Developing a decision model to define inventory policies for products of DSM Food Specialties
M.C. Smit
15 July 2005
Inventory
Policy
Decision
Making
Inventory policy decision making
Developing a decision model to define inventory policies for products of DSM Food Specialties
Groningen, 15 July 2005
Thesis
Technology Management
Faculty of Management & Organisation University of Groningen
The Netherlands
Author: M.C. Smit
Supervisors : prof. dr. J. Wijngaard (University of Groningen) dr. D.P. van Donk (University of Groningen) A. Pier (DSM Food Specialties)
The author is responsible for the contents of this thesis. The copyright of this thesis rests with the author.
Preface
In January 2005 I started my internship at DSM Food Specialties (DFS). During this research, that takes part of my graduation, I did not want to develop theoretical insights only. I wanted this research to result in practical insights for the company as well, for DFS to be able to implement and use. Looking back at the last six months I believe I have reached this objective. Carrying out the research gave me a better understanding of the range of issues connected to supply chain management. At the same time, it gave me more insight in the focal issues that need to be addressed when working on a long-term project like this. I would like to thank several people that have supported me during the continuation of this research.
First of all I would like to thank Jan Smit for pointing out this opportunity and helping me with setting up the first contact with the Demand and Supply Chain knowledge management systems technology, DSCM was a very suitable place for this internship. Furthermore I would like to thank Ed Kooijman at DSCM for giving me this chance. I thank Alex Pier for his support during the start and the continuation of this research. At the starting phase Alex helped me with setting up the research definition and goals. When the research was set up he always kept a critical view and gave me constructive input to further improve the research. Furthermore I would like to thank my supervisors from the university: prof. dr. J. Wijngaard and dr. D.P. van Donk. Special thanks go to prof. dr. J. Wijngaard. I am very grateful for his effort and his enthusiasm during the last six months. Dr. D.P. van Donk provided me with constructive help to further improve this thesis as well. Finally I would like to thank all people at DSCM that answered the, sometimes long, lists of questions that I had for them. Thanks go also to Frank Venema, Martijn de Haas, Peter Jan Faber, Fatima Zoundri, Ardy van Erp and Gerard Vermeulen, who attended my workshops and gave me input to further develop the decision model that I developed.
Maartje Smit Delft, 2005
Summary
The research described in this thesis is performed at the Demand and Supply Chain Management (DSCM) department of DSM Food Specialties (DFS). DFS produces and markets value-added ingredient solutions for the international food and beverage industries. The products include for example enzymes, yeast extracts and cheese coatings. DSCM is supporting DFS in activities concerning purchasing, planning, logistics and customer service. During a first analysis of current issues at DSCM, it appeared that the inventory policies were based on rules of thumb and experience of the people involved. There was no overview on all the issues that influence an inventory policy. This information is not only needed to determine a ‘fitting’ inventory policy for each product (to match for example with a certain customer service level or to minimize total costs). More insight in the issues that are involved also shows management which measures can be taken to adapt inventory levels.
The goal of this research was therefore to develop a decision process for selecting inventory policies at DFS. This design resulted in a decision model as well as an implementation plan.
Based on a diagnosis of the DFS products and the organisation, a decision model is developed that consists of nine decision groups. For each decision group it is described what decisions need to be made and how they can be made. The nine decision groups are:
• Strategy definitions. This decision group is concerned with the definition of some strategic concepts that set the organisational context. Defining customer service measures and height is and example of a decision in this group.
• Strategies on product level are other strategic issues that are less general than the previous decisions. The decision whether or not to sell a product in the first place is such a decision.
• Stratification decisions. Product stratification can be used to classify the important products that need extra management attention. Different stock policies can be used for the different product classes. The customers can be prioritised as well.
• Chain design decisions involve those decisions that are related to the design of the supply chain of the products (like inventory points, transport routes and so on). After making a picture of the current supply chain, the chain is evaluated and (when possible) simplified.
• Concept chain development. The production model uses concept chains. Concept chains are general supply chains of DFS products. A concept chain does not have to resemble a complete supply chain. It can be compared to a piece of a puzzle. Several pieces resemble a complete chain. Each DFS supply chain can consist of a different combination of pieces compared to the supply chain of other products.
Using concept chains simplifies the inventory policy decision by regarding small parts of the chain individually, before integrating them. During the research there are eight concept chains defined for DFS. These are called Purchasing, Distribution, Cycled production, Production with capacity problems, Production with capacity problems and a divergent material, Tolling, Make to Order and Assembly. The main issues, managing parties and managerial instruments are described for each concept chain, together with the data requirements of the necessary decisions.
• Concept chain matching involves choosing a matching concept chain with an actual chain of a product. By using the concept chain it is shown what are the important decision variables in this case, what information is needed to make these decisions and who should manage these decisions.
• Production process decisions involve the decisions on the elements of the production process that influence the inventory policies, like for example batch sizes, packaging and replenishment quantities. These decisions can be influenced by several cost elements.
• Control system design decisions are concerned with setting up the chosen control system.
These decisions are mainly concerned with which ordering method is used.
