Causes of the bullwhip effect : A study of the bullwhip effect in the Volvo Group Service Market Logistics’ supply chain

(1)

Linköping University | Department of Management and Engineering | Master’s thesis, 30 credits Mechanical Engineering with the Master’s Program in Industrial Engineering and Management

Spring 2021 | LIU-IEI-TEK-A--21/04179—SE

Causes of the bullwhip

effect

– A study of the bullwhip effect in the Volvo Group

Service Market Logistics’ supply chain

Klara Dahlin

Oscar Säfström

Supervisor: Fredrik Stahre

Examiner: Mikael Malmgren

Supervisors at Volvo Group: Christian Beckers and Debbie Lau

Linköping University SE-581 83 Linköping, Sweden +46 013 28 10 00, www.liu.se

(2)

Abstract

The bullwhip effect is defined as an upstream amplification of demand variability and has received interest within multinational companies for decades. As early as in the 1950’s, Forrester (1958) discussed what is today known as the bullwhip effect, which has a negative impact on the customer service, costs, and inventory investment in a supply chain (Lee et al., 1997). Even though the bullwhip effect has been noticed in various industries, the consequences, in form of decreased availability and increased costs the further up the supply chain the bullwhip goes, still remain. The employees at Volvo Group Service Market Logistics suspect that their supply chain has been affected by the bullwhip effect and want to know if it is correct and subsequently know why it has occurred. Therefore, this master’s thesis highlights the root causes of the bullwhip effect and presents strategies to mitigate it. To

understand how the bullwhip effect affected the Volvo Group Service Market Logistics’ supply chain, the purpose was formulated as follow:

The purpose of this study is to identify events in the Volvo Group Service Market Logistics’ supply chain where the bullwhip effect has occurred, its root causes, and how to reduce or eliminate the bullwhip effects.

The studied flow was from the Central Distribution Center (CDC) in Ghent, to the Regional Distribution Center (RDC) in Brazil, to the Dealers associated to the RDC in Brazil, and the customers. Data was collected from each node and events were studied to find bullwhip events. After sorting out the part numbers that passed the criteria for bullwhip events, the amount of data had to be reduced even more. A couple of different conditions were applied which resulted in four suitable bullwhip events.

Thereafter, the authors conducted interviews with Logistics Managers at each node of the supply chain to find the root causes of the bullwhip effect in each studied event.

Among the several found root causes, lack of information transparency was the most frequent occurring root cause, found in three out of four studied bullwhip events. Insufficient communication and lack of information sharing cause bullwhip effects, and the authors found that improved communication both between and within the nodes will contribute to better planning, and consequently avoided bullwhip effects. Other root causes found were issues with the ordering system, lack of learning and experience, neglected lead times, fear of empty stock, price fluctuations, and phase-out of the spare part.

To reduce or eliminate the bullwhip effect, the focus was on mitigating the root causes since the root causes create opportunities for the bullwhip effect to occur. Four suggestions were given with suitable mitigation strategies found in the literature, where the four suggestions were sales campaigns, prepare for boosts, keep track of manually placed orders, and ordering system and Logistics Manager

behavioural issues. The suggestions could then be connected to the different found root causes. The stated suggestions and mitigation strategies focused on mitigating the root causes in a long-term perspective and consequently the bullwhip effect itself.

(3)

Acknowledgement

This master’s thesis is the final project of the Mechanical Engineering program with the master’s program in Industrial Engineering and Management at Linköping University. It was conducted over 20 weeks at the Service Market Logistics department at Volvo Group by the two authors.

The authors first want to thank their supervisors at Volvo Group, Christian Beckers and Debbie Lau, for believing in us and being a great support throughout the spring semester, and the employees at Volvo Group who took good care of us. Moreover, the authors want to thank their supervisor from the university, Fredrik Stahre, for coming with invaluable advice. Also, a thank you to our opponents who took their time to go through our report and give feedback.

(4)

Nomenclature

Abbreviations

CDC: Central Distribution Center DIM: Dealer Inventory Management DIP: Demand and Inventory Planning EOQ: Economic Order Quantity JIT: Just In Time

LPA: Logistics Partnership Agreement RDC: Regional Distribution Center SDC: Support Distribution Center SML: Service Market Logistics TPO: Time Planned Order VMI: Vendor Managed Inventory

(5)

Table of figures

Figure 1 - The specified system. ___________________________________________________________________ 2 Figure 2 – An overview of how the CDCs, RDCs and SDCs are located across the world. _______________________ 4 Figure 3 – The order flow in the Volvo SML supply chain. _______________________________________________ 5 Figure 4 – The responsibilities of the DIP Team, the Refill Team, and the DIM Team. _________________________ 5 Figure 5 – Illustration of a bullwhip effect and description of the amplification. _____________________________ 7 Figure 6 - Causes of the bullwhip effect in a supply chain. _____________________________________________ 13 Figure 7 – A summary of the strategies to mitigate the bullwhip effect. __________________________________ 18 Figure 8 - An illustration of an ongoing bullwhip effect. _______________________________________________ 20 Figure 9 – An illustration of a smoothening of the bullwhip effect. ______________________________________ 20 Figure 10 - The studied system. __________________________________________________________________ 21 Figure 11 – Displaying an increase in order variance in each node of the supply chain. ______________________ 22 Figure 12 – Illustration of the fictional example’s supply chain. _________________________________________ 23 Figure 13 - Example of orders in the fictional supply chain. ____________________________________________ 24 Figure 14 - A summary of the purpose of the study and the study's questions. _____________________________ 27 Figure 15 - The phases of the study. _______________________________________________________________ 28 Figure 16 - An illustration of how the questions related to Research Question 1 depend on each other. _________ 31 Figure 17 – A simplification of a bullwhip event, which shows customer sales data, orders from the Dealer to the RDC and orders from the RDC to the CDC. __________________________________________________________ 33 Figure 18 - An illustration of how the questions related to Research Question 2 depend on each other. _________ 35 Figure 19 - An example of the Five-Why method. ____________________________________________________ 36 Figure 20 – An illustration based on Latino’s (2002) logical tree. ________________________________________ 37 Figure 21 – Suggestions of how to reduce or eliminate the bullwhip effect, where each suggestion is connected to the affected root causes. ________________________________________________________________________ 41 Figure 22 – An illustration of how orders at each node were compared. __________________________________ 45 Figure 23 – How the air freight and backorder data was presented. _____________________________________ 45 Figure 24 - A graph of bullwhip event A ____________________________________________________________ 47 Figure 25 - A graph of bullwhip event B ____________________________________________________________ 48 Figure 26 - A graph of bullwhip event C ____________________________________________________________ 49 Figure 27 – A graph of bullwhip event D ___________________________________________________________ 50 Figure 28 –The availability in the CDC. During the weeks with a negative inventory level, there were backorders at the CDC. _____________________________________________________________________________________ 52 Figure 29 – Sales and orders from each node, and the inventory level at the RDC in Brazil. ___________________ 52 Figure 30 – An overview of information and material flow of the Volvo SML supply chain. ___________________ 53 Figure 31 – An illustration of placed orders in each node and the availability at the RDC in Brazil on a monthly interval, for Part Number A. _____________________________________________________________________ 58 Figure 32 – An illustration of the availability at the CDC for Part Number A on a weekly interval. ______________ 59 Figure 33 – Increased orders and sales in September and October due to a Top 1000 boost. __________________ 60 Figure 34 – An illustration of placed orders in each node and the availability at the RDC for Part Number B on a monthly interval. ______________________________________________________________________________ 61 Figure 35 – An illustration of the availability at the CDC for Part Number B, on a weekly interval. _____________ 62 Figure 36 – Preparation for the sales campaign. _____________________________________________________ 63 Figure 37 – An illustration of placed orders in each node and the availability at the RDC in Brazil for Part Number C, on a monthly interval. __________________________________________________________________________ 64 Figure 38 – An illustration of the availability at the CDC for Part Number C, on a weekly interval. _____________ 65 Figure 39 – An illustration of the increase in orders at each node and the overreaction from Logistics Managers at the RDC. _____________________________________________________________________________________ 66 Figure 40 – An illustration of the availability at CDC for Part Number D on a weekly interval. _________________ 67

