Abstract:
Purpose: The aim of this thesis is to find an organized approach to cleaning up excess
inventory, particularly at the case company X. The practical problem is that most retailers and distributors struggle to have the right stock-keeping unit (SKU) at the right place, in the right quantity, and at the right time to meet the demands of their customers. In most warehouses there are thousands of SKUs listed as inventory stock. It is important to have a strategy for managing this inventory.
Methodology: The focus of this thesis is on the SKUs that occupy most of the inventory
space. This criterion is used to classify inventory with the aim of cleaning up the inventory system. It is impossible to take all of the SKUs into account; to manage inventory efficiently, one should use the available resources in the best way by focusing on the most important inventory SKUs. The case company’s data was analyzed in Excel, and eight approaches to determining the right SKUs to consider were identified. A Pareto distribution of 2/30 was used for 49 of the 2,963 SKUs. These 49 SKUs have been dealt with by applying the Q,r inventory control model to reduce inventory costs while maintaining a tolerable service level.
Findings: This study contributes to the literature by exploring different approaches to
Preface
This thesis has been written as part of the requirements for graduation from the master’s in Technology and Operations Management program at the University of Groningen. I would like to express my gratitude to the people who offered their support and feedback to me during this thesis project.
First, I would like to thank my supervisor, Dr. van Foreest, for his support and help during this project. We had many fruitful discussions that helped me write this thesis. His feedback on my drafts was very useful. Without his support, it would not have been possible to complete this project. In addition, I would like to thank my co-assessor, Professor Ir. Wortmann, for his feedback during my presentation.
1. Introduction
Inventory management is one of the key elements in supply chain management. Since there can be substantial costs involved in staying above and below the optimal inventory range, careful inventory management can make a huge difference in the profitability of a business (Abdelaziz and Mejri, 2013). Therefore, a small savings in inventory will yield a crucial benefit for the organization. Through proper inventory control techniques, the risk of stock-out and overstock situations can be minimized (Biswas et al., 2017). Inventory is defined as the stock of physical goods held at a specific location at a specific time (Seetharama et al., 2003).
However, the problem in practice is that most retailers and distributors struggle to have the right stock-keeping unit (SKU) at the right place, in the right quantity, and at the right time to meet the demands of their customers (Christopher and Ryals, 2014). An SKU is an individually identifiable item stored in a specific location. When a company takes inventory of its stock, it counts the number of items it has of each SKU (Sawaya and Gianque, 1986). Increasing product variety creates operational challenges for companies and results in higher inventory levels. A large number of SKUs deteriorates decision quality and leads to inventory problems (Wan and Sanders, 2017). A logical inventory classification is necessary for managers to plan efficiently and control the items (Chen, 2012).
A well-known inventory classification technique is ABC analysis, which is based on the Pareto principle; it classifies SKUs into three groups: A, very important; B, moderately important; and C, relatively unimportant. Traditional ABC analysis is based on a single criterion, such as annual dollar usage. However, in practice, the various and numerous customer demands lead to an increasing variety of inventory SKUs, which may not be homogeneous; thus, the main difference among them may also lie in other criteria (Chen, 2012). It has been generally recognized that lead time, average unit cost, scarcity, and demand distribution, for example, are also important criteria in deciding the importance of a SKU (Ramanathan, 2006).
during this study at the case company. The inventory classification criterion used in this paper is: amount of inventory space per SKU. To better control inventory, it is important to look closely at the SKUs that occupy most of the inventory space. These SKUs have the highest holding costs because they occupy the most pallets or space.
Case company X is a wholesaler that sells carpets. In the carpet industry, it is common for companies to coordinate their activities through a push system. This means that most companies use a make to stock policy (MTS), holding inventory at the end of the supply chain. The order portfolio of company X consists of about 3,000 different SKUs. These 3,000 different SKUs are held as inventory in the warehouse. Currently, company X has no idea how to fill this warehouse or how many SKUs they should have in stock in order to minimize inventory costs. Relatively few SKUs occupy most of the inventory. To significantly reduce inventory costs, we focus on the SKUs that occupy most of the inventory space.