• Quantitative decisions involve the actual calculations of how much and when to order. The calculation method is dependent on the chosen control system and the concept chain.
This paper concludes with recommendations on the implementation of the decision model within the DSCM organisation. The responsibilities of the decision elements of each decision group and the frequency of the decision moments are presented. It is argued that a project team should be formed in order to implement the decision model as well as to take care of the continuation of the usage of the decision model. Fitting information systems need to be installed as well to support the decision takers in their tasks.
Contents
Part I Introduction to the research ... 2
1 Company profile... 3
1.1 Introduction ... 3
1.2 DSM... 3
1.3 DSM Food Specialties ... 4
1.4 Demand and Supply Chain Management ... 6
2 Problem definition ... 9
2.1 Background ... 9
2.2 Research objective and main question ... 9
2.3 Research methodology... 12
2.4 Deliverables and scope setting ... 13
Part II Diagnosis phase ... 15
3 Diagnosis methodology and definitions ... 16
3.1 Introduction ... 16
3.2 Supply chain definition... 16
3.3 Product selection criteria ... 17
4 Results: General characteristics ... 20
4.1 Introduction ... 20
4.2 DFS products and organisational structure ... 20
4.3 Performance measures ... 22
5 Results: Other supply chain characteristics... 27
5.1 Introduction ... 27
5.2 CDT ... 27
5.3 SG and SP... 29
5.4 WF ... 31
5.5 WR... 32
5.6 Enz... 33
6 Diagnosis phase: conclusion ... 35
Part III: Design phase ... 36
7 Introduction to the decision model ... 37
7.1 Introduction ... 37
7.2 Decision groups ... 38
7.3 Decision model ... 40
8 Strategic decisions ... 42
8.1 Strategy definitions ... 42
8.2 Set strategies on product level ... 44
9 Tactical decisions ... 49
9.1 Stratification decisions... 49
9.2 Chain design decisions... 51
9.3 Concept chain development ... 55
9.4 Concept chain matching ... 70
10 Operational decisions ... 71
10.1 Production process decisions... 71
10.2 Control system design decisions ... 73
10.3 Quantitative decisions ... 76
Part IV Change phase ... 79
11 Tasks and responsibilities ... 80
11.1 Introduction ... 80
11.2 Overall responsibility ... 80
11.3 Strategy definitions ... 81
11.4 Set strategies on product level ... 82
11.5 Stratification decisions... 83
11.6 Supply chain design decisions ... 83
11.7 Concept chain development and matching ... 84
11.8 Production process decisions... 84
11.9 Control system design decisions ... 84
11.10 Quantitative decisions ... 85
11.11 Concluding ... 85
12 Implementation ... 86
12.1 Introduction ... 86
12.2 Implementation method ... 86
12.3 Dealing with resistance to change... 87
12.4 Rate of response to change ... 88
12.5 Use of information systems ... 88
13 Conclusion... 91
Bibliography ... 94
Appendix A: Organization chart Demand Chain Planning ...Error! Bookmark not defined.
Appendix B: Researched elements...Error! Bookmark not defined.
Appendix C: Case descriptions...Error! Bookmark not defined.
Appendix D: Map...Error! Bookmark not defined.
Appendix E: Decision variables from the cases... Error! Bookmark not defined.
Appendix F: Concept chain development ... Error! Bookmark not defined.
Appendix G: The decision model and all its relations ... Error! Bookmark not defined.
Part I Introduction to the research
This thesis is split up in four parts. This first part gives an introduction to the research. The second, third and fourth part describe the actual research itself. The conclusion follows adjacent to the fourth part.
In this first part DSM Food Specialties, the organisation where the research is performed, is shortly described (chapter 1). Following from this description the research goal is determined.
The deliverables and the research steps are described as well (Chapter 2).
1 Company profile
1.1 Introduction
This chapter gives a short introduction to the DSM corporate organization and the main parts of DSM that play a central role in this research: DSM Food Specialties and DFS Demand and Supply Chain Management.
1.2 DSM
DSM is an international group of companies with end markets that involve applications such as human and animal nutrition and health, cosmetics, pharmaceuticals, automotive and transport, coatings, housing and electrics & electronics (E & E). The group has annual sales of approximately EUR 8 billion and employs approximately 25,000 people around the world. The headquarters of DSM are located in the Netherlands, with locations in Europe, Asia, South America and North America. DSM has a decentralized organizational structure built around four clusters that are empowered to carry out all business functions. These clusters are: Life Science Products, Performance Materials, Industrial Chemicals and DSM Nutritional Products (a recent acquisition).
The activities of the Life Science Products cluster are mainly targeted at the pharmaceutical, food and agrochemical industries. The products of the Performance Materials cluster are used in a wide variety of end-use markets, such as the automotive sector, the electronics industry, coatings and building construction. The Industrial Chemicals cluster consists of capital-intensive businesses that are situated at the beginning of the value chain and require the use of large-scale production facilities. The industrial chemicals products include for example fertilizers. DSM Nutritional Products is a supplier of vitamins, carotenoids (i.e. colorants and antioxidants) and other biochemicals and fine chemicals to the food, feed, health and cosmetics industries.
