FREEZING THE FLUCTUATING DELIVERY RELIABILITY
Researching the causes of the fluctuating delivery reliability in order to diminish waste in a heat-exchanging company.
Author: M. Pronk
Student number: 1578790
1st supervisor: Dr. G. A. Welker
2nd supervisor: Prof. Dr. J. Wijngaard
Company supervisor: Drs. J. ter Doest
Key words: Delivery reliability, due date assigning, departmental control Copyright 2007 by Martijn Pronk
University of Groningen Alfa Laval Groningen
PREFACE
This thesis represents the end of my master study Operation & Supply Chain at the university of Groningen. Although it was my initial intention to study Change Management I got acquainted by the interesting topics covered by the course Operations Management. This course made me decide to broaden my knowledge in the field of Operation and Supply Chain. Until today I never regretted the decision I made to change my objectives.
The end of a period is also the beginning of a new one. Curious on what this new period has to offer I am not afraid to say goodbye to my student-days. Before doing so, I would like to thank some people who helped me to write this “crown” on my efforts to gradate.
First of all, I would like to thank all discussion partners of Alfa Laval Groningen for their time and their pleasant approach in telling me the ins and outs of their functional tasks. Their kindness, patience and efforts helped me to integrate into a special company community. Also, I definitely would like to thank Gera Welker for her time and efforts spend, on reading and placing critical comments at my concepts. Her flexible and timeless way of guidance, certainly helped me in writing this thesis. Next to that, I would like to thank Jacob Wijngaard for his critical review at the end of my thesis trajectory. Furthermore, I would like to thank Joost ter Doest for shaping my thoughts and helping me to develop this thesis. Our timeless conversations triggered me to adopt different points of view. Finally I would like to thank my parents for my carefree study time. Their unconditional support and love, within my life but especially during my seven years of study, helped me to develop myself to the person I am today.
EXECUTIVE SUMMARY
Different products at Alfa Laval Groningen, further referred to as ALG, are processed through two different order cycles. These order cycles are set up according to lean philosophies and are aiming to fulfil high customer demands. High customer demands results in many product variations in low volumes to produce. Not ideal considering the fact that leanness operates at best in a low demand variation and high volume environment. The high variation in demand combined with inflexible, but efficient, production processes causes negative influences on the delivery reliability. Constant measuring showed that the delivery reliability fluctuates between 60% in busy periods and 100% in calm periods. The fluctuating delivery reliability leads to control measurers such as:
Control activities between sales and production. Extra transport costs due to split delivery
Rush-orders in order to make up lost time of the previous production processes disturb the production line.
Subsequent delivery of components.
These control measures can be seen as forms of “waste” of the order cycles. In order to remove these kinds of wastes the following main research question has been developed:
What are the main causes of the fluctuating delivery reliability and how can these causes be overcome?
In order to answer this main research question it was presumed that the delivery reliability of an company is determined by due date assigning and internal control. With this premises the order cycles of ALG were analyzed. This analysis indicated that the assigned due dates at ALG are too uncertain and that the different departments of ALG are not fully controlled. Together these factors make it impossible to generate a reliable delivery performance, especially in busy times. Further analysis in the uncertainty of the assigned due dates turned out that this uncertainty is fed by the following causes:
1.) Lack of insight in the component availability of non-configurable products at the due date assigning.
2.) Lack of insight in the capacity of the different processes required to manufacture the product.
3.) The lack of a central and formalized feedback function about the due dates.