(9)

Figure 41 – An illustration of the delayed season and the double placing of air- and sea freight orders. _________ 68 Figure 42 – Mapped strategies to reduce or eliminate the bullwhip effect. ________________________________ 74 Figure 43 – Mapped strategies to reduce or eliminate the bullwhip effect. ________________________________ 80

(10)

1

1. Introduction

In this section, an introduction of the study will be presented. The chapter will include background, purpose, the scope of the study, and directives and delimitations.

1.1 Background

Volvo Group is a Swedish company that operates within the area of for example infrastructure solutions, and offers trucks, buses, construction equipment, power solutions for marine, and industrial

applications (Volvo, 2021). From now on, Volvo Group will be entitled as Volvo in this report. Volvo Group Service Market Logistics division, called Volvo SML in this report, serves the demand of spare parts to multiple brands all over the world. The customers of Volvo purchase service contracts to ensure that maintenance will be available at the time when it is needed. For the customer, these service contracts reduce the risk of loss in income caused by long lead times in repairing vehicle breakdowns, andVolvo will simultaneously gain value through their offered services. To substantiate the demand of spare parts, Volvo has developed a wide logistical network with multiple actors (Lau, 2021).

A bullwhip effect is the phenomenon where the orders upstream the supply chain are of a larger scale than the actual sales to the customers. With every actor in the upstream supply chain, the

misrepresentation will increase (Lee et al., 1997) in cases of production and inventory decisions (Hu, 2019). To supply the customers, Volvo SML tries to keep a high stock availability and at the same time keep the supply chain costs as low as possible. Volvo SML has previously had problems with reduced availability of spare parts in its Central Distribution Center which the employees believe is caused by the bullwhip effect, although they do not have a lot of knowledge about it. Therefore, they want to

investigate if the bullwhip effect occurs in their supply chain and if it does; know why it occurs. Forrester (1958) discussed the bullwhip effect already in the 1950’s but the phenomenon still occurs frequently in supply chains. Since the bullwhip effect is still occurring, this study is not only relevant to Volvo, but also to anyone who is interested in the reasons behind the bullwhip effect in other supply chains.

According to Forrester (1961), the loss of efficiencies refers to the increasing swings in inventory in response to fluctuations in customer demand, and the further up the supply chain the bigger impact it has. Previously, the lack of stock in the warehouses in Volvo SML has caused problems since it will then not reach the aimed service level, which might lead to losing customers to the competitors. To secure future competitiveness in the service market, Volvo wants to improve the understanding of the bullwhip effect by identifying events where the bullwhip effect has occurred in the past and investigate why the bullwhip effect occurs in the Volvo SML supply chain. The Volvo employees also want to know how the bullwhip effect can be reduced or eliminated in the Volvo SML supply chain. Therefore, the purpose of the study will be as displayed in section 1.2.

1.2 Purpose

The purpose of this study is to identify events in the Volvo Group Service Market Logistics’ supply chain where the bullwhip effect has occurred, its root causes, and how to reduce or eliminate the bullwhip effects.

(11)

2

The term event is in this study defined as the order flow for a certain part number during a specific time period. A bullwhip event is an event that includes one or several bullwhip effects, meaning that at least one bullwhip effect has occurred during the specified time period and part number. Every bullwhip event can have one or several underlying root causes.

This study defines root cause as the main cause of the bullwhip effect. In other words, it is the root cause that triggers the bullwhip effect to occur. However, there can be a combination of root causes that trigger the bullwhip effect, and they must necessarily not cause a bullwhip effect when isolated. An example of a root cause to the bullwhip effect is that a Dealer places excessive orders due to the fear of getting a stock-out.

1.3 The scope of the study

In the Volvo SML supply chain, customers order their parts from the Dealer warehouse. The Dealer then places an order to the Regional Distribution Center (RDC). However, if the order is urgent and within Europe, the Support Distribution Center (SDC) supports the RDC with urgent orders. SDCs are in general located closer to the customers in Europe than the RDCs. The Central Distribution Center (CDC) provides spare parts to RDCs, SDCs, as well as straight to the Dealers, and the CDC manage their contact with needed suppliers. The Volvo SML supply chain covers suppliers, CDC, SDCs, RDCs, Dealers’ warehouses, and customers, but this study will focus on the flow including RDCs and excluding the SDCs, as per the directives in the next section. A node is defined as each member of the supply chain that a certain spare part pass and has the opposite direction than the echelon levels.

Figure 1 - The specified system.

1.4 Directives and delimitations

To be able to perform the study, some directives and delimitations must be made. These will be based on directives from Volvo and suggestions from the supervisors from both Volvo and Linköping

(12)

3

1. The studied spare parts belong to Volvo’s brand Volvo Trucks. These are the parts that have the highest demand at Volvo SML and the largest revenue among the brands in Volvo’s service market. Another reason is that the demand is less affected by trends and seasons than other brands in Volvo.