Life Science Products
Performance Materials
Industrial Chemicals
DSM Nutritional
Products DSM
Figure 1: Organization chart DSM
1.3 DSM Food Specialties
Each of the above-mentioned four clusters consists of several business groups. This research is done at one business group of the DSM Life Science Products cluster. DSM Life Science Products consist of the following business groups:
− DSM Fine Chemicals
− DSM Pharmaceutical products
− DSM Anti-Infectives
− DSM Food Specialties
− DSM Bakery Ingredients
This research focuses on DSM Food Specialties (DFS). DSM Food Specialties produces and markets value-added ingredient solutions for the international food and beverage industries. The products include for example enzymes, yeast extracts, cheese and meat coatings and antibiotic residue tests. These products are mainly manufactured with the aid of fermentation and enzyme technology. DFS comprises five business units (see figure 2). Each business unit consists of several Product Market Combinations (PMC’s). In total there are 14 PMC’s as shown below the business units in the figure. The PMC’s ARA and Peptopro are separate entities next to the
‘regular’ PMC’s, due to their importance for the company.
DSM Dairy Ingredients supplies a range of ingredients and additives for the dairy industry, such as starter cultures, coatings for cheese and dairy product tests. DSM Savoury Ingredients is a producer of ingredients for flavourings and flavour enhancers (such as yeast extracts) used in products such as soups, sauces, instant meals and snacks. DSM Enzymes produces all enzymes. This includes for example enzymes used in dairy products or alcoholic beverages.
DSM Functional Food Ingredients produces ingredients for baby food, food supplements and functional foods. DSM Ingredients Development focuses on the development of new, innovative food ingredients. The main production sites of DFS are in Seclin (France), Capua (Italy), Delft (Netherlands), Zaandam (Netherlands) and Moorebank (Australia). All sites mentioned in this paper are displayed on the map in appendix D.
The core products, their markets and the main technologies that are used at DFS are displayed in figure 3. Here it is shown that the same product category can be used in several markets. Yeast extracts for example are used both in Savoury as in Beverages.
Figure 3: Core technologies, products and markets of DFS Core
Markets Core Technologies
Core Products
Fermentation Technology Enzyme Technology
Enzy- mes
Cultures Test kits Yeast extracts
Flavou- rings
PUFA’s, β- carotene
Dairy Cheese Yoghurt Desserts Ice cream
Savoury Sauces Soups Snacks RTE Meals
Beverages Juices Wine Beer Alcohol
Nutrition Infant Sports Special Health DSM Food Specialties
Finance, Control & ICT Research & Development
QESH & M
Legal Affairs
Demand & Supply Chain Management Human Resources
Strategy & Marketing Services
Dairy Ingredients
Test Cultures &
Media Preservation
Savoury Ingredients
Yeast extracts Processed Flavours AMC/ Diacetyl Wine Yeast
Functional Food Ingredients
Probiotics Enzymes
ARA
Dairy Enzymes Nutritional Enzymes Beverages Enzymes Baking Enzymes
Ingredients Development
Peptopro
Figure 2: Organization chart DFS
1.4 Demand and Supply Chain Management
Supply chain management is in theory been described as being involved with the integrated management of the flow of goods and information throughout the supply chain, in order to make sure that the right goods are delivered at the right place and quantity at the right time (Giannoccaro and Pontrandolfo(2002)). This is basically what the Demand and Supply Chain Management (DSCM) department at DFS does. DSCM is supporting DFS in activities concerning purchasing, planning, logistics and customer service. DSCM has been set up in the beginning of 2004, in order to achieve greater overall efficiency. Before 2004 the DSCM activities were done for each business unit separately. The mission of DSCM is to provide the Business Units and their customers with a reliable and cost-effective demand and supply chain.
The DSCM department of DFS (hereafter called DSCM) manages approximately 4000 different products and approximately 8000 different materials. They are stored in more than 40 warehouses all over the world, counting up to more than 100 inventory points (including those of tollers1). DSCM is currently decreasing the number of inventory points. The DSCM head office is situated in Delft (Netherlands). There are also some local offices in, for example, Australia and the U.S. However, DSCM activities will be more and more centralised in Delft in the upcoming year. Approximately 120 employees are working at the DSCM department. The organizational structure of the department is depicted in the figure below. As shown in the figure, the Supply Chain Planning department works next to (and together with) Customer Service, Purchasing and Logistics.
1The word ‘tolling’ used in this thesis refers to ‘outsourcing’. A ‘toller’ is the party that performs the outsourced activity.
Director DSCM
Supply Chain Planning Manager
Secretary
Logistics Manager Global Purchasing
Manager Customer Service
Manager
See appendix A
Figure 4: Organization chart DSCM
Within the Supply Chain Planning department each PMC has a Demand Chain Planner that is responsible for all planning activities within a PMC, including the consultation of the other three departments. One person can be the Demand Chain Planner of multiple PMC’s. The distribution of the 14 PMC’s over the Demand Chain Planners is shown in appendix A. Please note here that appendix A is a representation of the functional relationships within the department. The hierarchical organization chart looks slightly different.