The aim of this thesis is to find an organized approach to cleaning up the inventory system so that company X can have better control over the inventory. This paper contributes to the literature on management practice by providing companies with an organized approach to dealing with a large number of SKUs and excess inventory. The Q,r inventory control model has been tested for the SKUs that occupy most of the inventory space with the aim of reducing inventory costs. This policy also leads to certain inventory costs; in this paper, we compare these costs with the company’s current costs. Further, it is important for company X to deliver a tolerable level of service to maintain customer satisfaction. Company X requires a service level of at least 95%. The main research question of this thesis is as follows:
How can excess inventory in the carpet industry be cleaned up to minimize inventory costs while providing a tolerable level of service?
2. Theoretical background
This chapter provides the theoretical background of inventory management, inventory classification, inventory costs, the Q,r inventory control model, and the different service levels.
The role of inventory management is to ensure that stocks of raw material or other supplies, i.e. finished goods, are kept at levels that provide maximum service levels at minimal cost (Sternman and Dogan, 2015). The inventory in a given organization can contain a large number of SKUs. Increasingly, product variety is creating operational challenges for companies and resulting in higher inventory levels. A large number of SKUs deteriorates decision quality and leads to inventory problems (Wan and Sanders, 2017). A logical inventory classification system is necessary for managers to efficiently plan and control SKUs. Inventory classification directly affects a company’s ability to satisfy customer orders with its inventory. Poor inventory classification results in inventory buildup of those SKUs that are not needed to meet customer demand. On the other hand, it can also lead to inventory shortages of SKUs when they are needed the most.
2.1 Inventory classification
Further, ABC analysis is based on Pareto analysis, which means that 20% of the SKUs are A-items. These A-items are the most important SKUs in terms of annual dollar usage of the company and should be intensively checked. However, in practice, 20% leads to too many A-items to reasonably check. Imagine that a company has 100,000 SKUs; this means that there are 20,000 A-items that must be intensively checked. In practice, this is not possible (Durlinger, 2009). Additionally, large firms usually have to manage huge inventories consisting of thousands of SKUs with limited resources, like money and time. To manage inventory efficiently, it is necessary to use the available resources in the best way by focusing on the most important inventory SKUs (Hatefi, Torabi and Bagheri, 2014). Furthermore, traditional distributions have changed in ways that are disruptive to different things. The Pareto 80/20 rule has decayed into an empirical anachronism. An anachronism is something that does not quite fit in its time. The Pareto 80/20 rule states that 80% of effects come from 20% of causes. However, firms are increasingly seeing Pareto proportions closer to 1/25, 2/30, 5/50, 10/90. This means that causes are having a greater impact on the effects. Pareto’s ‘vital view’ has thus become a ‘vital viewer’ (Schrage, 2017), which means that only a small part of the SKUs are responsible for a larger portion of sales, inventory costs, or other criteria.
2.2 Inventory costs
2.3 Q,r inventory control model
In this thesis, the Q,r inventory control model is tested. The two problems that company X has is when to place an order and how much to order. The Q,r inventory control model focuses on these two decision variables, which are under the control of the manager. These decisions variables are considered when aiming to minimize total inventory costs (Hadley and Whitin, 1963).
Depending on how costs and customer service are represented, Q and r interact in terms of their effects on inventory, production or order frequency, and customer service. However, it is important to recognize that the two parameters generate two fundamentally different kinds of inventory. The replenishment quantity Q affects cycle stock. Cycle stock represents stock held between ordering cycles. The reorder point r affects the safety stock. Safety stock is inventory that protects against stock-outs due to fluctuations in demand. Given these definitions, the entire inventory held in the EOQ model is the cycle stock, and the entire inventory held in the base stock model is the safety stock. The EOQ model is an inventory model that calculated the order quantity that minimizes the total ordering costs and holding costs. The base stock model is an inventory model where inventory is replenished one unit at a time as random demands occur, so that the only issue is to determine the reorder point. In a sense, the Q,r model represents the synthesis of these two models (Hopp and Spearman, 2011).
