University of Groningen
New cooperation forms for web shops
Which cooperation form can help web shops to make a combined shipment to the customer?
By
Jutha van den Broek j.h.van.den.broek@student.rug.nl
Student number: 1682660
Supervisor: Prof. K.J. Roodbergen Second assessor: Dr. M.J. Land
Preface
The paper in front of you is my master thesis for the master Business Administration, specialization Operations & Supply Chains. With this research I have finished my master at the University of Groningen. I performed a research on the possibility of making a combined shipment to the customer when two web shops are involved. The research was an experience where I have learned a lot on the subject. There are several people who I would like to thank for helping me with my thesis.
First my thanks goes to prof. K.J. Roodbergen. He gave me useful insights and feedback during my research. Second of all I would like to thank my second assessor, dr. M.J. Land for helping me with the right formulation of my mathematical model.
Next to my assessors, I would like to thank my classmates Elise and Tessa for giving me critical comments on my research. Also I would like to thank my parents, boyfriend and friends for supporting me through my research.
List of tables and figures
List of figures
Figure 1: Variables influencing a combined shipment p. 11
Figure 2: E-commerce fulfillment models p. 13
Figure 3: Fulfillment model in this research p. 15
Figure 4: Options for a company p. 21
List of tables
Table 1: Subdivision of categories p. 18
Table 2: Web shop combinations p. 19
Table 3: Share of product of combinations on total volume p. 19 Table 4: Formula of standard normal distribution for statistical chance of
warehouse A going out of stock p. 24
Table 5: Products central in this research p. 28
Table 6: Share of products for web shops Omorfiá p. 29
Table 7: Transportation costs p. 30
Table 8: Demand per product in combination with other web shops p. 30
Table 9: Total transportation costs per web shop p. 31
Table 10: Handling cost per product p. 31
Table 11: Variable inventory costs per product p. 32
Table 12: Savings per product p. 33
Table 13: Total savings per product option 1 p. 33
Table 14: Total savings per product option 2 p. 34
Abstract
Content
1. Introduction ... 6
1.1 Motivation ... 6
1.2 Problem exploration... 6
1.3 Research objective and research question ... 8
1.4 Method of the research ... 9
1.5 Structure of the research ... 9
2. Theoretical Background ... 11
2.1 Inventory management ... 12
2.2 Transportation ... 13
2.3 Transportation and inventory management ... 15
2.4 Conclusion ... 16
3. Methodology ... 17
3.1 Assumptions ... 17
3.2 Data modeling ... 18
3.3 Choosing of the web shops ... 19
3.4 Conclusion ... 21
4. Model formulation ... 22
4.1 Model formulation ... 22
4.2 Defining the variables. ... 26
4.3 Context ... 28
4.4 Conclusion ... 28
5. Results ... 29
5.1 Defining the products ... 29
5.2 Variables ... 30 5.3 Proof of concept ... 34 5.4 Discussion ... 36 5.5 Conclusion ... 37 6. Conclusion ... 38 6.1 Conclusion ... 38
6.2 Limitations and further research ... 39
6.3 Managerial implications ... 40
References ... 42
Appendix 1 ... 48
1. Introduction
1.1 Motivation
In the Netherlands, internet sales are rapidly rising. Where in 2002 only 28 percent of the population (people aged between 12 and 74) purchased online, this percentage has grown to 71 percent in 2010. About 23 percent of the population has never bought products online (CBS). This percentage will decrease because of the rising trend of internet sales. Because of the growing popularity of buying online, more stores are shifting from selling the old fashioned way to selling online. To measure up to the demand of online sales some companies restructure their work floor. A company visit at Centraal Boekhuis learned that at first they only had a small space for internet sales. However book sales in stores dropped and the internet sales increased. To handle this change, Centraal Boekhuis changed the lay-out of the work floor. Space which was first used for to provide stores with books is now used in to fulfill internet orders.
Different products can be bought online like: books, clothes, media, jewelry, household appliances, and so on. Most web shops are specialized in one branch to fulfill customer needs. This means that when a customer purchases items online they have a different account for each web shop. When ordering at multiple web shops, the customer receives the products at different times. For the customer it is inconvenient to be at home at the different times to receive their packages. Of-course some packages fit in the mailbox, but this is not always the case. Nuzakelijk.nl executed a research into customers’ likes and dislikes when it comes to delivery services. Nuzakelijk.nl states that customers want more choice in home delivery, like more freedom for when it will be delivered. Another option for the customer is to receive their purchases simultaneously, even if they order at multiple online retailers. For customers this means that they only have to be home at one particular time of the day. 1.2 Problem exploration
the warehouses of the individual retailers to the customer? After the order is received the products are picked and shipped to the customer. If the products are stored at separate warehouses, transportation between these warehouses might be needed before the products can be shipped simultaneously. Shipping to the customer, or the actual order fulfillment, is an important aspect in this paper. Order fulfillment is one of the weakest links in e-commerce (Ayanso et al. 2006), therefore it is important to examine what is needed to fulfill the orders of the customer when buying at separate online retailers.
1.2.1 Practical relevance
As stated in the motivation, customers wish to have more freedom in their choice of home delivery. The research of Nuzakelijk.nl has concluded that customers would like more option when it comes to home delivery. They would like to have the possibility of collecting their package at the retailer or, in the case of home delivery, the customer wants to choose the time of delivery, in case of home delivery. This research examines whether part of the wishes of the customer can be met by making a combined shipment from multiple web shops to the customer. With delivery of orders in one shipment, customers only have to take care that someone is at home for one moment of time instead of multiple times, therefore increasing customer’s satisfaction.