Each Demand Chain Planner is supported by a Demand Forecast Analyst (DFA), a Supply Planner (SP) and an Inventory Planner (IP). This is shown in figure 5. This is however the distribution of activities and not of functions: sometimes the Demand Chain Planner also takes care of the DFA, SP or IP activities.
The DSM Corporate organization developed a Demand Chain Management organization design, depicted in figure 6. DSCM is currently in the middle of the implementation process of this organization design. According to this design the main responsibility of Demand Chain Management is to align the conflicting interests between demand and supply. Whereas the sales department pursues flexibility and high Customer Service Levels (CSL’s), the manufacturing department pursues an uninterrupted order flow resulting in cost optimisation and a continuous and stable production process. This conflict of interests needs to be balanced by Demand Chain Management. This is done during the Sales and Operations Planning (S&OP) meeting.
The new organization design also includes the implementation of SAP, an enterprise resource planning (ERP) software. The SAP system however still needs refinement in the defined system structure and the reliability of the parameters used.
Demand Chain Planner
Supply Planner Demand Forecast
Analyst
Inventory Planner
Other departments:
Customer Service Global Purchasing
Logistics
Figure 5: Supply chain planning
Strategic
Define Supply Strategy based on Agreed Customer Service Levels
Tactical
Supply Planning Inventory Planning Sales &
Operations Planning Proposed
Demand Plan Demand
Planning Proposed
Supply Plan Proposed Inventory Plan
Agreed Demand Plan Agreed Inventory Plan Agreed Supply Plan Operational
Order Processing
Capacity/
Production Planning Master Production
Scheduling Demand Management Inventory
Planned receipts Planned issues
Production orders
Figure 6:The demand chain planning process
2 Problem definition
2.1 Background
In addition to the organizational design changes described in the previous chapter DSM Food Specialties wants to achieve a drastic reduction of inventories. The DSM Corporate organization instructed its business groups to reduce their Operational Working Capital (OWC, calculated as accounts receivable, plus inventory, less accounts payable). DSM Food Specialties and DSM Corporate organization have agreed on a target where the OWC contains 25% of the third parties sales. Within this target they set the inventory level target on 20% of the third parties sales. Since early 2004 a project has started to reduce inventory levels. Since then the inventory levels have decreased with 34%. The aim for this year is to further bring down the inventory levels with 26%
compared to the inventory levels of the beginning of this year. Note though that bringing down the inventory levels should not be a goal on itself. Even though for some products this might be very well possible, for other products this can lead to significant problems. Consider for example lost sales to important customers or production shut down because a basic material is not available.
Lower inventory costs will in that case result in higher overall costs, which is obviously not desirable. The actual inventory levels should therefore rather be compared to target inventory levels. To know these target levels and to know the effect of bringing down certain inventory levels, insight is needed in the decisions that lead to inventory policies. This is however a problem within DSCM. There is no decision process defined to determine inventory policies. The inventory policies that are used are based on rules of thumb and experience of the managers involved.
There is no clear view if these policies match the optimal inventory required. This research therefore focuses on the definition of an inventory policy decision process. There are several advantages of a clearly defined, or formalized, decision process. First of all it makes clear why the inventory levels have specific targets, since these targets are obtained by going through several decision steps. In other words; there is a clear reasoning behind these targets. When the targets are set based on a formalized decision process, it is clear how the inventory levels can be brought down, for which products this is possible and, moreover, what are the effects on issues like for example customer service. A decision process allows furthermore what-if simulation before any policy changes are implemented. Based on the decision process, responsibilities of decision making can be assigned to people within the organisation. By knowing exactly what decisions need to be made it is clear as well which information is necessary to make these decisions. This allows better use of information systems, like SAP.
2.2 Research objective and main question
The previous paragraph described the absence of a formalized decision process that leads to inventory policies within DFS. It also described the advantages of a defined decision process.
This research examines the issues that should be regarded to make inventory policy decisions. It describes the decision process, with a special focus on the situation of DFS. The objective of this research is described as:
Develop a decision process for selecting inventory policies at DFS.
Before continuing with describing the research questions the two central concepts that are used within this research need to be defined. First of all, what is an inventory policy? An Inventory policy is the outcome of decisions made by Inventory Management. In general Inventory Management is regarded to determine two things: how much to order (the size) and when to place the order (timing). These are often described as the basic or routine decisions of Inventory Management. (see for example Stock and Lambert (2001), Vollmann et al (1997), Slack et al (1995), Peterson and Silver (1979), Fogarty et al (1991)) This research considers inventory policy in a broader sense however. Here Inventory policy includes not only determining what to order (what), determining the size (how much) and timing (when) of orders, but also the placement of products (where to have inventory points).
What leads to these decisions of ‘what, how much, when and where’ is the decision process. A decision process basically consists of set of decision rules (what decisions need to be made).