delivery performance. The causes of the unsatisfactory level of internal control are summarized below:
4.) More orders are accepted and released than the order processing capacities can handle
5.) There is no detailed planning for the order processing 6.) The capacities of the order processing are relatively fixed
7.) The planner often releases more orders than production capacity can handle
8.) The mix-flexibility of the production capacities is too low to vary capacity on the short term
9.) Finished products are often kept on stock for more than seven days. Next to the causes of the uncertain due date and the unsatisfactory level of internal control it was found that the delivery reliability was measured wrongly. This results in a kind of blindness for the actual delivery reliability considered by the customer. The causes for the wrong measurement are:
10.) Wrong point of measuring the delivery reliability
11.) Wrong variables used to measure the delivery reliability
TABLE OF CONTENTS
TITLE PAGE PREFACE
EXECUTIVE SUMMARY
PART I INTRODUCTION
CH 1 ALFA LAVAL GRONINGEN
§ 1.1 Facts & Figures 9
§ 1.2 The organization 9
§ 1.3 The Product & order cycles 10
§ 1.4 The customer order decoupling point 12
§ 1.5 Conclusion 13 CH 2 RESEARCH DESIGN § 2.1 Problem statement 14 § 2.2 Research objective 15 § 2.3 Research question 15 § 2.4 Research model 15 § 2.5 Conceptual framework 16 § 2.6 Sub-questions 18 § 2.7 Theoretical framework 18 § 2.8 Overview research 23 § 2.9 Conclusion 24
PART II CONCRETIZE PERFORMANCE
CH 3 DELIVERY RELIABILITY
§ 3.1 Making delivery reliability operational 26 § 3.2 Relative importance delivery reliability 27
§ 3.3 Delivery reliability ALG 27
PART III DIAGNOSE PERFORMANCE
CH 4 ORDERPROCESSING
§ 4.1 Concretization order processing configurable orders 32 § 4.2 Due date assigning configurable orders 34 § 4.3 Internal control at processing configurable orders 34 § 4.4 Order processing non-configurable orders 35 § 4.5 Due date assigning non-configurable orders 38
§ 4.5.1 Requested due date 38
§ 4.5.2 Promised due date 38
§ 4.5.3 Planned due date 39
§ 4.5.4 Discussion 39
§ 4.6 Internal control at processing non-configurable orders 40 § 4.6.1 Work order acceptance/release decision 41
§ 4.6.2 Capacity detail planning 42
§ 4.6.3 Capacity assignment 42
§ 4.6.4 Priority dispatching 44
§ 4.7 Conclusion 45
CH 5 PRODUCTION & EXPEDITION
§ 5.1 Concretization production process 47
§ 5.2 Internal control production process 48
§ 5.2.1 Work order acceptance/release decision 48
§ 5.2.2 Capacity detail planning 48
§ 5.2.3 Capacity assignment 49
§ 5.2.4 Priority dispatching 50
§ 5.3 Concretization expedition 50
§ 5.4 Internal control expedition 51
§ 5.5 Conclusion 52
PART IV PERFORMANCE PROBLEMS
CH 6 CAUSES FLUCTUATING DELIVERY UNRELIABLITY
§ 6.1 Causes due date assigning 55
§ 6.2 Causes internal control 56
§ 6.3 Delivery reliability 58
§ 6.6 Conclusion 58
PART V REDESIGN
CH 7 IMPROVEMENTS
§ 7.1 Improve due date assigning 61
§ 7.2 Improve internal control 62
§ 7.3 Improve delivery reliability measurement 64
REFERENCE LIST APPENDICES
Appendix I: Legend with Actor Activity Diagram
Appendix II: Due date list determined on available production capacity Appendix III: Time till configurable order gets planned due date
CH 1 ALFA LAVAL GRONINGEN
This chapter serves as a first acquaintance with the products and the order cycle of Alfa Laval Groningen. The first paragraph discusses some facts and figures of Alfa Laval Groningen. In paragraph 1.2 the organizational structure is discussed. Paragraph 1.3 describes the products and the order cycles through which the products are processed. Paragraph 1.4 pays attention to the decoupling point of Alfa Laval Groningen. Finally, this chapter is concluded with a brief summary in paragraph 1.5.
§ 1.1 Facts & Figures
Alfa Laval Groningen BV (ALG), formerly known as Helpman BV, is on itself a relatively small organization with 120 employees and a turnover of 22 million Euros. The company exists of two production facilities, one in Bulgaria and one in Groningen. The company already exists more than 80 years and has faced many ups and downs. Due to its flexibility the company managed to survive the downs and celebrate the ups. Many different core products, through out the years, have contributed to this success. In its early years Helpman BV produced, for example, small bakery instruments and transport installations for flour mills, it even produced window- and doorframes. Nowadays ALG tries to strengthen the position of the Alfa Laval Group (Mother Company) by developing, producing and selling heat exchangers. The Alfa Laval Group is a world leader in heat transfer, centrifugal separation and fluid handling. Alfa Laval Group acquired Helpman BV in March 2007. Alfa Laval Groningen primarily concentrates on heat transfer markets within Europe. In this geographical area ALG tries to be an ideal partner and supplier for wholesalers and medium sized till large installers by producing high quality products and having superior know-how of the application areas.