The total cost is a function of the ordering quantity Q and reorder point r. Different values of Q and r result in different costs (Kao and Hsu, 2002). Figure 1 depicts a representation of the Q,r model. Q and r are the decision variables. A stock-out could occur when the reorder point is relatively low or when demand during lead time is suddenly very high.
2.4 Service level
3. Methodology
This section describes the methodology that was used to answer the research question. First, the case company and research design are described. Next, the data collection method is explained. Then, the reliability, validation, and verification processes are explained. In Section 3.5, the Q,r inventory control model is explained. Finally, in Section 3.6, a sensitivity analysis is described.
3.1 Case description
The case company in this study is company X, a carpet wholesaler. Company X has a large product portfolio, and they distinguish themselves through customization and good customer service. Company X owns three subsidiaries: company Y, company Z and company A; an overview of the supply chain is provided in Figure 2. Company Y is the manufacturer of the raw materials. Company Z and company A are manufacturers of the carpets. The slogan of company X is “Flooring is our product, service our business.” The company’s unique selling points are the quality of its carpets and fast delivery. Company X has indicated that it currently has to deal with huge inventories in the warehouse. Company X has a warehouse that contains 16,000 pallet spaces. If the company implements appropriate inventory policies for their SKUs, they will likely achieve lower inventory costs and less inventory.
system (Hopp and Spearman, 2011). Many firms in the carpet industry are part of a group and are subsidiaries of a wholesaler. The tension in this chain is that none of the organizations want to hold large inventories, but they also do not want to take on the risks involved with maintaining small inventories.
3.2 Research design
To investigate this problem, numerical analyses were conducted. In this case study, a quantitative analysis was conducted using Excel 2016. First, a numerical analysis, which is laid out in Chapter 4, was conducted to decide which SKUs should be taken into consideration. An inventory policy was tested, and then the inventory costs and service level were calculated. The results with respect to the inventory costs and service level are provided in Chapter 5.
3.3 Data collection
The data is divided into three categories, namely: A, B, and C –data. A-data are data that were derived directly from the company. B-data were not immediately available but collectable, and C-data were not collectable and available at the company. So, if C-data is needed, it should be assumed (Robinson, 2014). All the data are from 2017, and contain information about inventories, demand, and costs. The data were used for the numerical analysis. All the data were tested with the Q,r inventory control model and were compared with the current situation with respect to inventory costs. In this thesis, the data concerning inventories, demand, and costs are category A-data because they were directly available. The following A-data were used in the model:
The following C-data were not available. Therefore, the following assumptions and simplifications were used for the model: - Capacity of one pallet is 650 items. The A and C –data were used for the numerical analyses in Chapter 4 and for the results in Chapter 5. 3.4 Reliability, validation, and verification
- Holding cost per unit per year – hc - Shortage cost per unit – p Decisions variables: - Order quantity – Q - Reorder point – R
The objective of the model was to minimize total expected annual ordering, shortage, and holding costs. The outputs of the model were the optimal policies for Q and r, minimal total annual expected ordering, shortage, and holding costs, and service level of the optimal policy. The model calculated how much the company should order and at which reorder point. It was about a single inventory item at a single price under uncertain lead time and demand conditions (Cobb, 2013). Lead time demand (X) is defined as the sum of distributed periodic demand values (Cobb, 2013). The standard deviation and mean of lead time demand were calculated as follows:
(1) μX=μL·μAD
(2) σX= μL·σA2D+μ2AD·σL2.