1.2.2 Scientific relevance
This research investigates the possibility of storing products in a warehouse of a business partner. This makes it possible for two retailers to combine orders in one warehouse and therefore orders can be delivered in one shipment. For this to happen, first of all products must be selected which will be stored at a warehouse of a business partner. Product demand plays a role in deciding which products to relocate. In this research the following variables will be researched: inventory management, space allocation and transportation and in which way they influence each other. During this research no literature was found about the subject of storing packages at a warehouse of a business partner to be able to deliver one package to the customer instead of several packages.
model has a history of about 50 years. Multi-echelon models search for the most efficient way of ordering policies for all locations in a supply chain network (Kochel & Thiem, 2011). In these models the ordering policies are defined for all stages in a supply chain, in this case no research has been done into inventory management. For this reason this model will not be adequate in this research.
Transportation is another important aspect in this situation. Transportation can be up to 50 percent of all logistic costs (Swenseth & Godfrey, 2002). For this reason transportation can be important to take into account at an early stage when exploring new opportunities. When transporting products between warehouses extra costs will be made (Kukreja et al. 2001). Extra transportation costs are made, because from one distribution center (DC) several production facilities are replenished. In this research transportation brings products from location A to location B and the other way around, while by pooling of stock the products only gets transported from location A to location B. Transportation costs in this case therefore will be lower because the costs can be shared by more products. Yao et al. (2010) have done research into serving the customer directly from the warehouse and directly from the production facility. In this research the customer will be served from one or multiple warehouses, but the customer will only receive one shipment.
1.3 Research objective and research question
Following from the problem exploration, the research objective and research question can be defined.
1.3.1 The objective of the research
To gain insight in product location and product selection for multiple web shops, with each having its own warehouse, working together in ensuring a combined shipment to the customer, by performing a literature research and by building a model.
1.3.2 The research question of the research
1.3.3 Sub questions
The research question will be answered in two ways. First of all a literature research will be performed into the variables influencing combined shipment to the customer. Secondly a model will be built to research the possibility of storing products in the warehouse of a business partner to realize combined shipment.
For the literature research the following sub question will be answered: Which variables influences combined shipment according to literature? For the methodology section the following sub question will be answered:
How is data gathered for examining the possibility of making a combined shipment? For building and testing the mathematical model the following two sub questions will be answered:
1. How will the mathematical model be formulated and how will the variables of the literature research be formulated in the model?
2. What is the outcome of the mathematical model, as formulated in chapter 4, when data of a case is used?
1.4 Method of the research
First a literature research is performed into inventory management and transportation. After the literature research is performed, a mathematical model is built. The model will be used to examine a case. Data to be used is based on data of Neckermann, therefore some tables are excluded in the public version. The case will support whether the mathematical model will work. The calculations with the data and the proof of concept will be performed by using the program excel.
1.5 Structure of the research
2. Theoretical Background
To make combined shipment to the customer possible, three variables are important to consider (figure 1): inventory management, space allocation and transportation. Space allocation is concerned with allocating products to space (Desmet & Renaudin, 1998). Inventory management is concerned with the planning and control of inventories (APICS dictionary, 2004). Transportation costs are the costs associated with the transportation between warehouses and transportation to the customer. In this research only attention is given to the two variables transportation costs and inventory management. A research is done after the possibility of transporting products and storing products at the warehouse of a business partner. If the results shows that it worthwhile for companies to exchange products in making a combined shipment to the customer, companies must decide which products to store at the warehouse of a business partner. Space allocation can be used to decide which products to store in the other warehouse. A company must know how much space there is to store products in, and has to keep space free for products which they will receive from the other company. Because it is not yet known if it is worthwhile for companies to exchange products, no attention is given to space allocation in this research. However some background information can be found in appendix 1.
Figure 1 Variables influencing a combined shipment
2.1 Inventory management
As stated above is inventory management the department of business management which is concerned with the planning and the control of inventories (APICS dictionary, 2004). It includes the tasks in the inventory control process for the logistics network (Schönsleben, 2007). Keeping inventory is important because it avoids that a company is out of stock, it gives protection when a supplier is unable to deliver and inventory gives security if something goes wrong in the production process. Also inventory is kept because a certain batch size of products is needed for the production process or for transportation. The dilemma with inventory management is that on the one hand a high inventory level is needed to satisfy customers demand and so generate revenues; on the other hand high inventory levels lead to high inventory costs.
2.1.1 Inventory costs
Inventory costs can be divided into the following types of costs (Schönsleben, 2007, Huang et al. 2011):
1. Setup & ordering costs (the costs associated when ordering products)
The higher the carrying cost of a product, the faster a company wants to sell that product. Carrying cost of a product is defined by three variables (Yim, 2003):
1. Demand of the product
2. Desired customer service level of a product
3. Order frequency and lead time of replenishment of a product
A reduction of the lead time will decrease inventory costs and will allow a company to use more accurate demand information. This can improve the customer service level and can reduce safety stock requirements. When reducing lead-time, certain risks may arise, like high purchasing costs and out of stock of supplies (Liao & Shyu, 1991).
Deterioration can take place for several products at once at the end of a planning horizon (like clothes in the fashion industry), or during a planning horizon. Two different categories exists (Raafat, 1991):
1. Products that have a fixed shelf life, like foods.
2. Items which continuously decay, like radioactive materials.
When inventory deteriorates, extra costs are associated with it (Huang et al. 2011). These costs are variable, because the longer a product is in stock the more chance deterioration can take place and therefore inventory costs will increase.
Inventory costs can be divided into fixed inventory costs and variable inventory costs. Fixed inventory costs are the costs which do not change over time, for example ordering costs. Variable inventory costs are the costs which change over time, for example carrying costs; the longer a product is stored in a warehouse the higher the costs will be of storing that product.
2.2 Transportation
In case there are more than two warehouses, an important consideration is the transport between those warehouses. Gallego & Simchi-levi (1999) argue that direct shipping is more desirable, this means that only one warehouse at a time is served. However Hall (1992) argues that stopping at multiple suppliers and buyers provides cost savings, therefore serving multiple warehouses. When stopping at multiple warehouses, vehicle routing is an important aspect. Vehicle routing is about the route a truck drives between warehouses and is usually defined in the transportation policy (Archetti et al, 2011). Most of the time the vehicle routes are fixed because it is about the transport between warehouses. The route the vehicle takes to transport to the customers is more difficult to manage because different products are transported to different locations in an area. This means that the vehicle route is different every day. One solution to solve this problem is to make vehicle routes based on areas (Mar-Ortiz et al. 2011). For customers who order online, several ways exists in how the product is delivered to the customer. According to Skjøtt-Larsen et al. (2007) six e-commerce fulfillment models exists (figure1). The different models show the information and product flow when a customer orders a product on the internet.