The decisions require input (information where the decision is based on) and result in output (what is the outcome of the decision made). This is depicted in the process model below. To develop a decision process the decision rules need to be defined. A decision rule can be defined as a function for selecting an action (Giannoccaro and Pontrandolfo (2002)). The type of decision rules within the decision process are based on the outcome. The output of the decision process for inventory policies consists of the answers to the ‘what, how much, when and where’ questions stated above.
The first part of the objective of this research is to develop a decision process. From the definition above this means that the decision rules need to be defined. The second part of the objective is to develop a decision process that is particularly suitable for DFS. This means that the decision rules that are selected are adapted to the situation of DFS and can be implemented within DFS.
This results in the following research question:
Input Information to
make the decisions
Decision process Decision rules
Output What, how much, when
and where Figure 7: Process model of a decision process
What decision rules determine an inventory policy for products of DFS and how can these be implemented within DFS?
It was already explained that the decision process consists of a set of decision rules. What decision rules are suitable is influenced by the context the decisions are made in (see the conceptual model below). Inventory policy decisions take part of supply chain management. The context of the decision process consists thus of supply chain characteristics. The supply chain is referred to here from a broad organizational perspective, and not only from the inventory management perspective. Zomerdijk and De Vries (2003) illustrated that including the organization perspective (and thus taking a broad view) on inventory control leads to better inventory policy decisions. An example of a supply chain characteristic is therefore also the organization’s strategy. Another example is the production process involved. Before the decision process can be developed, the elements of the supply chain (the supply chain characteristics) need to be defined and described for DFS in particular. Only after this is clear, the set of decision rules can be defined. This results in a decision model. The next step is to determine how the decision model can be implemented within DFS. This results in the following sub questions:
A. What supply chain characteristics influence inventory policy decisions?
B. What are the supply chain characteristics of DFS?
C. What are the decision rules that can be used?
D. How can the decision process be implemented within DFS?
Figure 8: Conceptual model
Decision process Context
set of decision
rules
2.3 Research methodology
There are several types of research. It is important to define which type of research is concerned here to determine the appropriate research methodology and research steps (De Leeuw (2001, p.
69)). This research can be described as a practical research. Practical research focuses on the need for knowledge of a specific party, in contrast to scientific research that focuses on the general need for knowledge. This research can furthermore be characterised as problem solving.
A problem solving research results in a complete representation of all necessary knowledge to solve a specific situation. This can be represented in a system design or in recommendations (De Leeuw (2001)). A typical problem solving process has three main phases, displayed in figure 9.
The first phase, the diagnosis phase, involves describing the current situation and evaluating it.
It is important here to look at the situation from multiple perspectives. Part of the description of the current situation has already been described in chapter 1 and paragraph 2.1. Chapter 1 described the organisation, based on internal documents and interviews with DFS representatives. This chapter described the problem background based on interviews with the Demand Chain Planners and with the Supply Chain Planning manager (head of the DSCM department). The next three chapters (chapter 3-6) complete the diagnosis phase. First, research question A is answered by defining the supply chain characteristics. This is done with help from literature. Then research question B is answered by describing the supply chain characteristics for DFS. This is done based on case studies that were developed with help from internal DFS documentation and interviews.
Figure 9 Problem solving approach (source: De Leeuw (2001) p. 182) Determine
direction
Evaluate
DESIGN DIAGNOSIS CHANGE
Multiple perspectives
Change Design Evaluate
Describe
The second phase (chapter 7 - 10) is the design phase. In this phase the decision process is designed. This involves the definition of the decision rules, resulting in an answer to research question C. The design is based on information both from theory (literature) and practise (the case studies from the previous phase).
The third and last phase (chapter 11-12) involves the implementation of the developed decision process within the DSCM organisation, answering research question D. This includes determining where the necessary information should come from and who should be the decision taker(s) and when (or how often) the decisions are to be made. This research results in tailor made recommendations for DSCM on this subject. After the results are implemented, DSCM should keep evaluating the design and, if necessary, make adaptations. This is however, as will be described in the next paragraph, not part of this research.
2.4 Deliverables and scope setting
The decision process is presented in a framework: the decision model. This framework takes the decision taker(s) through the necessary steps, making sure all concerned issues are taken into account in order to obtain an inventory policy. The framework should be adapted to the DFS organisation. The research performed should not take longer than six months. In order to perform the research within this timeframe the research is limited in several ways. First of all, the focus of this research is mainly on the organisational and structural side of the decision process. It shows what decisions need to be taken, by who and with help from which information.
This research does not include the calculation of inventory policies for specific products. The development of detailed calculation methods (formula) is not included either. This research does include tailor made recommendations on the implementation of the decision model within DFS. The actual implementation and evaluation of the decision model are not included in this research. DSM Food Specialties has a great number of products and supply chains and each chain has its own dynamics and complexity. This research is limited to the supply chains connected to the warehouses in Australia and New Zealand. This area is, compared to other areas of DFS, not too extensive or complex, but still contains all elements of concern for this research (production facilities, warehousing, customers and so on). Since all elements are present at these warehouses, it is still possible to use the results on the other areas of DFS as well. However, even the warehouses in Australia and New Zealand contain almost 350 different products. This research does not try to make an individual picture of the supply chains of each of these 350 products. Instead a smaller number of products is selected that represent general DFS products. The research done for the selected products is generalized a later phase of the research.