The decision variables order quantity and reorder point are under the control of the manager. These decisions variables were chosen to minimize total inventory costs, which were calculated as follows (Hadley and Whitin, 1963):
(3)
(4) Figure 3 provides an overview of the Q,r inventory control model calculation in Excel. This figure consists of values of the number one SKU for the average inventory per year at company X. The expected shortage per cycle is the shortage times the probability of a shortage. The expected shortage per cycle was calculated in cell G1. The shortage and probability of a shortage are displayed in columns F and G. The shortage was calculated by taking the maximum of “0, lead time demand – reorder point”. If the reorder point is greater than the lead time demand, there are no shortages. If the reorder point is smaller than the lead time demand, a positive value is shown. The probability of a shortage was calculated using this formula: =norm.dist. (lead time demand + 0,5, μX, σX, 1) minus norm.dist. (lead time demand - 0,5, μX, σX, 1).
Further, the area under the normal curve was 99.73% with three standard deviations of the mean. In this Q,r model, this is the ‘effective maximum’ for lead time demand. In column E, the lead time demand values are presented. The formula in cell G1 takes the values between the mean and the effective maximum. In this case, the mean for lead time is 5,114 items and the standard deviation is 3,269 items, thus the effective maximum is 14,921 items.
The expected shortage is 3.04 items per cycle and there are 12,63 order cycles per year. Therefore, the shortage costs are €30.03 because the shortage cost per unit is €0.781. Ordering costs are €1,579.26 because there are 12.63 order cycles per year and the ordering costs per order are €125.00. Finally, the holding costs are €3,198.00, because the holding costs per item is €0.12, and the average inventory on stock is 26,650 items. These three costs components combined total €4,807.29 in inventory costs per year.
3.5.1 Service level
- Service level ≥ 95%. It is important for company X to have a tolerable level of customer service. Thus, the service level should be at least 95%.
3.6 Sensitivity analyses
Sensitivity analysis is fundamentally important for risk analysts, especially in the presence of difficult computational models with uncertain inputs. The reason to perform a sensitivity analysis is to rank the relative importance of the different input variables, to uncover relations between input variables and output variables in difficult models, to identify areas of investigation and refinement in the model, and to perform reasonableness checks and quality assurance (Iman and Helton, 1998; Frey and Patil, 2002; Saltelli, 2002; Helton and Davis, 2002).
4. Numerical analysis in Excel
In this section the numerical analysis in Excel is explained. First, the selected SKUs that should be considered are described. In Section 4.2, the different approaches are discussed. In this subsection, the researcher explains the selection process and how to clean up inventory. Thereafter, the Pareto distribution is explained and an analysis of the selected SKUs is described.
4.1 Selected SKUs to consider
Large firms usually deal with inventories that may contain thousands of SKUs, while the required resources to manage them, such as money and time, are often limited. To achieve efficient inventory management, the reasonable action is to try to use the available resources in the best way by focusing on the most important inventory SKUs (Hatefi, Torabi and Bagheri, 2014). Company X has 2,963 SKUs, and it seemed impossible to take all these SKUs into consideration. However, the first step to cleaning up the inventory was to identify the amount of inventory space occupied per SKU in the inventory. Every SKU that is held as inventory has a minimum of one pallet place. The capacity of one pallet is 650 carpets. In order to get the amount of pallet spaces per SKU, the average inventory per SKU was divided by the capacity of one pallet. Further, SKUs that occupy at most one pallet space could not be reduced in terms of holding cost because there are fixed holding costs per pallet space. It does not matter if there are 200 or 600 items on a pallet with respect to holding costs. Thus, in terms of holding costs, it is not possible to reduce the costs of an SKU that occupies one pallet space. However, reducing the holding costs of one pallet space is more or less an end-of-life issue; the SKU must be completely disposed of because the inventory should be reduced to zero. However, this is not considered in this paper.
total number of SKUs with the total number of pallets. Figure 4 provides a distribution overview that occupies two ore more pallets places. The green bars indicate the total number of pallets, and the values are provided on the left y-axis. The red line indicates the total number of SKUs, and values are displayed in the figure and on the right y-axis. For instance, there are 231 SKUs that occupies two pallets places, and one SKU that occupies 224 pallets places.