M: Manufacturer E: E-commerce site C: Customer R: Retailer I: Inventory FC: Fulfillment center DC: Distribution center
Physical flow Information flow
Figure 2 E-commerce fulfillment models
goods are collected and delivered to multiple warehouses, the truckload needed should be determined beforehand. It is a waste of time and money, if a truck is driving half empty. On the other hand it is also undesirable when the truck is full while more products are waiting to be shipped. Determining the truckload is easily done when there is only one type of product to consider. When different types of products need to be shipped it gets harder to determine the truckload. Sancak & Salman (2011) solve this problem by estimating the number of trucks needed in a period, using the pallet count of the product types. The fleet of vehicles is also an important consideration. When there are not enough trucks or the trucks are too small or too large, there is a chance that not all products can be transported or that there is a truck driving half loaded (Etezadi & Beasley, 1983).
2.2.1 Transportation costs
As mentioned before there are two types of carriages, for-hire carriages and private carriages. When using a for-hire carriage the transportation costs are fixed and determined by the shipper. When using a private carriage, depreciation costs arise and these costs must be taken into account in the calculation of the transportation price. Next to depreciation costs, fuel costs and driver wages must be taken into account. Carriers refer to these costs as operating costs (Toptal & Bingol, 2011). When using private carriers, a company also has to pay road taxes (www.belastingdienst.nl). The EU wants to charge heavy freight transport with emission rights. However at the time of writing this is not yet approved by the European council (www.europa.eu).
2.3 Transportation and inventory management
known transportation costs to the customer can decrease. If the product location is known it is more easily to get that product than it changes warehouses all the time.
2.4 Conclusion
As shown in this chapter space allocation, transportation and inventory management are important variables for making a combined shipment to the customer. However space allocation is excluded in this research because first the transportation costs and storage costs of products must be defined. When this has happened and the research shows it is worthwhile then space allocation becomes important to research. It shows that in inventory management, it is important to make a trade-off between the amount of products to have in storage and the customer service level a company want to achieve. For transportation several e-commerce models exist (see figure 2), in this research a new type of e-commerce model will be researched (figure 3), the model shows that the product to the customer is delivered from one of the two warehouses.
Information flow Product flow
Figure 3 Fulfillment model in this research
E-commerce site A
E-commerce site B
3. Methodology
3.1 Assumptions
When performing a research, not everything can be investigated. In this research a mathematical model will be built to examine if it is worthwhile to store products at a warehouse of a business partner. For this model the following assumptions hold:
There are two web shops. Because this is the first time a research is conducted into web shops working together for delivering one package to the customer, two web shops will be used.
As stated in the introduction the number and location of warehouses can be complex and can be a research on its own. That is why there is a limit of one warehouse per web shop.
Seasonality is left out. Seasonality is not discussed because it will make the model too complex. For seasonality and how to determine optimal inventory the research of Xiao et al. (2009) should be examined.
The products are non-perishable, as perishable products are not commonly sold on the internet and make it more difficult in defining products which to store in the warehouse of a business partner because deterioration can take place, and it is not sure how the demand is in the warehouse of a business partner.
Product delivery to the customer will be done by PostNL to assure fixed prices.
For the replenishment of a companies’ own warehouse, nothing will change. Because fast-moving products are stored at each other’s warehouse, there must be a replenishment strategy. Replenishment of the other warehouse will be cyclical by using a truck riding between the two warehouses.
3.2 Data modeling
To investigate if it is possible to store products in two warehouses, data of two web shops is needed. However web shops do not yet work together in making a combined shipment to the customer. This means that data is difficult to gather. Using data from two separate web shops cannot be used because there is only a small chance that a customer buys online in the two different web shops which you have chosen in the same time frame. For this reason only data is gathered from one large online retailer. The retailer must have several product categories. These product categories must be known to the customer and the customer is able to buy in the several categories in the same time frame. The different categories will be used to form the different web shops in this research.
For building the mathematical model data will be used from the online retailer Neckermann. Reasons for choosing Neckermann is that it is one of the largest e-commerce companies in Europe (www.thuiswinkel.org). Until 2006 Neckermann was a mail order company, which also had an internet site. In 2006 they updated their strategy from a mail order company to an e-commerce company. Neckermann is best known for their fashion, but they also sell electronics, lifestyle accessories, and sports accessories etcetera. (www.neck.nl). This data was gathered by a student who performed her research at Neckermann (Langkau, 2011).
Because Neckermann is a large company, a large number of orders come in every day. Those orders is the data which will be used in this research. Only for Neckermann is chosen because customers can order in several product categories which exist in Neckermann. Neckermann has 19 different product categories and in this way a customer buys more easily products of several categories at one time. In the case of Neckermann customers buys at different categories because it is easy to buy all products in one web shop, instead of visiting multiple web shops. The different product categories can be used to model individual web shops. In this way there is assurance that customers buy products in a certain time frame and in different categories. The different categories will function in this research as separate virtual web shops.
category. With the help of those four items, there is an overview of which products are ordered by a customer and in which product category. The combination of product categories which is ordered most in, are the two web shops for this research.
3.3 Choosing of the web shops
With the help of excel the data of Neckermann is analyzed. In Neckermann there are 19 product categories. This is a large number and will result that some categories will only have a small amount of data. This means that there is a small chance that a lot of combinations are made with other web shops. For this reason four large categories are made and every category represents one virtual web shop. All categories which have to do with clothing are assigned to the web shop Fashion. The category jewelry and watches is not chosen for this category because these are accessories and not real clothing. The categories of electronics are in one section because several web shops sell different categories of electronics. The lifestyle accessories category is chosen because all these categories have something to do with decorating a house. The categories exist of products for the bathroom, the bedroom and the living room. The category other is chosen because the last categories could not fit easily in one of the other large categories. Because these categories are too small to let them be their own category, one web shop other is made. Table 1 shows the four large categories and the categories which fall under those web shops.