The wanted reduction of inventory levels did initiate the problem statement as was shown in the first paragraph of this chapter. It is however not considered a requirement for the research. This means that the decision model does not necessarily have to lead to inventory reductions for all products.
Whether SAP is a suitable information system will not be questioned here. The management of DSM has already made a choice for SAP and this is not likely to be changed.
Any company sensitive information (like product names, costs or sales volumes) collected during this research will not be described in this thesis or will be presented less specific (sales volumes in percentages of total sales, products are made anonymous).
Part II Diagnosis phase
The first part of this thesis introduced the basic elements of this research. First an introduction to the company was given. The research objective, the research questions, the scope setting and the research steps were described. There were three research phases defined in chapter 2. This part of the thesis discusses the first phase, the Diagnosis phase. Chapter 3 describes the methodology that is used in this phase. It describes as well the definition of the supply chain characteristics that were presented in the previous chapter. Based on this definition the supply chain characteristics are described for DFS in chapter 4 and 5.
3 Diagnosis methodology and definitions
3.1 Introduction
As explained, the Diagnosis phase contains a description of the supply chain characteristics of DFS that influence the decisions that need to be made to obtain inventory policies. In order to do so, the supply chain characteristics need to be defined. This is done in paragraph 3.2. With help from current literature five groups of characteristics are defined. The next step within the diagnosis phase is concerned with describing the supply chain characteristics for DFS. This is done with help from a selection of products that reflect ‘general’ DFS products. These products are used as case studies to represent the DFS context in general. This chapter describes the selection criteria that are used and the products that result (paragraph 3.3).
3.2 Supply chain definition
Here the supply chain is defined as “the internal supply chain that integrates business functions involved in the flow of materials and information from inbound to outbound ends of the business”
(Harland (1996)). It is important to look at the supply chain from a broad organizational perspective, and not only from the inventory management perspective. “Inventory management objectives, policies, and decisions should be consistent with overall organizational objectives and should be consistent with marketing, financial and manufacturing objectives” (Fogarty et al.
(1991)). Besides, as explained by Peterson and Silver (1979) questions concerning the inventory policy issues are difficult to answer because “they involve overall company policy and cannot be answered from the limited point of view of inventory management or cost accounting alone”.
Therefore it is important that the set of characteristics is not taken too small. They are supposed to reflect all aspects that are directly and indirectly related to the inventory policies. Besides the traditional points of attention in inventory management literature, such as order quantities and replenishment strategies, other aspects need attention as well. To make sure all relevant elements are included in this research the theory of Brevé (1990) is used. Brevé (1990) distinguishes five groups of characteristics and several elements within these groups that can be used in a logistical diagnosis (and for the supply chain analysis as well). The five groups of characteristics are:
A. General characteristics
These characteristics consist of basic information about the company and give a global view of the current situation of the company. The general characteristics shape the context of the characteristics in the other groups.
B. Product characteristics
This group of characteristics involves all aspects that are typical for the products of the organization, including raw materials and intermediates.
C. Production and process characteristics
This group of characteristics involves all information that is concerned with the flow of goods.
D. Planning and control characteristics
The planning and control characteristics give more information about the management of the flow of goods, the planning processes that are used, the information systems involved and other related items.
E. Purchasing characteristics
Several aspects, like the production process, the product quality and the flexibility of the company, are greatly influenced by the suppliers. This includes elements like quality of the purchased (raw) materials, the price and the flexibility from the suppliers.
Appendix B gives a detailed overview of the elements in each of the five groups that are used to describe the supply chains of DFS. This list of elements is used in interviews to obtain the information to describe the supply chains of DFS. The General characteristics of DFS are described in chapter 4. The other four groups of characteristics are shortly described in chapter 5.
A more elaborate description of these characteristics can be found in appendix C. The characteristics described in chapter 5 are based on case studies on a selection of six products.
The selection criteria to obtain these six products are described in the next paragraph.
3.3 Product selection criteria
The supply chain characteristics are described based on a selection of products. This paragraph describes these products and the criteria that are used to obtain the selection. Since the focus of this research is on inventory policies it is important to select products that resemble different situations with probably also different inventory policies. According to Tersine (1994, p. 9-12) inventory situations can be classified based on five characteristics. These are the repetitiveness of the inventory decision, the source of supply, the knowledge of demand, the knowledge of lead time and the type of inventory system. These characteristics are used as a basis here to make the product selection.
The products that are selected are all of a repetitive nature: they are not ordered only once.
Products that represent single orders are not taken into account. By doing so the selected products are more representative for all DFS products. Another advantage is that there is more information about for instance demand patters and inventory levels.
Tersine divides the source of supply into outside and inside supply. Here a broader view is taken by looking at the entire supply chain. There is a distinction made between products with a relatively clear and simple supply chain and products with a more complex one. Complex chains are here defined as chains with more than two stages and/ or with highly divergent points in the chain.