2. The study will only involve parts that are stored in, and transported via Ghent in Belgium, which is Volvo SML’s central distribution center in Europe. Most of the spare parts pass through the CDC in Ghent and therefore it is more important to study. The CDC in Ghent will be called the CDC in this study.

3. The scope includes RDCs and excludes SDCs and orders that goes straight from the CDC to the Dealers. The SDCs are only represented in Europe and are therefore not as influenced by the lead times as the RDCs which are located further away from the CDC. When studying the flow where the spare parts go straight from the CDC to the Dealers, the bullwhip effect has less impact on the supply chain performance as there are a fewer number of nodes.

4. The type of spare parts that have a large economic impact on Volvo SML are parts classified as fast-moving. Therefore, the study will have a focus on analysing this category of spare parts. Fast-movers are spare parts which have a high turnover rate, the demand is high and there is a constant flow of incoming orders. Each node of the supply chain has different spare parts that are classified as fast-moving, but in this study the classification is based on what the CDC defines as a fast-moving spare part.

5. Some spare parts are assembled with other spare parts to create a kit. The kit then gets a different identification number and will not be seen in the sales statistics of the original spare part. Therefore, the spare parts that are part of kits will not be included in this study.

(13)

4

2 Case situation

In this part of the report, a brief description of the situation and the company will be presented.

2.1 Organization

Volvo was founded in 1927 in the Swedish city of Gothenburg. In the beginning, the focus was on cars and then expanded to other areas. Volvo sold Volvo cars to Ford Motors in 1999, and a new Group was created with a focus on the industry of commercial automotive (Volvo Group, 2021). The company has close to 100,000 employees and distributes products to about 190 countries.

Today Volvo consists of ten different business areas: Volvo Trucks, Renault Trucks, Mack Trucks, Volvo Construction Equipment, Volvo Buses, Volvo Financial Services, Volvo Penta, Volvo Autonomous

Solutions, Volvo Energy, and Arquus (Volvo Group, 2021). This study will be executed under Volvo Trucks Operations, which belongs to the area of Volvo Trucks. Volvo Truck Operations is divided into different areas, where this study will be focused on the Service Market Logistics which works with the spare parts to the aftermarket.

2.2 Service Market Logistics

Volvo SML has CDCs, SDCs, and RDCs in several locations across the world. Where these are located can be seen in Figure 2.

Figure 2 – An overview of how the CDCs, RDCs and SDCs are located across the world.

For the spare part aftermarket, there are two types of sales: over counter sales and repair sales. The customers of Volvo SML are mostly companies that own trucks from Volvo. The order flow of the Volvo SML supply chain starts with the customer placing an order to their correlating Dealer. The Dealer then orders from the RDC, SDC, or straight from the CDC. The RDC and SDC will order from the CDC, which in turn order from suppliers. However, the SDCs and the orders going straight from the Dealers to the CDC

(14)

5

will be out of scope of this study and will therefore not be discussed any further than this. See Figure 3 for the order flow in the supply chain.

Figure 3 – The order flow in the Volvo SML supply chain.

The responsibilities of the inventory and forecasting in the different nodes of the supply chain are divided into three sub-divisions: Demand and Inventory Planning (DIP), Refill Team, and Dealer Inventory Management (DIM). In Figure 4, the different areas of responsibility are displayed. The DIP Team controls the inventory in the CDC where they forecast the demand flow between the CDC to the RDCs. The Material Planners collaborate with the DIP Team and supply the demanded quantity of spare parts to the CDC. The Refill Team’s responsibilities are to fill up the RDCs and receive demand data from the DIM Team. Usually, the DIM Team supports the Dealer with inventory strategies and inventory levels, but in some cases the Dealers control their own inventory. The customers place their orders directly to the Dealer Warehouses, which in turn base their demand on historical customer data patterns. The different teams and their corresponding responsibilities are presented in Figure 4.

(15)

6

3 Theoretical framework

This chapter aims to create an understanding of previous research and literature that are related to this study. The first section of the chapter focuses on explaining what the literature says about the bullwhip effect, the second section presents literature about the causes of the bullwhip effect, the third how to mitigate the bullwhip effect, and the fourth explains the bullwhip effects’ impact on the availability.

3.1 Bullwhip effect

Companies want to coordinate the flow of products from one end of the supply chain – part suppliers – to the other – customers – but it is almost impossible to coordinate activities in multiple companies in a cohesive way (Oskarsson et al., 2013). Companies that are a part of a supply chain need to cooperate with other companies to secure material supply and to have a trustful distribution (Oskarsson et al., 2013). Lack of communication and misinterpretations within the supply chain might lead to reduced availability and extra inventory costs due to the bullwhip effect. According to Bhattacharya and Bandyopadhyay (2010), the bullwhip effect can have an impact on the supply chain in various ways, where some of them are that the total cost for the whole supply chain increases, lower availability of products and a loss of revenue.

According to Jeong and Hong (2019), small adjustments in the orders from a customer can amplify an increase in the order quantityin the supply chain when the information is transferred upstream. When this phenomenon occurs, it is recognized as the bullwhip effect, see Figure 5. In manufacturing, this means that the demand pattern coming out of a node in the supply chain is greater than the variation of the demand that came into the same node. Wang and Disney (2015) define the bullwhip effect as the amplification of order volatility throughout the supply chain. By taking Wang and Disney’s (2015) definition into account, the bullwhip effect can be detected by measuring the volatility and the

parameters’ variance, coefficient of variation, or standard deviation. The variance is commonly used in mathematical analysis and it is therefore convenient to measure the bullwhip effect by comparing the variance between actual demands and orders (Wang & Disney, 2015). With these measurements a bullwhip effect can be identified with two conditions: one condition that studies variance, see equation 1, and the second that studies the amplification, see equation 2. The number of nodes in the supply chain is denoted as n in both the equations below, whereas the nodes are counted from the customers towards the suppliers (customer: n = 0), meaning that the variance and amplification increase upstream the supply chain.

𝑉𝑎𝑟𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛< 𝑉𝑎𝑟𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛+1 < 𝑉𝑎𝑟𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛+2 (1)

𝐴𝑚𝑝𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛 < 𝐴𝑚𝑝𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛+1 < 𝐴𝑚𝑝𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛+2 (2)

If the ingoing signal is mainly a stationary one with considerable noise, the noise amplifies the signal in each node. The amplified noise at each node will in the end lead to order swings which end up in a bullwhip effect (Tang et al., 2020). An example of amplification between a customer and a supplier is illustrated in Figure 5.

(16)

7

Figure 5 – Illustration of a bullwhip effect and description of the amplification.