Electronics Fashion Lifestyle accessories Other
Consumer Electronics Lady fashion Sleeping Jewelry and watches Household appliances Men fashion Bedroom textiles Sports and leisure
accessories Computer electronics Kids fashion Bath and HH basics Gifts
Sports fashion Home accessories Games and toys Under and night
wear
Home textiles Beauty
Big sizes
Table 1 Subdivision of categories
Category Electronics
Fashion X Fashion
Lifestyle X X Lifestyle
Other X X X
Table 2 Web shop combinations for products
This table shows that the most combinations of products are made between the web shop Fashion and the web shop Other. Data of the web shops Lifestyle and Electronics will be excluded from now on. Those data will be excluded because the research is about the exchanging of products between two web shops. Table 3 shows the percentage of orders made in combination with the total amount of products ordered in one web shop.
Web shop % of orders in
combination with products in total
Other 24.61 %
Fashion 1.71 %
Table 3 Share of product of combinations on total volume
Using the name web shop Other makes the text difficult to read, other names for the web shops will be used, which are the following:
Other: ‘Syllogí’ (Greek for collection) Fashion: ‘Omorfiá’ (Greek for beauty)
suitable solution because 141 customers ordered in three web shops and 9 of them ordered in all four web shops.
3.4 Conclusion
4. Model formulation
4.1 Model formulation
For a company who wants to store inventory at a business partner, two options exists. In the original situation both companies stored their products in their own warehouse. The first option is that a company can decide to store extra inventory at the warehouse of a business partner. This means that inventory is added and the total amount of products increases. The second option for a company is to keep the total inventory the same but relocate some inventory to the warehouse of a business partner, figure 4 gives an example with 70 products.
Figure 4 Options for storing products
In this paragraph the mathematical model will be explained. First the option of adding inventory is discussed, after which the option of replacing inventory is discussed. The following variables are used in this research.
Originial
situation
Total: 70 Warehouse A: 70 Warehouse B: 0Option 1
Total: 77 Warehouse A: 70 Warehouse B: 7Option 2
Total: 70 Warehouse A: 63 Warehouse B: 7S1 Total savings per shipped product for option 1 per period S2 Total savings per shipped product for option 2 per period Y Gross savings of transportation to the customer per product T Transportation costs between warehouses per product If Fixed inventory costs (handling costs) per product Iv Variable inventory costs (carrying costs) per product Ti Time in inventory per product
Option 1 Adding extra inventory
The formula is built upon the idea that transportation savings to the customer are made because products are transported to the customer from one warehouse instead of two warehouses. However when shipping from one location, other costs arise because the products must be transported and stored at the other warehouse. Extra costs are inventory costs and transportation costs. In this option extra transportation costs, fixed and variable inventory costs are made because extra inventory is added. The following formula shows the total savings for this option.
S1 = Y – (T + If + (Iv * Ti)) (1)
When there is a positive result it means that it is worthwhile to transport the products to, and store the products at the other warehouse. It shows the amount of money which will be saved per product. If there is a negative result it means that for a company it is better to transport the products directly to the client instead of storing the products at a warehouse of a business partner. For option 1 there is no extra chance that warehouse A goes out of stock because the inventory level stays the same.
Option 2 Relocation of inventory
S2 = Y – (T + If) (2)
With option two there is a chance that warehouse A is earlier out of stock. Inventory level at warehouse A decreases because of the relocation of inventory. The customer service level will not change because in total there are still 70 products available for the customer (see the example of figure 4). For the company however extra transportation costs are made between the two warehouses because a package is sent to warehouse B while it is needed in warehouse A. There are three ways in how extra costs can arise, product A will be used as example.
A. Only product A is ordered, so it does not have to be combined with other products. In this case only S2 is the extra cost which is made.
B. Product A is ordered in combination with products of web shop A. This means that product A which is stored in warehouse B must be send back to warehouse A. This means that the costs of S2 arise twice, for sending the product back and forth.
C. The product is ordered in combination with products of web shop A and B. Here there are two options:
I. Products are first shipped from warehouse A to warehouse B, this means that S2 is no extra cost, because for the other products also S2 occur.
II. Products are shipped from warehouse A to warehouse B, this means that for the product of A S2 is counted twice (because of bringing it back and forth). For the products of web shop B, only S2 arise.
In the old situation there were 70 products in warehouse A. The warehouse has, for example, a chance of going out of stock of 5 percent. In the new situation there are 63 products in warehouse A and 7 in warehouse B. The total chance of going out of stock stays 5 percent. However the chance of warehouse A going out of stock will change. It is not to say of this will be higher or lower, because the demand will also change for warehouse A. To calculate how the chance of going out of stock will change, a standard normal distribution can be used. The following variables are used:
P Probability of going out of stock 𝑥 Demand of warehouse A
σ Standard deviation µ Average of demand V Total inventory
v Inventory at warehouse B
Table 4 shows the standard normal distribution per situation. Products Customer service level Formula Old Situation Total products 70 P1 𝑃 = (𝑥 − µ σ < 𝑉 − µ σ ) Warehouse A 70 P1 𝑃 = (𝑥 − µ σ < 𝑉 − µ σ ) Warehouse B 0 - New Situation Total products 70 P1 𝑃 = (𝑥 − µ σ < 𝑉 − µ σ ) Warehouse A 63 P2 𝑃 = (𝑥1− µ1 σ1 < 𝑉 − 𝑣 − µ1 σ1 ) Warehouse B 7 P3 𝑃 = (𝑥2 − µ2 σ2 < 𝑣 − µ2 σ2 )
Table 4 Formula of standard normal distribution for statistical chance going out of stock
changes also. For calculating the chance that a product is sent to warehouse B while it must lay in warehouse A can be done by P1-P2 (= P3). For a company it means that P3 must be as low as possible to make sure that no extra costs are made in delivering the products to the customer because the products are stored in the wrong warehouse.
