 41.356 Manufacturer 12 86 7.17 8.043 64.697 The sum (total) column shows an increase across the supply chain; starting from 106 units to 115 units then 137 in the customer, retailer and wholesaler levels respectively. After the wholesaler the demanded quantity decreases to 103 units at the distributor level and then further drop to 86 units at manufacturer level. This change is not the same as the one experienced in SC1 results in Table 1.0. To analyze the bullwhip effect across the supply chain, we consider the variability (the measure of variance) across the supply chain. The column of variance shows that as we move from the consumer to the manufacturer the variability in demand/ order increases. This is in agreement with the definition of the Bullwhip effect. Therefore, across the supply chain the variance of demand (order) increases. The Bullwhip effect per level is calculated using the standardized demand quantities for the incoming and outgoing process. The standard deviations are divided by the means of the demand values to get the bullwhip effect for all levels in the supply chain. The results of these quotients are shown in the table below. Table 2.1: Results of standard deviation, mean and bullwhip effect of SC2 Level Consumer Retailer Wholesaler Distributer Manufacturer Standard deviation (𝜎) 1.899 3.370 4.981 6.431 8.043 Mean (μ) 8.830 9.580 11.420 8.580 7.170 Bullwhip effect (𝛚) 0.215 (21.5%) 0.352 (35.2%) 0.436 (43.6%) 0.750 (75%) 1.122 (112.2%) The results show that the bullwhip effect increases from 21.5% (0.215) at consumer’s level, and then 35.2% (0.352) at retailer’s level, 43.6% (0.436) at wholesaler’s level, 75% (0.750) at distributer level and eventually to 112.2% (1.122) at manufacturer’s level. These results concur with our expectations that along supply process amplitude of bullwhip increases. Comparing the results of SC2 (Table 2.0) and those of SC1 (Table 1.0), the bullwhip effects in the first four levels of SC2 are lower than their respective ones in SC1. These results show that there are some interventions (remedies) that were in place in process SC2 that reduced the bullwhip effect in the first four levels. The bullwhip effect in level 5 of SC2 is higher than that of SC1 and it is even abnormally above 100%. The second part of analysis of SC2 supply chain involves looking at the inventory and shortage expenses and how their trends changes across the supply chain. The results below show the totals of inventory and shortage accompanied by the respective costs across the supply chain. Table 2.2: Results of inventory and shortage costs in SC2 supply chain Level Retailer Wholesaler Distributor Manufacturer Inventory 62 38 33 152 Shortage 4 14 43 19 Inventory cost 31 19 16.5 76 Shortage cost 4 14 33 19 Total cost 35 33 49.5 95 The total cost changes from 35 to 33 then 49.5 and lastly 95. Therefore, the cost of running the inventory and the shortage costs increases across the supply chain. Since the cost of running a shortage is twice as expensive as the cost of running a unit inventory then, it is better to maintain stock than incurring higher expenses on shortage. This is one of the contributors of bullwhip effect which keeps on magnifying itself along the supply chain. 3.0 The findings on above analysis According to Lee et al (1997), updating of quantity demanded through forecasts, fluctuations of prices due to shortages, effect of ration and the batching of orders are main factors behind bullwhip effect. In our discussion of the above analysis results, we will also base our discussion on the same factors. 3.1 Demand forecast updating From the above results it is realized that the connecting levels in the supply process have future anticipated demands depending on the orders they get from their previous levels. Increase in order given will pave way to forecast of higher demand and this affects the next level through ordering larger quantities. This occurrence will seek increased demands hence setting of different forecast and affecting the information flow throughout the supply process. The reverse also applies; that is, if the consumers decrease their demands then the forecast on demand by the next level is negatively affected. The possible remedy in countering the effect of forecast is to advise consumers/ customers to order directly from the origin without passing through intermediaries. On the other hand, if proper records of consumer demands are shared within the supply levels and singular forecast is done for the whole supply process then this problem can be curbed. 3.2 The order batching In order batching, our analysis shows that as Din quantity comes in and stock are in place, the next supply level will not order. This leads to piling up of demands before placing order since costs of ordering and delivery process are minimized. This factor can be addressed using the electronic device inventory. The electronic inventory will provide clear records or data about the supply chain that can be used to regulate quantities demanded and quantities supplied. 3.3 The price fluctuations The other common cause that can be attributed by our analysis of SC1 and SC2 results in bullwhip effect is price fluctuations. This factor is due to different business issues like business promotions through advertisements, business deals etc. When the prices of goods go up and down, this prompts variation in demand along the supply process. If prices are lower, the consumer purchases in large scale than expected and when the prices normalizes the consumer purchases less hence leading to stagnating stock in the stores. Methods of sales promotion like advertisements, market reorganization and other business changes will cause changes in price. Therefore when it is anticipated that the prices are to go down then orders go up and the opposite applies. Stabilizing prices through government interventions, sensitization process, minimizing number of sales promotions and decreasing advertisements are ways of decreasing this effect. 