3.2 Causes of the Bullwhip Effect

Why a bullwhip effect arises is either connected to operational or behavioural actions. Operational causes refer to the physical and institutional structure (Sterman J. D., 2005). The physical structure handles the placement of inventories throughout the supply chain, time delays in production, transportation and so on. The institutional structure includes the coordination among firms in a horizontal and vertical way, availability of information between organizations and departments, and incentives faced by each decision maker (Sterman J. D., 2005).

The behavioural causes include how the Logistics Managers base their decisions. These behavioural causes encompass attitudes of the decision makers, attributes about other actors, and the routines they use to analyse information and make decisions about production, capacity, and planning (Sterman J. D., 2005). According to Croson et al. (2014), behavioural causes of the bullwhip effect arise from suboptimal decision-making which will have a negative impact on the supply chain performance. In this section will operational and behavioural causes be stated, followed by a description of the types of failures that trigger the root causes.

3.2.1 Operational causes

Demand forecast updating

These recurrent adjustments create chaotic responses at the end of the supply chain. Order managers are in use of replenishment rules which have a stabilizing and smoothing effect on orders (Disney & Labrecht, 2008). Usually, the decision makers do not react to the market, they react to stabilized orders from actors downstream. When the actors use single echelon inventory optimization to optimize and

(17)

8

forecast inventories, Logistics Managers upwards the supply chain reacts to stabilized orders. This causes disharmony in the overall supply chain and order swings occur (Disney & Labrecht, 2008). The selection of forecasting method has a significant influence on the bullwhip effect. This creates irregularities in demand which sometimes do not correspond to the actual demand, whose demand forecast information will be used further upstream the supply chain (Bhattacharya & Bandyopadhyay, 2010). Misleading demand forecast information will harm the supply chain performance when the information is used further up the supply chain. The existing forecasting methods can be categorized into the following divisions:

1. Time series models 2. Machine learning models 3. Agent-based models

4. Control engineering models

Chaing et al. (2016) investigated multiple demand forecasting models and their impact on the bullwhip effect. The models forecast the demand and a measure of the bullwhip effect were made by calculating the ratio between the variation of orders and the variation of demand. The investigation shows that the bullwhip effect varies a lot depending on the choice of forecasting method. Some of the methods are consistently creating greater bullwhip effects and the investigations show that the bullwhip effect can be multiple times more serious depending on which forecasting technique being used.

Ordering policy and batching

Ordering policies and batch sizes have a principal role in the amplification of order variability compared to fluctuations in customer demand (Lee et al., 1997). There are two main types of batching that can be used. Time-based, or periodic, batching is a strategy where orders are placed within an interval, once a day for example. The second type is based on order quantities and is often a result of the use of an Economic Order Quantity (EOQ) ordering policy (Potter & Disney, 2006). The EOQ policy minimizes the ordering and holding cost (Bhattacharya & Bandyopadhyay, 2010). Each batch policy is used for different purposes and the periodic batch policy is commonly used in supermarkets due to the steady demand from customers. The batching policy based on order-quantities may be needed in industries where the customer has set a minimum order size.

Burbidge (1961) states that the optimal batch size is one unit, but there may be circumstances where this is not possible. Therefore, it is important to understand the impact of the batch size. Cachon (1999) states two operational approaches to reduce the bullwhip effect. The first is balancing retailer order intervals, which means that customers are placing orders on different days. The second is a flexible quantity strategy, which means that it is good to reduce the fixed order quantity and increase the time between orders. Order large quantities rather than EOQ while ordering production set-ups and

transportation will increase the variety of the size of each order. Large variety will make it difficult for the supplier to supply the demand, and this variety will increase the risk of a bullwhip effect arising (Disney & Labrecht, 2008).

(18)

9

Effect of inventory policies

The inventory policies are strategies of how to control the inventory levels. The decision policies specify at what time the replenishment of the inventory should be initiated and the quantity of it. What time each order is placed and what quantity is dependent on the inventory level. The inventory policy has an impact on the inventory level, which leads to the inventory policy having an impact on the bullwhip effect (Bhattacharya & Bandyopadhyay, 2010). There are studies on which inventory policies companies should use, based on how the business is structured. Disney and Towill (2003) studied inventory policies and compared the traditional way of running a supply chain against a vendor managed inventory (VMI) supply chain. While analysing the Bullwhip effect, Disney and Towill (2003) demonstrated that a VMI supply chain performed better than a traditional supply chain in responding to variation in orders.

Effect of lead time

Heydari et al. (2009) analysed the lead time variability in a supply chain and stated that the order variability increased with increased lead time variability. However, Heydari et al. (2009) also showed that the bullwhip effect does not have any impact on the order variability if the variability increased uniformly. Bhattacharya and Bandyopadhyay (2010) refer to Wang et al. (2008) who state that the bullwhip effect increases with increased lead time. Wang et al. (2008) also state that a reduction in lead time will be useful in the work of reducing safety stock, out of stock losses, and improving the service level. There are multiple principles for reducing the lead time and one of them, the time compress principle, is stated in section 3.3.

Rational and shortage gaming

Shortage describes a buyer’s approach to managing shortages in supply, and rationing is the suppliers' approach to rationing during shortage periods. In cases where shortage events occur, buyers tend to place irrational orders to fulfil their inventory position, where these irrational orders tend to increase the amplitude of the bullwhip effect (Bhattacharya & Bandyopadhyay, 2010). These enhanced orders increase the order variance and make it difficult to forecast the actual demand of products (Lee et al., 1997).

Price fluctuation

Forward buying is a result of price discounts, coupons, and other special promotions on the market. These offers result in that customers’ orders are of a larger quantity than needed (Disney & Labrecht, 2008). They will then stop buying until the inventory level has weakened. When the product price has returned to its normal level, the number of incoming orders will be less than during the sales campaign. As a result, the order pattern does not reflect the consumption pattern, and the variation of bought quantities is much larger than the variation of the consumption rate (Lee et al., 1997). These variations cause the bullwhip effect. Lee et al. (1997) also state that forward buying makes sense if the holding cost is less than the price differential, although they state that most companies mostly suffer when using it.

(19)

10

Lack of transparency

What is meant with information sharing is that demand information at a downstream site is transmitted upstream the supply chain. Bhattacharya and Bandyopadhyay (2010) claimed that improved information sharing and coordination between different nodes in the supply chain reduce the bullwhip effect. Distortions, lack of transparency and information sharing provides conditions for a bullwhip to arise. To reduce the risk of a bullwhip arising, it is according to Almeida et al. (2015) valuable to share information related to historical demand at customers, and scheduled order requests further down the supply chain. In terms of forecasting Trapero et al. (2011) state that information sharing improves forecast

performance and Chong and Liu (2012) state that information sharing reduces the fluctuations in inventory levels.