4.1.1 Discussion of the formula
As the formulas shows are the total transportation costs fixed and includes the costs of the truck and number of kilometers. Inventory costs (paragraph 4.2) are the costs associated of storing the products in the warehouse (variable costs) and the handling costs (fixed costs) of shipping the product. The formulas shows that transportation costs and inventory costs are two independent variables. If one of those variables increases, the other one is not affected. This is the same for variable and fixed inventory costs. The savings (Y) are the savings of transporting the products from one location to the customer instead of two locations. Those savings are defined by the costs which a company makes in transporting the products to the customer. The total savings (S) are lower than for option 2. This is explained by the fact that for option 2 the chance of a stock-out is not taken into account in the formula. Further research should look into this and find a method to fit the costs associated with stock-out into the model.
4.2 Defining the variables.
In this paragraph the variables of the formulas will be defined.
4.2.1 Inventory costs
The inventory costs can be divided into fixed inventory costs (If) and variable inventory costs (Iv). In methodology 2.1.1, the definition of the costs are discussed.
The fixed inventory costs exist of the following: Ifh Handling Costs
Ifso Setup and ordering costs
The fixed inventory costs exist of those two costs because over time these costs will not change.
The variable inventory costs exist of the following: Ivs Storage costs
Ivd Depreciation and decay costs
The variable inventory costs exist of these costs because over time, depreciation will get higher and storage costs will also rise. The longer a product is in a warehouse, the more money it will cost to let it stay there. For the variable costs a percentage will be taken from the price.
The following formula can be made for variable inventory costs. Iv = Ivs + Ivd
When certain costs are not applicable for the company, these costs can be left out of the formula.
4.2.2 Transportation costs
Transportation costs (T) exists of the following (methodology section 2.2.1):
Tt Costs of a truck (this can be the costs of hiring a truck or depreciation of a truck) Ttd Costs of a truck driver (per hour or day)
Tf Fuel price per km times number of km Trt Road tax (not always the case)
Ter Emissions rights (maybe in the future)
This gives the following formula for transportation costs. T = Tt + Ttd + Tf + Trt + Ter
When certain costs are not applicable for the company, these costs can be left out.
4.2.3 Savings
The savings (Y) are the difference in shipping the products from one warehouse to the customer instead of two warehouses. When shipping from one warehouse, transportation costs only have to be paid once instead of twice. This gives us the following definitions
This gives the following formula for savings: Y = Yt – Yo
4.3 Context
Now the formula is formulated it is important to pay attention to the allocation of products to space. As stated in chapter 2, this variable is not yet taken into account. This variable shows which products and the amount of products which must be stored in the warehouse of a business partner. One model which can be used for allocating space is the forward-reserve model (Hackman & Platzman, 1990). This model decides which products to store in a forward area and which products only in the reserve area. The warehouse of a business partner can be seen as an extra forward area, or as the only forward area. Further research must be conducted if this is true and if this model can be used, or if another model is more suitable.
4.4 Conclusion
5. Results
5.1 Defining the products
Table 5 shows the products which will be transported in this research. Web
shop
Product Price # of orders in combination with own web shop # of orders in combination with other web shop Total # of orders % of orders in combination other web shop
Syllogí A € 24.95 3 7 10 70 %
Syllogí B € 2.50 2 7 9 77.77 %
Syllogí C € 20.99 9 9 18 50 %
Omorfiá D € 19.99 62 8 70 10 %
Omorfiá E € 27.99 71 7 78 11.42 %
Table 5 Products central in this research
Table 5 shows that for web shop Syllogí two types of products have made more combinations in customer orders with web shop Omorfiá than with web shop Syllogí. Products C has made the same amount of order combinations with web shop Omorfiá and Syllogí. For product D and E is chosen because they had the highest share in order combinations with web shop Syllogí (table 6).
In this chapter the model formulated in chapter 4 is tested with the help of a case. First the products to be used are defined. Then the individual variables will be discussed. As last the model is tested. The research question which will be answered is:
Product How many orders in total
How many orders in combinations with web shop Syllogí
How many orders in combinations with web shop Omorfiá
% of combination with Syllogí compared to total orders D 70 8 62 10 % E 78 7 71 11.42 % F 114 7 107 6.14 % G 120 7 113 5.84 %
Table 6 Share of products for web shop Omorfiá
The products showed in table 6, are the products that had seven or more order combinations with web shop Syllogí. Omorfiá also sells other products, but these had six combinations or less with web shop Syllogí.
5.2 Variables1
In this paragraph the variables of the formula are made case specific.
5.2.1 Transportation costs
Fiat Doblo
Costs hiring € 30.72 (one part of a day)
Max. weight 874 kg
Truck driver € 14.42 per hour = € 14.42 * 2.5 hours = € 36.05
Costs per km € 0.14 per km = € 0.14 * 50 km = € 7,-
Total costs € 73.77
Table 7 Transportation costs
The amount of products to be shipped must be defined. Table 8 shows the demand for combinations made with the other web shop.
Web shop Product # products for the other web shops
Syllogí A 7
Syllogí B 7
Syllogí C 9
Omorfiá D 7
Omorfiá E 7
Table 8 Demand per product made in combination with the other web shop
Web shop # number of products
Transportation costs per product
Total costs per product Syllogí 30 € 36.89 / 30 = € 1.23 € 1.23 Omorfiá 20 € 36.89 / 20 = € 1.84 € 1.84
Table 9 Total transportation costs for each web shop for one product 5.2.2 Inventory costs
Inventory costs are divided into fixed and variable inventory costs. First the fixed costs will be discussed.