Figure 4 Distribution overview that occupies two ore more pallets places
which would not lead to a large reduction in inventory costs. Figure 5 provides an overview of the cumulative number of pallets in percentage form. For instance, about 300 SKUs occupies 50% of the pallets. It is obvious that the shape of the line is decreasing rise. This means that the first SKUs occupy the most space in terms of number of pallets.
Figure 5 Cumulative number of pallets in percentage
There are different approaches that could work for cleaning up the inventory. It is possible to consider the SKUs that occupy 30 pallets or more, the SKUs that occupy five pallets or more, the first 10 SKUs that occupy the most pallets, or the SKUs indicated by a Pareto distribution. The researcher considered these eight different approaches. Table 1 provides an overview of the different approaches. Each approach led to a certain number of SKUs (see column 2). In column 3, the percentage of total SKUs is provided. The total number of SKUs was 2,963. Thus, 10 SKUs amount to 0.34% of the total number of SKUs. The fourth column provides the total number of pallets. Thus, the first approach led to 10 SKUs occupying 1,003 pallets. This is 18.2% of the total number of occupied pallets (see column 5). The total number of pallets was 5,509.
Approach to selecting
SKUs to consider Number of SKUs Percentage of total SKUs Total number of pallets Percentage of total pallets places
1. First 10 SKUs 10 items 0.34% 1,003 pallets 18.2% 2. Items occupy 30 or more
pallet places 15 items 0.51% 1,163 pallets 21.2% 3. Pareto distribution, 1/25 25 items 0.84% 1,380 pallets 25.0% 4. Pareto distributions
2/30 49 items 1.65% 1,650 pallets 30.0%
5. SKUs occupy five or more
pallet places 140 items 4.72% 2,208 pallets 40.0% 6. Pareto distribution 5/50 304 items 10.26% 2,755 pallets 50.0% 7. Items occupy two or
more pallet places 526 items 17.76% 3,199 pallets 58.0% 8. Pareto distribution
10/90 2,285 items 77.12% 4,958 pallets 90.0%
It is clear that Approach 8 would lead to too many SKUs. 77.12% of the total SKUs occupy 90% of the total pallets places. This concentration is much lower in comparison with the Pareto distribution of 10/90, where 10% of the total SKUs should occupy 90% of the total pallets places. Approach 7 would also lead to too many SKUs according to Hatefi, Torabi, and Bagheri (2014) and Durlinger (2009). According to these two groups of researcher, 526 SKUs is too many to consider because the company should focus on the most important inventory SKUs.
Approach 6 is the Pareto distribution 5/50. However, the concentration at the case company is much lower because 10.26% of the total number of SKUs occupies 50% of the total pallet places. This means that Approach 6 would lead to too many SKUs according to Hatefi, Torabi, and Bagheri (2014) and Durlinger (2009).
Approach 5 would also lead to too many SKUs, according to Hatefi, Torabi, and Bagheri (2014) and Durlinger (2009), because 140 SKUs would occupy five or more pallet places.
The first four approaches would lead to a number of SKUs that would be realistic to intensively consider. Approaches 1 and 2 could be good options for cleaning up inventory. However, 10 and 13 SKUs would occupy 18.2% and 21.2% of total pallet places, respectively, and this percentage is quite low.
Approach 3 and 4 are left to choose for selected SKUs to consider. The researcher chose Approach 4 because this number of SKUs is best suited for the task of cleaning up excess inventory. Besides, the concentration at the case company is much higher than the Pareto distribution 2/30 because only 1.65% of the SKUs occupy 30% of the total pallet places. Thus, the Pareto distribution 2/30 approach was chosen.
4.3 Pareto distribution
4.4 Analysis of considered SKUs
5. Results and interpretation
In this section, the results are presented and interpreted. The results consist of the Q,r inventory policy that was tested for the different SKUs. The 49 SKUs selected for consideration were presented in Section 4. In this study, the research applied a Pareto proportion of 2/30. This means that about 2% of the SKUs represented 30% of the inventory space. These 2% were intensively investigated in order to reduce their inventory costs. These 49 SKUs are distributed as follows: company Z manufactures 32 SKUs, external suppliers in the Netherlands manufacture 13 SKUs, and company A in India manufactures four SKUs.