Geary et al. (2006) argue that the members of the supply chain should have access to data – such as work in progress, flow rates, and inventories – of the other members of the chain. This to minimize the double guessing and information delays from one member to the next. According to the study made by Lee et al. (2000), a supply chain member upstream the chain can gain cost reductions and inventory reductions by implementing information sharing with an actor one step down the supply chain. It can also lead to an improvement in the effect of demand distortion (Lee et al., 2000).

Information distortion leads to an increased bullwhip effect and information sharing leads to reducing the effect (Jeong and Hong, 2019; Dai et al., 2016). With a higher rate of information sharing at each echelon in a supply chain, the bullwhip effect can be reduced. However, the impact of the bullwhip effect differs from one layer of echelon to another and the influence of the information sharing rate decreases with each member down the supply chain (Jeong & Hong, 2019). According to Dai et al. (2016), if there is real-time information sharing in the supply chain, the experienced bullwhip effect by the manufacturer is decreased and so are the costs.

3.2.2 Behavioural causes

Fear of empty stock

The Logistics Managers at each node may place excessive orders due to the fear of getting a stock-out and losing customers. This behaviour might lead to a large variation in order sizes and cause a bullwhip effect (Bhattacharya & Bandyopadhyay, 2010). At the time of upswings in demand and the supplier is unable to increase the production speed, products becomescarce and customers are placed in allocation. Customers will then respond to the longer lead time and the unreliable suppliers by increasing inventory levels, order farther ahead, and request a surplus of orders due to the fear of getting stock-outs (Almeida et al., 2015). For example, if the supplier will be able to deliver say 70% of its order, the customer will then likely order 130% or even more of what they require next time to fulfil the demand of the customer. Firms that lose these allocation battles will probably lose market shares and customers can be left standing with too many products (Sterman J. D., 2005).

(20)

11

Logistics Manager’s responses

Sterman (1989) models managerial behaviour decisions based on the Beer Game, where he states that managerial behaviour has an impact on the bullwhip effect in parameters such as oscillation,

amplification, and phase lag. The result from the Beer Game problem shows that rounds ran by people are characterized with instability and oscillation and the effective inventory becomes negative, which indicates that the sectors have backlogs. Sterman (2005) formulates that the patterns between different games are similar to each other: it starts with inventory decline throughout the supply chain and the Logistics Managers will react to the upcoming backlogs. To reduce the number of backlogs, managers place a wave of orders, which are growing larger at each node. This wave of orders will reduce the stock-outs and probably overshoot the inventory. These overshoots are not acceptable from a cost-optimizing point of view and the managers will manage the inventories by abruptly stop placing new orders, where the inventory peaks then slowly decline. These behavioural regularities are even more interesting due to that there is no variation in customer demand, which means that the oscillation throughout the supply chain is a consequence of how the managers respond (Sterman J. D., 2005).

Lack of learning and training

Negligence and disregard from the Logistics Managers generally arise due to a lack of knowledge or training (Bhattacharya & Bandyopadhyay, 2010). Yan and Katok (2005) show that order variability decreased significantly when Logistics Managers had experience in a related logistical situation and when Logistics Managers developed team strategies collaboratively.

Neglecting time delays

Croson et al. (2014) state that the Logistics Managers do not adequately account for the time delays, feedbacks, and non-linearity in the system. The Logistics Managers then tend to base their decision on orders based on the difference between the stock on hand and their target level, and the already placed orders have not been taken into consideration which causes instability and variations in placed order quantities. The time delays may include delays in order, delivery, or shipping which should be cut down to reduce the bullwhip effect (Bhattacharya & Bandyopadhyay, 2010).

Instability and trust

The trust factor can be divided into two groups, affective trust and trust in competence. Ha et al. (2010) describe the affective trust as the level of emotion and personality developed in a long-term relationship and the trust in competence refers to rational decisions. The affective trust has an impact on

collaboration and information sharing between firms. If the suppliers trust their partners, they will gain trust and more frequent communication with their partners which will lead to improved information sharing (Almeida et al., 2015). Decisions in many areas, especially strategic ones, must be shared with partners who have enough technical knowledge to contribute to the supply chain performance. For this reason, trust in competence is important during collaborative decision-makings because wrong

(21)

12

same way, Fawcett et al. (2012) claim that trust in the supply chain is based on competence and performance commitments.

Trust between partners can have an impact on the supply chain performance (Sterman J. D., 2005). When Logistics Managers find their suppliers to be unreliable, delivery quotes are not met and the customer receives less than the placed order due to a shortage of supply. The blame and mistrust will then worsen the instability of the supply chain.

Coordination risk

Croson et al. (2014) define coordination risk as a behavioural cause of the bullwhip effect. The

coordination risk arises when members of the supply chain perceive a risk that the other members will not behave optimally and therefore compensate by placing disproportionate orders. A decision-maker who is uncertain if the partners make incorrect orders and decisions may decide to diverge from the original strategy to create a buffer to make up for a non-optimal behaviour. This will lead to a variation in demand towards the suppliers (Almeida et al., 2015). Such deviations might generate latent instability in the supply chain. When the optimal decision rule is not known between the members of the chain, this reaction from the decision-maker is especially likely to happen (Croson et al., 2014).

3.2.3 Summary of root causes

(22)

13

Figure 6 - Causes of the bullwhip effect in a supply chain.

To improve the reliability of the study, each reference used to state the operational and behavioural root causes in section 3.2.1 and 3.2.2 is presented in Table 1. In the table, all root causes from the literature study are presented and marked with the corresponding author.

(23)

14

Table 1 – Compiled root causes correlated with authors.