The ‘knowledge of demand’ is another criterion. Here three possible demand patterns are defined: (1) constant demand (with a clear trend), (2) seasonal demand and (3) lumpy demand (demand can be zero for specific periods and then go up significantly, without showing any pattern at all). The plants at Australia and New Zealand are using SAP since 2004. There is no information about demand of the periods before 2004. Therefore the determination of the demand patterns of the products is mainly based on the knowledge of the Demand Chain Planners.
Tersine divides the knowledge of lead time into constant and variable lead time. Lead times are usually constant within DFS. This characteristic is therefore not chosen as a selection criterion.
Finally, regarding the type of the inventory system, a distinction is made between ‘make to order’
(MTO) and ‘make to stock’ (MTS) products.
It is clear that there can be thought of more characteristics in order to make a selection of products. However, choosing more selection criteria results also in more possible combinations of characteristics. This results again in a greater number of selected products, which is obviously not the goal of the selection process. The criteria now chosen results in twelve possible combinations (see figure 10).
Figure 10: The selection criteria
1 2 3 4 5 6 7 8 9 10 11 12 Demand:
a. lumpy b. constant c. seasonal Supply chain d. simple e. complex Inventory system f. MTS g. MTO
a b c
d e d e d e
f g f g f g f g f g f g
The selection criteria were discussed during a workshop with all Demand Chain Planners involved with products in Australia and New Zealand. The first criteria, the demand pattern, divided the products into three groups. These three groups consisted of three Demand Chain Planning areas (see for the different areas the organization chart in appendix A). These areas are: Cultures and Media, Process Flavours & Yeast extracts and Wine Ingredients. The Cultures and Media showed a lumpy demand pattern, the Process Flavours & Yeast extracts showed a constant demand pattern and the Wine Ingredients were seasonal. As said, each area has a different Demand Chain Planner. Therefore the selection process continued in more detail with these Demand Chain Planners individually. This resulted initially in a selection of ten products.
After a first analysis of the supply chains of these ten products, followed by a second workshop with all involved Demand Chain Planners, it was shown that several products showed a lot of resemblance in their supply chains. Therefore the product selection was further brought down to five products. However, still one group of products was missing that represents a large part of the total product portfolio of DFS: the enzymes. Therefore a sixth product has been added to the product selection, which is an enzyme. These six products and their place within the selection criteria are presented in table 1. For these products it is expected that their supply chains are representative for ‘average’ DFS supply chains. The research done on these products can therefore be generalized at a later stage within this research and are useful for whole DFS.
There are no products found with a lumpy demand pattern that have either relatively simple chains (combination 1) or a make to stock inventory system (combination 3). No products with constant demand were found that were make to order (combination 6 and 8). As can be seen the combination numbers 10 and 12 have not been found either. This is due to the fact that there are no make to order products for seasonal products.
Alias Criteria combination Demand Supply chain Inventory system
CDT 4 Lumpy Complex MTO
SP 5 Constant Simple MTS
SG 7 Constant Complex MTS
WF 9 Seasonal Simple MTS
WR 11 Seasonal Complex MTS
Enz 7 Constant Complex MTS
Table 1: The selected products and their criteria combination
4 Results: General characteristics
4.1 Introduction
This chapter describes the first group of the supply chain characteristics: the general characteristics. These characteristics consist of basic information about the company and give a global view of the current situation where the company is in now. These characteristics are the same for all products. They shape the context of the characteristics of the other groups. The general characteristics are divided into the following elements (see also Appendix B):
− General information about the product
− General information about the market
− Production typology (what kind of industrial process are we dealing with)
− Total amount of products
− Organisational structure
− Strategic decisions (performance measures)
The first five elements are described in the following paragraph. Paragraph 4.3 discusses the performance measures that are linked to the strategic decisions of DSM Food Specialties.
The information described in this chapter is obtained using internal DFS documentation and by interviewing company representatives. These were mainly the head of the DSCM department and one of the Demand Chain Planners, Alex Pier, who supervised this project.
4.2 DFS products and organisational structure
The organizational structure of DFS and DSCM was already described in chapter 1 and will not be repeated in detail here. However, some of the focal points need special attention.
DSM Food Specialties produces and markets value-added ingredient solutions for the international food and beverage industries. DSM Food Specialties is, as the name already suggests, active in a specialty business. This means: many different products, many customers and low volumes. There are approximately 4000 different products and 8000 different materials.
Most of the finished products are Make to Stock.
DFS has a decentralized organizational structure. The Demand and Supply Chain Management (DSCM) department at DFS is supporting DFS in activities concerning purchasing, planning, logistics and customer service. The Supply Chain Planning department works next to (and
together with) the Customer Service department, the Purchasing department and the Logistics department. Within the Supply Chain Planning department each PMC has a Demand Chain Planner that is supported by a Demand Forecast Analyst (DFA), a Supply Planner (SP) and an Inventory Planner (IP). Each Demand Chain Planner can be responsible for several PMC’s.