5.1 Inventory costs
First of all, the current inventory costs for 2017 were compared with the inventory costs calculated through the Q,r inventory control model. This Q,r inventory control model is considered the new situation. As stated in the theoretical background section, inventory costs consist of holding costs, ordering costs, and shortage costs. These three cost components are explained separately below.
5.1.1 Holding costs
Figure 8 Ordering costs per year per SKU
5.1.3 Shortage costs
5.1.4 Difference inventory costs
In the sections above, the inventory costs were explained per part. In this section, the total inventory costs of the new situation are compared with those of the current situation. The current inventory costs are €473,200 because ordering costs are €344,500 and holding costs are €128,700. The inventory costs in the new situation would be €79,818.98 because ordering costs are €31,974.08, holding costs are €47,268, and shortage costs are €576.90. Figure 10 presents an overview of the inventory costs per year per SKU. The blue bars represent the inventory costs in the current situation, and the red bars represent the new situation. There is a reduction of 83.1% in inventory costs, or a cost reduction of €393,381.02, which are huge savings in inventory costs. Figure 10 Inventory costs per year per SKU The following causes are behind these huge savings. In the current situation, company X orders every week for every SKU from their subsidiary company Z and from other external suppliers in the Netherlands. In the new situation, as developed using the Q,r inventory control model, the average frequency of ordering is 4.05. In the current situation, the ordering costs are 73% of the total inventory costs, and in the new situation, the ordering costs are 40% of the total inventory costs. This means that there are relatively large savings in ordering costs in comparison with the other two cost components.
policy), the optimal average inventory has been calculated for each SKU. This means that the model considered the capacity of one pallet.
5.2 Service level
In this section, the current service level is compared with the service level that was calculated with the Q,r inventory control model. The current service level is almost 100%, but it has not been accurately recorded in the last several years. One of the unique selling points of company X is that delivery is fast and on time. It is very important in the carpet sector to have high delivery reliability. The average service level in the new situation would be 98.67%. Figure 11 provides an overview of the service level per SKU in the new situation.
Figure 11 Service in the new situation
The sensitivity analyses of the holding costs determined that they have the greatest impact on total inventory costs. If the holding costs increase by 10%, the inventory costs increase by 5.37%. This means that the holding costs are an important part of the inventory costs and that this cost component should be given more attention in comparison with the other two costs components. Appendix B provides an overview of the three inventory costs elements for the new situation, including the sensitivity analyses with a 10% increase in holding costs applied. It is obvious that the ordering costs increased the most. The reason for this is that when the holding costs are increased by 10% per item, there are more orders per year because the model calculates the minimum inventory costs. The sensitivity analyses determined that ordering costs have the second greatest impact on inventory costs. When the ordering costs were increased by 10%, the inventory costs increased by 4.69%. It is important for company X to give enough attention to this cost component because the transporting of the carpets is outsourced, and company X does not have total control over these costs. Appendix C provides an overview of the three inventory costs elements for the new situation, including the sensitivity analyses with a 10% increase in ordering costs applied. It is obvious that if the ordering costs increase by 10%, the holding costs will also increase because the model calculates the minimum inventory costs. The model demonstrates that the company should have more inventory in stock and order less often.
6. Discussion
This section presents a discussion of the results and validity of this research. In Section 6.1, the limitations are provided.
and that have excess inventory. The approach and inventory control model used here could be used to reduce inventory costs at other companies while maintaining a tolerable service level.
6.1 Limitations
This section discusses the limitations of this research. The first limitation is that the analyses were applied to one case company in the carpet industry; therefore, the results are not immediately generalizable to all other industries. However, this research is useful for companies that must deal with many different SKUs, and high inventory per SKU.