Demand forecast updating Ordering policy and batching Effect of inventory policies Effect of lead time Rational and shortage gaming Price fluctuation Lack of transparency Fear of empty stock Logistical managers’ responses Lack of learning and training Instability and trust Neglecting time delays Coordination risk Bhattacharya and Bandyopadhyay (2010) X X X X X X X X X X X X X Sterman (2005) X X X X X X X X X

Disney and Labrecht

(2008) X X X X X X

Lee et al. (1997) X X X X X

Potter and Disney (2006) X X X X X

Devika, Jafarian, and

Hassanzadeh (2016) X X X X X

Geary et al. (2006) X X X

Dai, Li, Yan, and Zhou

(2016) X X X

McCullen and Towill

(2001) X X

Lee, So and Tang (2000) X X

Croson et al. (2004) X X Chiang et al. (2016) X Burbidge (1961) X Cachon (1999) X Heydari et al. (2009) X Wang et al. (2008) X

Jeong and Hong (2019) X

Sterman (1989) X

Disney and Towill (2003) X

Almeida et al. (2015) X

Yan and Katok (2005) X

Ha et al. (2010) X

Croson et al. (2014) X

Total literature sources

to each root cause 7 9 3 11 6 6 7 7 2 2 4 3 2

(24)

15

3.2.4 Type of failures that trigger the root cause

Failures can arise in numerous areas such as machine failures, process upsets, administrative delays, quality defects, and customer complaints and so on. According to Latino (2002) there are two simple ways to classify failures: sporadic and chronical. A sporadic failure is usually connected to a dramatic event that is not usually common, such as fires at the production plant or if the company lost a long-term contract to a competitor (Latino, 2002). Dramatical sporadic failures are most likely to happen a few times, but when the failure occur it tends to have a very dramatic impact and may be very costly for the company. Even though the sporadic events occur rarely, it is important to take these into

consideration due to the disturbance in the business and that the economic losses can be distributed over many years.

Chronical failures are the opposite of sporadic failures and are failures that occur repetitively. The impact of these failures is less than the impact for the sporadic ones, but due to the occurrence of these failures over time, they will cause a big impact on the total cost. Companies usually accept these

chronical failures, but it is important to investigate why they occur and their frequency. From the total cost perspective, it is of importance to analyse and investigate if there are patterns in upcoming chronical failures (Latino, 2002).

3.3 Mitigating the bullwhip effect

This section will initially present mitigation strategies of the bullwhip effect found in the theory, and thereafter present a summary of the stated mitigation strategies and which sources that mentioned them.

Knowledge and training

Negligence from Logistics Managers and decision-making problems arise from the lack of knowledge and training (Bhattacharya & Bandyopadhyay, 2010). To improve the knowledge about the bullwhip effect Yan and Katok (2005) suggested two types of practices: role specific training and system wide training. In the role-specific training, Logistics Managers attend to a role-play, in for example The Beer Game, where they can make decisions to solve supply chain problems. Through this type of training, Logistics

Managers learn how to make critical decisions and how to act in a situation that may cause a severe output (Bhattacharya & Bandyopadhyay, 2010). An early unintended error from a Logistics Manager could cause order swings and an uncontrollable outcome of the supply chain performance (Croson et al., 2014).

Normally, for managers in companies with large supply chains that are set to work at different

departments, a system-wide training will improve the understanding of supply chain structures and the interplay between nodes in the organization (Yan & Katok, 2005). A suggestion of system-wide training is when Logistics Managers act as central manager for a period in a practice game, and place orders for all nodes in the supply chain which directs to improve the understanding of underlying interrelationships (Bhattacharya and Bandyopadhyay, 2010; Yan and Katok, 2005). Senge and Sterman (1992) argue that simulation is a successful element to develop system thinking and encourage an overall understanding of the organization.

(25)

16

Trust, information sharing, and collaboration

Behavioural causes are causes directed to the negligence of decision making while the Logistics

Managers are placing orders. As mentioned above, behavioural causes are related to lack of education, training, fear of empty stock, and lack of coordination in decision making (Almeida et al., 2015). Tied up inventories is one of the main problems that the bullwhip effect has an impact on. The risk of tied up inventories can increase when the Logistics Managers have a lack of external trust with connected supply chain partners. To mitigate the problem Logistics Managers must be assured that their supply chain partners are making optimal decisions while placing orders (Almeida et al., 2015). If that is not the case, Logistics Managers need to compensate for the performance further down the supply chain where the order decisions are deficient. These compensations create order variations and to reduce these compensations, it is important to have trust in each supply chain member.

Collaborating and creating trustful relationships with partners can reduce the bullwhip effect. When the information is shared between each node and the inventories are structured and managed in a way to make the entire supply chain perform well, and not just a single node, will reduce the risk of a bullwhip effect. By applying VMI, just-in-time (JIT), and a centralized demand system may reduce the bullwhip effect (Shao et al., 2008). Externally VMI practices provide greater visibility in the supply chain which leads to better coordination between companies. The transparency will keep the inventory levels low and at the same time improve the availability (Almeida et al., 2015). Towill et al. (1992) described that a quicker and more accurate data capture and sharing the information between the nodes, will lead to a faster reaction to the bullwhip effect and may reduce the risk of a bullwhip effect.

Trust will have an impact on improved forecasts through information sharing and transparency between nodes in the supply chain. This type of information sharing mitigates demand amplification throughout the supply chain and hence reduce the bullwhip effect (Almeida et al., 2015).

Reducing uncertainty and reliance on forecasting

The best solution for mitigating the bullwhip effect is to create transparency throughout the supply chain, a centralized demand. With transparency throughout the supply chain, the companies have an opportunity to base their decisions on actual demand data instead of forecasts. Using real data will reduce many of the problems regarding the reliance on forecasting (Shao et al., 2008). Bullwhip effects are created when supply chain members base their demand input on their immediate downstream member when creating forecasts (Lee et al., 1997). The members downstream are at the same time basing their forecasts on their closest member downstream, and this makes multiple demand forecast updates. When forecasts are needed, using aggregated forecasts may reduce the risk of using inaccurate forecasts and consequently reducing the risk of a bullwhip effect (Shao et al., 2008).

Reducing variability and stabilizing the price

Price fluctuation is stated as a root cause of the bullwhip effect. Sales campaigns create variation in customer demand and by reducing the number of campaigns the risk of an occurrence of a bullwhip

(26)

17

effect can be reduced. Sales campaigns can increase short term sales and over time increase fluctuations at the customer site, and even larger fluctuations further up the supply chain. By using a consistent price, rather than using sales campaigns, is a way to reduce the retail forward buying (Lee et al., 1997). By eliminating these price fluctuations, the risk of dramatic shifts in demand that occur together with the sales campaigns will be eliminated. Therefore, minimizing price promotions can reduce the number of bullwhip effects (Shao et al., 2008).

Lead time reduction

Shao et al. (2008) state that reducing order lead times and information lead times, the bullwhip effect throughout the supply chain can be significantly reduced. Order lead times can be reduced by

implementing cross-docking, direct-shipping electronic data interchange, or advanced information systems (Shao et al., 2008).

Time compress principle: Every activity conducted in the supply chain should be made with the lowest possible time and the activities should be of value. This means that non-value-added time should be removed from the chain. This also covers the aspect of delivering on time (Geary et al., 2006). This principle can lead to a re-engineering of the business processes to shorten the lead time for both material and information (McCullen & Towill, 2001). The primary method for reducing the bullwhip and the lead time is a reduction of the order batching time (Devika et al., 2016).