Fixed costs
The fixed costs exist of handling costs and setup and ordering costs. The setup and ordering costs are unknown, so only the handling costs will be used in this research. For the handling costs the following time periods count for taking a product out of the warehouse of Neckermann (Leveling, 2010, p. 43).
Average picking time per product 0.8 s Sorting start-up per product 22 s
Sorting product 7.6 s
Packing order 51 s
Sorting/ clearing all products 1m 24 s = 84 s Average distance sorting station 25 m
Average speed 0.4 m/s
Average products per order 2.2 Average # of products in one routing
68
Table 10 Factors determining handling costs
0.8 + 22 + 7.6 + (51/2.2) + (84s/68) + (25/0.4/68) = 62.74 seconds to get a product out of the warehouse and to make it ready for shipment. If an employee costs € 14.42 per hour (same number as for the truck driver), one minute will costs a company € 0.24. Handling costs (and therefore fixed inventory costs) are €0.25 because it takes an employee is 62.74 seconds to get an order from the warehouse.
Variable costs
The variable costs exist of storage and depreciation costs. For the variable inventory costs halve of the selling price will be used, because the cost price of the products is unknown. Storage costs, including depreciation costs, for Omorfiá will be 200 percent for one year, because fashion is a very seasonal product. For web shop Syllogí the storage costs, including depreciation costs, will be 50 percent for one year. Syllogí has products with lower seasonality and lower demand, therefore the variable inventory costs is lower. Table 11 shows the variable inventory costs.
Web shop
Product Price Halve of the price
Storage and
depreciation cost per year
Variable inventory cost per week
Syllogí A € 24.95 € 12.475 € 6.238 € 0.12
Syllogí B € 2.50 € 1.25 € 0.625 € 0.01
Syllogí C € 20.99 € 10.495 € 5.248 € 0.10
Omorfiá D € 19.99 € 9.995 € 19.99 € 0.38
Omorfiá E € 27.99 € 13.995 € 27.99 € 0.54
Table 11 Variable inventory costs per product
The statistical representation of the warehouse going out of stock cannot be calculated with the data which is used in this research. One week of orders is just too few to calculate the chance of going out of stock.
5.2.3 Savings
Price shipping once Price shipping twice Savings (St-So)
€ 6.75 € 13.50 € 6.75
Table 12 Savings per product
In this research the price is the saving per product. The savings seems high but only the commercial price of sending a package is known. For companies, sending packages is probably cheaper but those prices are unknown to the public.
5.3 Proof of concept
As stated in chapter 3, the mathematical model build in chapter 4 will be judged by using a case, in paragraph 5.2 the individual variables were defined for this case and in this paragraph the model will be tested.
5.3.1 Option 1: extra inventory is owned by the company
The formula of option 1 was about subtracting savings with transportation costs, variable inventory costs and fixed inventory costs (Formula 1, paragraph, 4.1). Table 13 shows the outcome of the formula.
Table 13 Total savings per product option 1 5.3.2 Option 2: inventory is relocated
The formula for option 2 was about subtracting savings with transportation costs and variable inventory costs (formula 2, paragraph 4.1). Table 14 shows the outcome of this formula.
Product Savings Total costs per product S2 A € 6.75 € 1.48 € 5.27 B € 6.75 € 1.48 € 5.27 C € 6.75 € 1.48 € 5.27 D € 6.75 € 2.09 € 4.66 E € 6.75 € 2.09 € 4.66
Table 14 Total savings per product option 2
5.3.3 Trade-off inventory costs and transportation costs
When examining option 2 and changing the inventory costs and transportation costs a table can be made (appendix 2). This table shows the trade-off between inventory and transportation costs. The table shows that the higher the inventory costs the more products must be shipped to let it be worthwhile to transport products between warehouses. Table 15 shows the highlights.
Inventory costs # of products Grow factor compared to the number before € 0.50 6 € 1 7 1,166667 € 1.50 8 1,142857 € 2 8 1 € 2.50 9 1,125 € 3 10 1,111111 € 3.50 12 1,2 € 4 14 1,166667 € 4.50 17 1,214286 € 5 22 1,294118 € 5.50 30 1,363636 € 6 50 1,666667 € 6.50 148 2,96
Inventory costs are the fixed and variable costs added together. This means that for several products € 1.00 can be a high inventory costs while for other products this is a low inventory cost. This table only shows which amount of products must be sent to ensure lower costs than € 6.75. The grown factor increase, which means that the higher the total costs of inventory and transportation, the more products must be exchanged to make transportation worthwhile.
5.4 Discussion
5.4.1 Choice of products
First of all there were only 1500 combinations made between web shop Omorfiá and web shop Syllogí. This is not a lot considering the amount of products web shop Omorfiá sells in one week. As shown in the results 1.71 percent of the products sold of web shop Omorfiá were in combination with web shop Syllogí. For web shop Syllogí however 24.62 percent of the products were sold in combinations.
Second of all, a note must be made about the choice of products. Only information of one week was available. This means that for some products it is possible that the products sold were on discount or was a promotion product. When investigating another week the product choice can be totally different. This can still happen, because one week is only a small time frame to measure.
5.5 Conclusion
In the case of Neckermann it can be concluded that it is worthwhile for companies to work together in making a combined shipment to the customer. In this research huge savings (Y) where made in making a combined shipment to the customer. In the real world however the costs of transporting to the customer are probably lower, because in this research high prices were used in sending the package to the customer. This means that savings (Y) are lower and therefore total savings (S) will be lower.
6. Conclusion
6.1 Conclusion
The research was about the possibility of two web shops working together in making a combined shipment to the customer. The main focus was on transportation costs and the costs of storing products in the warehouse of a business partner.
As shown no research was ever done on delivering multiple orders from multiple companies in one shipment to the customer. Also storing products of a company into the warehouse of a business partner was not researched before in this situation. Because no literature existed, a mathematical model was formulated to see if it is worthwhile for companies to work together. Variables which are important in making a combined shipment to the customer are inventory management, transportation costs and space allocation. As stated before is space allocation not taken into account in this research. Transportation costs and inventory management are important because companies must know if it is worthwhile to exchange products between warehouses. Transportation costs are the costs made for transporting the products between warehouses. Inventory costs are divided into fixed and variable costs. Fixed costs are ordering and handling costs and variable costs are carrying and depreciation costs.
products must be shipped to let it be worthwhile to store products in the warehouse of a business partner.