The second limitation is that the case company had not tracked current shortage costs and service level at the time of the research. Thus, these two components were estimated for company X. If these two components had been tracked, the results would be more reliable.
7. Conclusion
This section presents the conclusion of the thesis, recommendations for the case company, and suggestions for future research.
As stated in the introduction, the research question is: How can excess inventory be
cleaned up in the carpet industry to minimize inventory costs while maintaining a tolerable service level? Specifically, the aim of this thesis was to find an organized
approach to cleaning up excess inventory at the case company X. In the sections containing the numerical analysis and results, this question has been answered.
The problem at the case company was that they had too much inventory for most of the SKUs. Company X had 2,963 SKUs in their product portfolio, and this is too many SKUs to intensively consider. The inventory classification ‘inventory space per SKU’ was used to classify and order the SKUs. Then, eight different approaches were taken to help the researcher decide which SKUs should be considered because considering all the SKUs was not possible. The Pareto distribution 2/30 approach was chosen, and 49 SKUs were considered as a result of this approach. These 49 SKUs were tested with the Q,r inventory control model in Excel. The results indicate that there would be a cost reduction of 83.1% compared to the current situation and a tolerable service level of 98.7%. This means that company X could save €391,383.02 per year.
A sensitivity analysis was also conducted to investigate which of the three costs components of inventory costs are dominant. The results indicate that the holding costs have the greatest impact on the total inventory costs. When the holding costs were increased by 10%, the inventory costs increased on average by 5.37%. The sensitivity analysis of the ordering costs determined that ordering costs have the second greatest impact on the inventory costs. An increase of 10% in ordering costs results in an inventory cost increase of 4.69%.
These results are useful for companies with a high number of SKUs that have to deal with excess inventory. These companies could use this organized approach to reduce inventory costs.
7.1 Recommendations for case company X
(49 SKUs) is way too high. The recommendation for company X is to focus on these 49 SKUs because this group of SKUs has a significant amount of excess inventory. Company X should use the Q,r inventory control model for this group of SKUs. This will lead to lower inventory costs with respect to service level. The holding and ordering costs will decrease significantly for the 49 SKUs. The results also indicate that company X should order less often in comparison with the current situation. This would lead to lower holding costs because these 49 SKUs would occupy only 606 pallet places in this new situation, and they currently occupy 1,650 pallet places. Another recommendation for company X is not to fill up the warehouse completely. Company X can use the Q,r inventory control model to avoid too much stock. An option for company X is to rent out a part of the warehouse. This would lead to more profit for the organization.
Furthermore, in this study, the tolerable service level was 98.7%. The current service level is almost 100% because the company has excess inventory and can immediately deliver using stock on hand within a replenishment cycle. A recommendation for company X is to calculate the trade-off between shortage costs and holding costs. If the shortage costs are very high, it might be preferable to have a service level closer to 100%. This could be calculated using the Q,r inventory control model.
7.2 Future research
There are several suggestions for future research. In this study, the focus was on the first 49 SKUs, and inventory space per SKU was used as the inventory classification. A suggestion for future research would be to investigate perhaps the first 5% or 10% of the SKUs instead of just 2%, as was the case here. However, this would depend on how much SKUs the company has and what the distribution of the SKUs is in terms of inventory space.
A second suggestion for future research is to calculate the rent revenue that could be yielded by renting out a part of the warehouse to other companies. The owner of the warehouse would still make money for the part of the warehouse that they are not using. In this research, the focus was on decreasing the amount of inventory and thus creating empty space in the warehouse. A future study could analyze how to turn that empty space into profit through rent.
8. References
Abdelaziz, F.B. and Mejri, S. 2013. Decentralized bilevel for shared inventory management. Production planning and control. 24 (8-9), p. 684-701
Alfares, H.K. 2015. Production and inventory planning with variable holding cost and all-units quantity discounts. Information systems and operational research. 53 (4), p. 170-177
Biswas, S.K., Karmaker, C.L., Islam, A., Hossain, N. and Ahmed, S. 2017. Analysis of different inventory control techniques: a case study in a retail shop. Journal of supply
chain management systems. 6 (3), p. 35-45
Chen, J. 2012. Multiple criteria ABC inventory classification using two virtual items.