For all plants involved forecast, inventory planning and supply planning is done. Depending on the PMC the planning procedures are more or less intensively supported by SAP. The planning processes are usually slightly different for each Demand Chain Planner. The detailed descriptions of these processes can be found in Appendix C. However some basics are similar and are described here. In the figure below the working areas of supply planning, inventory planning and forecast analysis are shown.
The forecast analyst a demand planning based on forecasts of the independent requirements of end products. There are several forecasting methods that can be used. They are described for example by Brown (1977) and Silver et al. (1998). These methods are not discussed in detail here. Within DSCM the possible forecasting methods are already studied and evaluated on their usefulness. The most suitable methods are implemented within the organisation. Each month there is a forecasting-round (with exception of seasonal products, where the forecasting is done twice a year). First an unrestricted forecast is developed by the forecast analyst. This forecast is based on historical information. The historical information that is used only takes the actual sales into account. This way of forecasting thus disregards the lost sales. There is a separate forecast done for A-customers (the most important customers) and all other customers (called collaboration customers). The unrestricted forecast of the A-customers is send to the concerned Sales Managers. The Sales Manager can adapt the planning if necessary, based on experience or on already made deals with customers. Within a week the DFA receives the (adapted) forecast
Figure 11: Working areas in the planning process Production
plant
Working area of inventory planning
Independent requirement
As a result of forecast STO
independent requirement
Order
Working area of supply planning
Local ware house Central
ware house
back from the Sales Managers and enters them into SAP (this is called the sales proposed). After the S&OP meeting it is decided which of the sales proposed are turned into sales agreed.
The inventory planner takes care of the inventories and determines the dependent requirements between the plants. The Inventory planner is mainly concerned with the local warehouses. For each warehouse the Inventory Planner can develop a list (with help from SAP) that shows the coverage that is still available at a warehouse for each product (the issue of coverage returns in chapter 4). Based on the urgencies or special exceptions the Inventory planner decides what should be replenished. This leads to a replenishment proposal and eventually in a STO: a stock transfer order. A significant part of the DFS products suffer from low forecast accuracy. This complicates the work of the Inventory planner, since the coverage numbers are based on the forecasted demand.
The supply planner is on his part in charge of the supply of finished products and intermediates from the central warehouses. The local production plants take care of the detailed production planning.
The Demand Chain Planner has chain responsibility and is responsible for the whole process.
Every month there is a Pre-S&OP and a S&OP (Sales and Operations Planning) meeting. For some PMC’s these meetings are called differently, but the main goals are the same. At the Pre- S&OP the Demand Chain Planner forms a proposal for the Sales and Operations Planning meeting (S&OP) together with the Supply Planner and the Production planner(s). During a Pre- S&OP the consequences of the forecasted demand on inventory and production as well as possible obstacles will be discussed in detail. The current situation is evaluated, keeping in mind the goal of optimal chain performance. In principle the meeting is concerned with balancing supply and demand.
The S&OP meeting is planned one week later than the Pre-S&OP. During the S&OP the situation is discussed in less detail and definitive decisions are made. The attendants of this meeting include the Business Manager (who carries overall responsibility of the PMC), the Demand Chain Planner (who carries the responsibility of the chain performance of the PMC, including the inventory levels), the Production planner and the Sales manager.
4.3 Performance measures
Chapter 1 described the mission of DSCM as “to provide the Business Units and their Customers with a reliable and cost-effective Demand and Supply Chain”. How this is translated into performance measures is explained in this paragraph. A key issue in inventory management is
determining the performance levels for the inventory system (Vollmann et al. (1997)). Setting the performance levels influences the choice for specific inventory policies. There are several measures to determine inventory system performance. Examples of the multitude of measures that are most used by organizations are (CSC/ University of Tennessee (1999)):
Evidently the same company does not use all of these measures at the same time. Each company should make a selection of the measures that best match with the characteristics and needs of the company (Stock and Lambert (2001) p. 607). Note that next to measuring the performance at one specific moment, there should also be looked at the development of the performance over time. Armstrong (1985) calls this the combination of ‘snapshots’ and ‘moving pictures’. The development over time illustrates possible trends or historical relationships and the influence of past decisions. This can help management to better understand the situation and make better inventory decisions. The performance measures currently used at DSCM that effect the inventory system are described below.
Operating Working Capital (OWC)
Logistics management is fundamentally linked to the working capital requirement within the business (Stock and Lambert (2001) p. 679). As explained in chapter 1, DFS is currently striving for a drastic reduction in inventories in order to reduce the OWC. The OWC is calculated as accounts receivable plus inventory less accounts payable:
AP Inv AR
OWC = + −
With:
OWC = Operating Working Capital AR = Accounts receivable Inv = Inventory AP = Accounts payable
− Outbound freight costs
− Inventory count accuracy
− Order fill
− Finished goods inventory turns
− On-time delivery
− Customer complaints
− Over/short/damaged
− Out-of-stocks (of finished goods)
− Returns and allowances
− Line item fill
− Inbound freight cost
− Back orders
− Inventory obsolescence
− Order cycle time
− Incoming material quality
− Overall customer satisfaction
− Inventory carrying costs
− Logistics cost per unit versus budget
− Invoice accuracy