In this research only 5 products were stored at the other warehouse. It is possible that at the moment it is not worthwhile for most web shops to store products at the warehouse of another online retailer. This lays in the fact that not a high percentage of total turnover is made in combination with other web shops. The question which has to be answered is: Is it worthwhile for the companies to go to the extra trouble to make a combined shipment to the customer for that small amount of products? When starting with storing products at warehouses of another company there are start-up costs to consider. These costs are not discussed in this research, but are important for companies. Because internet sales are rising, it may be well possible to store products at each other warehouses in the near future. In the future internet turnover will rise and therefore working together becomes more worthwhile for companies.
Because of the small turnover, companies may not be willing to start with storing products at another company. However for companies who have just merged, this research can be important. These companies have multiple warehouses and do not know the best way in making one shipment to the customer. With the help of the model, management can decide to relocate or add extra inventory at some warehouses to ship products more easily to the customer.
When web shops decide to work together it is recommended to do a marketing research to investigate what the target of both web shops is. If one web shop is focused on students while the other is focused on elderly people, it is probably not fruitful to work together. Elderly people will buy other products online than students and therefore there is a low chance that combinations of products are made. Web shops can decide to share information about their customers to examine if the same type of customer orders at the web shops. If this is the case, it can be possible for them to work together.
6.2 Limitations and further research
5.3) it is worthwhile for companies to work together. The following step is to take a look a space allocation. A model can be formulated or used, like the forward-reserve model, to allocate space to products. This model must decide for companies which products to store in the warehouse of a business partner. In this research the products were defined using the demand in combination with product of the other web shop. It may be possible that this method of assigning product is not correct.
In this research only five types of products are used, batches of ten products and only two DC’s are used. In this situation it was worthwhile to store products at the warehouse of a partner, however when more products are added, or when there are more DC’s, the system becomes more complex and other variables become important, like vehicle routes. More attention must be paid to this subject, especially when more web shops decide to work together. Additionally start-up costs are important to consider because it takes time and consultation between two companies to start to work together.
As stated in the discussion, the statistical chance of going out of stock has not been investigated, because there was a lack of data. The statistical chance of going out of stock should be further examined because it might be of influence for inventory management and the costs associated with it. There are higher transportation costs when a stock-out occurs in one warehouse.
6.3 Managerial implications
References
Articles
Abdel-Malek, L., Kullpattaranirun, T., Nanthavanij, S., 2005, A framework for comparing outsourcing strategies in multi-layered supply chains, International Journal of Production Economics, Vol. 97, issue 3, p. 318-328
Archetti, C., Bertazzi, L., Paletta, G., Speranze, M.G., 2011, Analysis of the maximum level policy in a production-distribution system, Computers & Operations Research, Vol. 38, issue 12, p. 1731- 1746
Ayanso, A., Diaby, M., Nair, S.K., 2006, Inventory rationing via drop-shipping in internet retailing: A sensitivity analysis, European Journal of Operational Research, Vol. 171, issue 1, p. 135-152
Battini, D., Faccio, M., Persona, A., Sgarbossa, F., 2010, Supermarket warehouses: stocking policies optimization in an assembly-to-order environment, International Journal of Advanced Manufacturing Technology, Vol. 50, issue 5-8, p. 775-788
Borin, N., Farris, P., 1995, A sensitivity analysis of retailer shelf management models, Journal of Retailing, Vol. 71, issue 2, p. 153-171
Campo, K., Gijsbrechts, E., 2005, Retail assortment, shelf and stock out management: issues, interplay and future challenges, Applied Stochastic Models in Business and Industry, Vol. 21, issue 4-5, p. 383-392
Cormier, G., Gunn, E.A., 1992, A review of warehouse models, European Journal of Operation Research, Vol. 58, issue 1, p. 3-13
Etezadi, T., Beasley, J.E., 1983, Vehicle fleet composition, Journal of the Operational Research Society, Vol. 43, issue 1, p. 87-91
Frank, R.E., Massy, W.F., 1970, Shelf position and space effects on sales, Journal of Marketing Research, Vol. 7, issue 1, p. 59-66
Gagliardi, J.P., Ruiz, A., Renaud, J., 2008, Space allocation and stock replenishment synchronization in a distribution center, International Journal of Production Economics, Vol. 115, issue 1, p. 19-27
Gallego, G., Simchi-Levi, D., 1990, On the effectiveness of direct shipping strategy for the one warehouse multi-retail R-systems, Management Science , Vol. 36, issue 2, p. 240–243.