International Journal of Production Research. 50 (6), p. 1702-1713
Christopher, M. and Ryals, L.J. 2014. The supply becomes the demand chain. Journal of
Business Logistics. 35 (1), p. 29-35
Cobb, B.R. 2013. Spreadsheet modeling of (Q,r) inventory policies. Decision sciences
journal of innovative education. 11 (2), p. 175-184 Durlinger P.P.J., 2009. Verhoog uw winst, ROI en ROA met een gedegen ABC-analyse- een 10 stappen plan voor productiebedrijven. Posterholt, Nederland: Durlinger Consultancy Frey, H.C. and Patil, S.R. 2002. Identification and review of sensitivity analysis methods. Risk analysis: an international journal. 22 (3), p. 553-578
Hadley, G.T. and Whitin, T.M. 1963. Analysis of inventory systems. Englewood Cliffs, United States of America: Prentice Hall
Hatefi, S.M., Torabi, S.A. and Bagheri, P. 2014. Multi-criteria ABC inventory classification with mixed quantitative and qualitative criteria. International journal of production
research. 52 (3), p. 776-786 Helton, J.C. and Davis, F.J. 2002. Illustration of sampling-based methods for uncertainty and sensitivity analysis. Risk analysis: an international journal. 22 (3), p. 591-622 Iman R.L. and Helton, J.C. 1998. An investigation of uncertainty and sensitivity analysis techniques for computer models. Risk analysis: an international journal. 8 (1), p. 71-90.
Janova, J. 2012. Crop planning optimization model: the validation and verification process. Central European journal of operation research. 20 (3), p. 451-462 Kao, C. and Hsu, W. 2002. Lot size-reorder point inventory model with fuzzy demands. Computers and mathematics with applications. 43 (10-11), p. 1291-1302 Ramanathan, R. 2006. ABC inventory classification with multiple-criteria using weighted linear optimization. Computers and operations research. 33 (3), p. 695-700 Robinson, S. 2014. Simulation: the practice of model development and use. (2nd ed.). New York, United States of America: Palgrave Macmillan. Salam, A., Panahifar, F. and Byrne, P.J. 2016. Retail supply chain service levels: the role of inventory storage. Journal of Enterprise Information Management. 29 (6), p. 887-902
Saltelli, A. 2002. Sensitivity analysis for importance assessment. Risk analysis: an
Seetharama, L.N., Dennis, W.M. and Pater, J.B. 2003. Production planning and inventory
control. (2nd ed.). India, New Delhi: Presentice Hall of India
Stanislaw, K. 2017. Influence of delivery quantity on service level in the stock replenishment system based on reorder level for irregular distributions of demand.
Scientific journal of logistic. 13 (4), p. 415-427
Sternman, J.D. and Dogan, G. 2015. “I’m not hoarding, I’m just stocking up before the hoarders get here”: behavioral causes of phantom ordering in supply chains. Journal of
Operations Management. 39 (40), p. 6-22
Tempelmeier, H. 2011. Inventory-management in supply networks – problems, models,
9. Appendices
Appendix A
Overview of total number of SKUs per number of pallets places
Number of pallets places Total number of SKUs Total number of pallets places
0 127 0 1 2310 2310 2 231 462 3 91 273 4 64 256 5 36 180 6 25 150 7 15 105 8 12 96 9 5 45 10 10 100 11 3 33 12 1 12 13 5 65 14 3 42 15 1 15 16 1 16 19 1 19 20 1 20 22 1 22 23 2 46 26 2 52 27 1 27 30 1 30 31 1 31 32 1 32 33 1 33 34 1 34 43 1 43 49 1 49 53 1 53 61 1 61 82 1 82 92 1 92 93 1 93 105 1 105 201 1 201 224 1 224