Material flow

A smooth material flow is expressed as the inverse of the bullwhip effect by Geary et al. (2006), which has the intention of identifying and eliminating poor material flow and replace it with effective material handling. Towill and Childerhouse’s (2006) paper aligns with this statement where they discuss that the principles will have an impact on the decrease of the bullwhip effect in a supply chain. McCullen and Towill (2001) agree with the previous statement. Three material flow principles are suitable in a supply chain that is affected by the bullwhip effect (Towill and Childerhouse, 2006; Towill et al., 1992) which are control systems principle, information transparency principle, and echelon elimination principle (McCullen & Towill, 2001).

Control system principle: There is a need to select the most appropriate control system best suited to achieving user targets. To gain a supply chain with dynamic stability, there is an option of different decision support systems (McCullen & Towill, 2001). The user targets need to be taken into mind when selecting the most appropriate control system (Geary et al., 2006).

Information transparency principle: If constantly shared data without noise and bias should be accessed by all actors in the supply chain, it will remove the risk of double guessing, and that the same action is made by multiple Logistics Managers. If inventories, work-in-progress, flow rates, and orders are visible throughout the supply chain, suitable decision systems can be used to optimize the material flow. Echelon elimination principle: The aim is to have the number of echelons as low as possible in the chain. It is not enough to have the stock optimized, but also that the stock is in the right place at the time it is needed (Geary et al., 2006).

(27)

18

Figure 7 – A summary of the strategies to mitigate the bullwhip effect.

A summary of all mitigation strategies is presented in Figure 7. To improve the reliability of the study, each reference used to state the mitigations strategies is presented in

(28)

19 Table 2 and marked with the corresponding author.

(29)

20

Table 2 - Compiled mitigation strategies correlated with authors.

3.4 Bullwhip effect and availability

A simple way to measure a bullwhip effect is to compare the variance in demand to orders and net stock. Disney and Labrecht (2008) and Chaing et al. (2016) formulate the bullwhip effect as how many times the variance of orders is greater than the variance of demand. The variance ratio is presented in equation 3. 𝐵𝑢𝑙𝑙𝑤ℎ𝑖𝑝 = 𝜎 2 𝑂𝑟𝑑𝑒𝑟𝑠 𝜎2 𝐷𝑒𝑚𝑎𝑛𝑑 = 𝑉𝑎𝑟(𝑂𝑟𝑑𝑒𝑟𝑠) 𝑉𝑎𝑟(𝐷𝑒𝑚𝑎𝑛𝑑) (3)

When the variance of orders is equal to the variance of demand, there is no amplification. A variance ratio larger than 1 indicates that the bullwhip is ongoing, see Figure 8, whereas a bullwhip smaller than 1

(30)

21

means that the orders are smoothed as seen in Figure 9. When smoothed orders occur, means that orders are less variable compared to the demand pattern and the bullwhip effect is dampening.

Figure 8 - An illustration of an ongoing bullwhip effect.

(31)

22

4 Study specification

This chapter clarifies how the bullwhip effect is defined in this study, and states Research Questions and Sub-questions to fulfil the purpose of the study.

The purpose of the study is presented below, and the studied system is presented in Figure 10. The purpose of this study is to identify events in the Volvo Group Service Market Logistics’ supply chain where the bullwhip effect has occurred, its root causes, and how to reduce or eliminate the bullwhip effects.

The purpose can be divided into three main areas: identifying bullwhip events, root causes of the bullwhip events, and investigation of how to reduce or eliminate the bullwhip effect. Each of the main areas will be the topic of one of the three Research Questions, where two Research Questions have associated Sub-questions.

Figure 10 - The studied system.

4.1 The definition of the bullwhip effect

Figure 11 displays the characteristics of the bullwhip effect. A small increase in sales leads to an increase in orders from the Dealers, which leads to an even bigger increase in orders from the RDC, which results in that the CDC places an even larger order. The fictional example in next section fulfils the requirements of a bullwhip event.

(32)

23

Figure 11 – Displaying an increase in order variance in each node of the supply chain.

The definition of a bullwhip effect in this report will align with an increase in order quantity in each node, based on equation 1 and 2. The increase in order quantity from the Dealers to the RDC, and the RDC to the CDC, does not need to be within the same month or week as the increase in sales, due to the lead times. To clarify how the definition of a bullwhip effect was composed, equation 1 and 2 will be presented, followed by the definition of the bullwhip effect for this study in equation 4.

𝑉𝑎𝑟𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛< 𝑉𝑎𝑟𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛+1< 𝑉𝑎𝑟𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛+2 (1)

𝐴𝑚𝑝𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛 < 𝐴𝑚𝑝𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛+1 < 𝐴𝑚𝑝𝐷𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑛𝑜𝑑𝑒 𝑛+2 (2)

The Sales quantity considers actual sales, which means the quantity of a spare part that the customer has ordered and received. The Order quantities between the Dealer to the RDC and the RDC to the CDC consider placed orders. It means what the Dealer wants from the RDC and what the RDC wants from the CDC and does not consider what has been delivered.

(33)

24

4.2 A clarification of a bullwhip effect

To explain a bullwhip effect, a fictional example has been created which can be seen below.

The fictional example: The supply chain has three echelon levels: the CDC, the RDC, and the Dealers. In this example, there are several nodes in each echelon level as seen in Figure 12.

Figure 12 – Illustration of the fictional example’s supply chain.

It starts with the customers placing orders to the Dealers. In this example, each customer places an order of 100 parts each to their associated Dealer. This is more than the expected demand of 90 parts per customer. Therefore, the Dealers think that the demand is on its rise and decide to increase their orders to the RDCs to make sure they will be able to cover the rising demand. The Dealers decide that they will order 105 parts each from the RDCs. The RDCs notice that the demand seems to be on its rise and therefore they want to make sure they will be able to deliver accordingly to the rising demand. They then increase their orders to the CDC and place orders of 250 parts each. The CDC now notice that there is a big increase in the demand and believes that the demand will increase even more. Therefore, they order many parts from their suppliers. As seen in Figure 13, the correctness of the actual demand and orders differs more and more in each echelon level of the supply chain. Echelon level 3 – the Dealers – have a correctness of 95%. In the next echelon level – the RDCs – the correctness has decreased to 80%, and in echelon level 1, the correctness of the demand compared to the orders is down to 73%. In each echelon level – from 3 to 1 – the demand and orders differ more and more. This is an example of the bullwhip effect in a three-echelon supply chain.

(34)

25

Figure 13 - Example of orders in the fictional supply chain.