Georgiadis, M.C., Tsjakis, P., Longinidis, P., Sofioglou, M.K., 2011, Optimal design of supply chain networks under uncertain transient demand variations, Omega-International Journal of Management, Vol. 39, issue 3, p. 254-272
Hackman, S.T., Platzman, L.K., 1990, Near optimal solution of generalized resource allocation problems with large capacities, Operations Research, Vol. 38, issue 5, p. 902-910
Hall, R.W., 1992, A note on bounds for direct shipping cost, Management Science, Vol. 38, issue 8, p. 1212-1214
Hoff, A., Andersson, G., Christiansen, M., Hasle, G., Lokketangen, A., 2010, Industrial aspects and literature survey: fleet composition and routing, Computers & Operations Research, Vol. 37, issue 12, p. 2041-2061
Kochel, P., Thiem, S., 2011, Search for good policies in a single-warehouse, multi-retailer system by particle swarm optimisation, International Journal of Production Economics, Vol. 133, issue 1, p. 319-325
Kukreja, A., Schmidt, C.P., Miller, D.M., 2001, Stocking decisions for low-usage items in a multilocation inventory system, Management Science, Vol. 47, issue 10, p. 1371-1383
Langkau, S., 2011, Product combinations in online shopping – what and how do customers buy online?, Master thesis Marketing at Vrije Universiteit Amsterdam
Leveling, R., 2010, Improving the order-picking process at Neckermann’s textile warehousing, Master thesis at Rotterdam Erasmus University
Liao, C.J., Shyu, C.H., 1991, An analytical determination of lead time with normal demand, International Journal of Operations and Production Management, Vol. 11, issue 9, p. 72-78
Liu, K.J., Zhou, Y.H., Zhang, Z.G., 2010, Capacitated location model with online demand pooling in a multi-channel supply chain, European Journal of Operational Research, Vol. 207, issue 1, p. 218-231
Lofti, M.M., Rabbani, M., Ghaderi, S.F., 2011, A weighted goal programming approach for replenishment planning and space allocation in a supermarket, Journal of the Operational Research Society, Vol. 62, issue 6, p. 1128-1137
Mahar, S., Bretthauer, K.M., Venkataramanen, M.A., 2009, The value of virtual pooling in dual sales channel supply chains, European Journal of Operational Research, Vol. 192, issue 2, p. 561-575
Martinez-de-Albaniz, V., Roels, G., 2011, Competing for shelf space, Production and Operations Management, Vol. 20, issue 1, p. 32-46
Petersen, J.A., Kumar, V., 2009, Are product returns a necessary evil? Antecedents and consequences, Journal of Marketing, Vol. 73, issue 3, p. 35-51
Raafat, F., 1991, Survey of literature on continuously deteriorating inventory models, Journal of the Operational Research Society, Vol. 42, issue 1, p. 27-37
Rabinovich, E., Sinha, R., Laseter, T., 2011, Unlimited shelf space in internet supply chain: treasure trove or wasteland?, Journal of Operations Management, Vol. 29, issue 4, p. 305-317
Sancak, E., Salman, F.S., 2011, Multi-item dynamic lot-sizing with delayed transportation policy, International Journal of Production Economics, Vol. 131, issue 2, p. 595-603
Swenseth, S.R., Godfrey, M.R., 2002, Incorporating transportation costs into inventory replenishment decisions, International Journal of Production Economics, Vol. 77, issue 2, p. 113-130
Tsai, M.T., Saphores, J.D., Regan, A., 2011, Valuation of freight transportation contracts under uncertainty, Transportation Research part E-Logistics and Transportation Review, Vol. 47, issue 6, p. 920-932
Toptal, A., Bingol, S.O., 2011, Transportation pricing of a truckload carrier, European Journal of Operational Research, Vol. 214, issue 3, p. 559-567
Yao, Z.S., Lee, L.H., Jaruphongsa,W., Tan, V., Hui, C.F., 2010, Multi-source facility location-allocation and inventory problem, European Journal of Operational Research, Vol. 207, issue 2, p. 750-762
Books
APICS, dictionary, 2004, 11th edition, Alexandria, VA
Foundations of inventory management, P.H. Zipkin, 2000, McGraw-hill, New York
Integral logistics management, P. Schönsleben, 2007, 3rd edition, Auerbach publications
Managing the global supply chain, T. Skjøtt-Larsen, P.B. Schary, J.H. Mikkola, H. Kotzab, 2007, 3rd edition, Copenhagen Business School Press
Appendix 1
Space allocation is concerned with assigning products to a certain amount of space. There are two different approaches in assigning products to space. The first approach is the bottom-up approach: space allocation is the process of allocating space to Stock Keeping Units (SKU) and assigning product categories to space. The second approach is the top down approach: using computing norms which use a local diagnostic of space allocation in each store (Desmet & Renaudin, 1998). The amount of space needed for storing products determines the number of warehouses and/or the size of the warehouses (Cormier & Gunn, 1992). If there is already a warehouse in place, the available space must be used in the best possible way. Allocation of products to space can be done with the help of stocking policies. A stocking policy can be simple with only one stocking point or as complex as a supply network with multiple stocking points. Companies can use a simple or complex network depending on their unique characteristics (Battini et al. 2010). Next to simple and complex policies, there are three types of stocking policies (Gagliardi et al. 2008).
1. Dedicated storage strategy: products are allocated to fixed locations.
2. Random storage strategy: products are assigned to several locations depending on the available space.
3. Class-based storage: products are assigned to a specific zone in a warehouse.
Several stocking models exists, one model which is worthwhile to mention for this research is the reserve model (Hackman & Platzman, 1990, Battini et al. 2010). The forward-reserve model divides the warehouses in two parts. In one smaller part (the forward area) products with high demand is stored, in the other bigger area (the reserve area) are products stored with lower demand. In this research the warehouse of a business partner can be treated as a second forward area. Space allocation is influenced by several variables, which will be discussed shortly in the following paragraphs.
1.1 Demand
different authors examined demand and shelf space but they all included different variables to measure optimal shelf space. They all state that shelf space is dependent on the demand but how the products must be allocated depends on several variables. Because of the different variables there is a different outcome. In the case of space allocation and the influence of internet sales no research has been conducted, therefore it is unknown which variable is the most important and which method is best to use.
1.2 Product returns
Product returns are also important to consider. When a customer buys a product on the internet, the customer cannot see the product and therefore know whether it meets their expectations. For this reason products are returned more easily (Petersen & Kumar, 2009). On the internet there is a larger choice of products, therefore customers make a wrong choice more easily (Rabinovich, 2011). Product returns have an influence on shelf space because the returned products have to be stored somewhere and must be taken into account in the stocking policy.
1.3 Safety stock
Appendix 2
The following table shows the amount of products needed if the inventory gets higher. In this table only whole numbers for inventory costs are used, to keep the table readable. From where the numbers are made bold, it is worthwhile to ship the products as in the case used in this research. Inventory cost is in this situation the sum of fixed and variable inventory costs. The transportation cost is € 36.89 is divided by the number of products. The inventory costs and transport costs are added up.