3.4 Effect of ration and shortage gaming The other cause of bullwhip effect analyzed in our study is ration and shortage gaming. If a good is demanded in excess of the present supply, the producer will reduce the goods through rationing while understanding that the consumers may demand in excess of what they require. After this occurs then later the supply may be in plenty and therefore lessening orders. This factor may lead to magnifying of orders at one point and reducing the orders at another point. 4.0 Other issues realized in the analysis of the simulated game results The research paper found out that, some issues should be considered when measuring the bullwhip effect. These issues are highlighted below: 4.1 Aggregation To accurately measure the total bullwhip effect we require the information of all orders in all supply levels at a particular time. The point at which the standard deviation of an order is calculated affects bull whip magnitude. When calculated at the production level it shows different value than when calculated at every supply level. In conclusion, to determine the bullwhip effect accurately we need to know the different aggregation level that we can adopt. 4.2 Incomplete data The data of demanded quantities may be missing in some supply levels or distorted in some levels. In other cases only data for greatly aggregated goods are available. Therefore calculations of bullwhip effect on such cases will be difficult. 4.3 The filtering for causes According to Fransoo and Wouters (2000), to determine bullwhip effect, one needs to actually determine the main causers of the effect. For them they realized that to filter the contributors of bull whip effect can be aggravating task. However they noted that changes in prices irregularly had negligible effect while rationing and shortage gaming had a greater contribution of bullwhip effect. 5.0 Conclusions of the study The aspects that influence the magnitude of the bullwhip effect contribute differently. Some of these aspects (factors) that our study deemed responsible for bullwhip effect in supply link are listed below: The lengths of supply chain levels and respective number of steps involved The velocity and required duration for an order to be fulfilled The conditions or incentives towards to over-reaction and panic ordering as inventory levels decrease Disorganization between each supply chain link; this causes demanding of uneven quantity of goods than is required. The uneven demand is because of different reactions to supply chain at hand. When there is no proper communication within the supply chain process, the process becomes cumbersome and difficulty. Lack of proper communication makes managers to perceive demand of a good in different ways between supply chain links. The choices of free to return whereby consumers on a free will over order because of expected supply shortages and then forfeiting the supplies when the supply normalizes. This results to demands in excess. Order batching, is when the incoming demand quantity is made and stocks are in place therefore the next supply level will not order. If this occurs then demands will increase before placing orders on the next supply level. This condition will produce an anomaly that creates variation in demand hence causing bullwhip effect. Frequent changes in prices due to market aspects like discounts given, government policies and seasonal changes can influence the pattern of purchases. Normally consumers buy highly when offered discounts and reduce their purchases when discounts are not given. This will cause irregular demand, irregular orders and eventually irregular supply Lack of data about current supply chain status and quantity demanded will fluctuate the quantity ordered. At times the managers will rely on the estimates of order (demand) from past records which may not necessarily reflect the current situation hence leading to fluctuations in demand. To overcome the bullwhip effect we need to employ countermeasures (remedies) to overcome it. Here are some of these solutions or simple remedies that can counter the bullwhip effect: Remove or decrease the intermediaries so as to shorten the supply process. Quicken the supply and decrease the fulfillment of demand time Make order sizes limited and give appropriate incentives Maintain correct inventory and order in every step of the supply process On order batching the appropriate countermeasure is to use electronic systems to check order batching. In cases of high order cost, the remedy is to employ use of electronic data interchange. In case of random ordering, it can be counterchecked with definite periodical deliveries. These two remedies will lead to increased frequency of smaller orders and consequently smaller variations will be realized upstream the supply chain. The countermeasure to shortage gaming is apportioning rationing schemes proportionately. This can be achieved through allotting quantities to the next supply level while considering past records. To avoid lack of information in the supply process, the supply chain information and capacity reservations can be shared within the whole link. The order size flexibility and the capacity reservoir should be decreased so that unrestricted ordering capability is checked. The countermeasure to the frequent changes in prices the minimal and maximum price setting can be by offering special ordering system that can be put in place so as to order regularly hence synchronizing the purchase and demand processes. The inaccuracies of forecasts of demands can be checked by providing accessibility to the sales records or data. On the choices of free to return whereby consumers demand on a free will without considering the later supply trends is not easily countered but, such choices can often be addressed through government interventions or punished through legal process to forewarn others. Read More