4.3 The identification and selection of bullwhip events

Oskarsson et al. (2013) compose that it is difficult to coordinate logistical activities at multiple companies and different echelon levels in a cohesive way. When each firm runs its business

independently, communication deficiencies will arise which cause increased supply chain costs and decreased availability (Forrester, 1961). To be able to identify the root causes of the bullwhip effect, it is of importance to identify events that have been affected by the bullwhip effect and to identify which ones that are suitable to study. Therefore, the first Research Question has been formulated as follows:

Research Question 1: How can previous bullwhip events be identified in the Volvo SML supply chain, and which events are best suited for further study?

According to Bhattacharya and Bandyopadhyay (2010), the bullwhip effect has different consequences such as lower availability and higher supply chain costs, which is of course something that a supply chain

(35)

26

wants to avoid. Volvo has suspected that the bullwhip effect has occurred in their SML supply chain before, and to establish that the bullwhip effects have occurred, previous events must be found to investigate the issue. Furthermore, the events must fulfil this study´s criteria of a bullwhip effect to be classified as bullwhip events. To identify whether these events have been affected by normal variation or the bullwhip effect, Jeong and Hong’s (2019) explanation of a bullwhip effect will be considered. To identify bullwhip events in the Volvo SML supply chain and ensure that the events fulfil the criteria of increased order quantity in each node according to equation 4, Sub-question 1.1 has been formulated as:

Sub-question 1.1: In what events has the bullwhip effect previously occurred?

To be able to analyse a big amount of information, a data reduction is needed, where it is important to make qualitative decisions to reduce the risk to study unimportant events. The study can only include a manageable amount of data and to avoid having too many events that passed the previous criteria, another sorting must be done. Consequently, the further study will focus on the most suitable representable events to gain the most relevant results. Thus, to find the most appropriate events, another reduction must be done based on reasonable parameters. Sub-question 1.2 will therefore be stated as:

Sub-question 1.2: Which bullwhip events are appropriate to investigate?

4.4 The causes of the bullwhip effect

To understand an undesirable outcome, it is of importance to understand why it happened. To find the actual root cause of the issue, it is needed to investigate what information was used to make the poor decision (Latino, 2002). An important part of the study is to identify the root causes of the bullwhip effect in the Volvo SML supply chain and to which nodes the root causes are associated. According to Sterman (2005), the occurrence of the bullwhip effect is either due to operational or behavioural causes, and several authors present different root causes within the two categories. To identify the causes of the bullwhip effect in the Volvo SML supply chain, Research Question 2 is stated as:

Research Question 2: What are the causes of the bullwhip effect in the Volvo SML supply chain?

Disney and Labrecht (2008) and Sterman (2005) emphasize that there are both operational and behavioural root causes of the bullwhip phenomenon. To get a deeper understanding of what has happened in previously identified bullwhip events in the Volvo SML supply chain, and what has caused the effect, both the operational and behavioural causes must be investigated. In the previous research question, the outcome was previous bullwhip events from the Volvo SML supply chain, which all have

(36)

27

root causes of different kinds. Sub-question 2.1 will therefore identify the root causes to the identified bullwhip events, hence Sub-question 2.1 is formulated as follows:

Sub-question 2.1: What are the root causes of each bullwhip event?

An investigation in which part of the supply chain that influence the rise of the bullwhip effect should be done. Both behavioural and operational root causes can be associated with a node in the supply chain, meaning where the bullwhip effect is triggered to occur. However, to fulfil the criteria of the bullwhip effect, the root cause must induce a bullwhip effect that has an impact on all included nodes of the supply chain. Even if the root cause arises in an upstream node – the CDC or the RDC – it must in one way or another influence the Dealer node to start the bullwhip effect to fulfil the criteria. With this said, the root cause can be associated to an upstream node but must have an impact on all nodes to be classified as a bullwhip effect. To connect each root cause to the origin in the supply chain, Sub-question 2.2 is stated as:

Sub-question 2.2: Which root causes are associated with each node of the Volvo SML supply chain?

4.5 Reduction or elimination of the bullwhip effect

After identifying the root causes of the bullwhip events, the next step is to investigate further how the bullwhip effect can be reduced or eliminated as an aid for Volvo to use. A bullwhip effect may cause insufficient availability, and the total supply chain cost could potentially increase (Bhattacharya & Bandyopadhyay, 2010). If the bullwhip effect is reduced or eliminated, the performance of the supply chain can be improved. In the literature, there are principles and strategies to mitigate the bullwhip effect. The principles and strategies are connected to the different root causes and must accordingly be applied to reduce or eliminate bullwhip effects in the future. To present how the bullwhip effect can be reduced or eliminated, Research Question 3 is stated as follows:

Research Question 3: How can the bullwhip effect be reduced or eliminated in the Volvo SML supply chain?

4.6 Summary of questions

The purpose of the study, the Research Questions, and Sub-questions stated above are summarized in Figure 14 below.

(37)

28

(38)

29

5 Methodology

This chapter describes the methodology used to study the bullwhip phenomenon. Firstly, a general approach of the study is presented, then the literature collection is described, and thereafter how the study was conducted. This is followed by a discussion of the study credibility.

5.1 Phases of the study

This study was divided into five phases: the introduction phase, preparation phase, data collection phase, analysis phase, and the conclusion phase. The phases are based on Patel & Davidsson (2011) and Lekvall & Wahlbin’s (2007) stated study approaches. The five phases are described below and an overview of the study is presented in Figure 15. A more detailed description of how the study was conducted can be seen in section 5.3.

Figure 15 - The phases of the study.

5.1.1 Introduction phase

The introduction phase was conducted for the authors to get introduced to the bullwhip effect, to Volvo, and to understand what the focus of the study was going to be, hence formulating a purpose. Together with the supervisors’ support, directives and delimitations were formulated which made it possible to present the scope of the study. The case situation was made to collect information about Volvo SML and to visualize a simplified supply chain. The simplified Volvo SML supply chain was used to understand the information and material flow, and at the same time visualize the scope of the study. The initial literature was collected to get an insight of what information that was of importance for the project, with the focus on the bullwhip effect. However, the literature covered a large wide of areas to get a broad and non-biased understanding before executing the study.

5.1.2 Preparation phase

The second part of the study was the preparation phase. The preparation phase consisted mainly of the study specification and a methodology. The study specification was based on the theoretical framework and included Research Questions founded out of the aim of the study and the formulated problem. With the presented theories in the theoretical framework, each Research Question was split into multiple Sub-questions. Each Research Question connected the Sub-questions with the purpose of the study. The

Causes of the bullwhip effect : A study of the bullwhip effect in the Volvo Group Service Market Logistics’ supply chain