§ 1.2 The organization
General Manager
Manager Sales & Marketing Manager Operations R&D HRM Finance Service After sales Sales Communication Credit control Shipping & Warehousing Purchasing &
Supply chain Production
Engineering Sheet steel unit
Maintenance
Figure 1: Organizational chart
§ 1.3 The product & order cycles
Alfa Laval Groningen BV designs, produces and sells heat exchangers. The heat exchangers are primarily used to cool, so called AGF products (Dutch abbreviation for: Potatoes, Vegetables, Fruit), meat and fish. Heat exchangers, produced at ALG, can be divided into coolers and condensers. Coolers are devices used to cool air in, for example, cold storage rooms. During this cooling process heat is produced, this heat is carried off by the condenser. The condenser often is placed on the rooftop of the cooled building.
The products of Alfa Laval are designed and produced in many different variants. To a great extent, the final application area determines the design of heat exchangers. Five different application areas can be distinguished. For each of these areas, ALG has defined a special priority series with air coolers optimized for this application:
- Shock cooling - Working areas - Cold storage - Frozen storage - Freezing rooms
HEAT-EXCHANGERS
Product family Family-type
Coolers - ZT
- TR - TX
Condensers - TC
Table 1: Heat exchangers subdivided according to their different families.
The difference between the products delivered to these two customer groups can clearly be seen in the size and capabilities of the products. The TX family, for example, is mainly designed and produced for the commercial market, where the TR family mainly serves the industrial customer. The production facility in Groningen only produces the products for the industrial customer, while the production of commercial products takes place at the Bulgarian facility.
Almost all products, manufactured at ALG, do have a standard design. This standard design can be adapted on customer demand. Due to this adaptation possibility the following different product categories are distinguished internally: (1) standard, (2) standard on request, (3) customer specified products and (4) Specials. Because these four product categories are quite important in creating an understanding of the ALG activities the definitions of these four product categories are stated below. These definitions are composed by a project team, which is trying to automate the order processing of the order cycle.
Ad 1. Standard product
A standard product, further to be referred as a standard, is a heat exchanger which can be configured by a salesman with the help of the Helpman Select1. These products
are catalogue products and do not need any engineering tasks/checks to be carried out. Inventories of components or a certain delivery lead-time need to ensure a maximal delivery time of 4 weeks.
Ad 2. Standard on request
A standard on request product, further to be referred as a SOR, is a heat exchanger which can be configured by a salesman with the help of Helpman Select. These products are catalogue products and do not need any engineering tasks/checks to be carried out. They however contain specific components which are normally not on stock. Before a due date is assigned the salesman thus first have to contact the purchase department.
Ad 3. Customer specified products
A customer specified product, further to be referred as a CSP, is a heat exchanger with customer specific add-ons which can be configured by a salesman in Helpman Select. Nevertheless this product category needs engineering tasks/checks to be performed.
Ad 4. Special
A special is a heat exchanger, totally build on customer demand and can not be configured in Helpman select. This product category needs engineering tasks to be performed. Delivery time becomes gradually known.
The four product categories are processed through two different order cycles. Both order cycles are presented together in figure 2.
Sales Work
Preparation Planning
Work
preparation Production Expedition
Engineering Work
preparation Engineering
Configurable (ad 1&2)
Non-configurable (ad 3&4)
Figure 2: Order cycles at Alfa Laval Groningen
According to the operation and logistics director, the sale of specials contributes to 1/3 of ALG’s turnover. The standard products, on their turn, take care of 2/3 of ALG’s turnover. Both specials and standards are delivered to commercial and industrial customers. In this research the order-cycles of Alfa Laval are divided into three different stages. In the first stage all non-physical production processes are performed at the so called order processing. In this stage a customer order is transformed into a production order. Once the transformation has taken place the second stage transforms the production order into a physical end-product. The third and last stage of the order-cycle is the expedition, which distributes the end-product to the customer.
Order processing Production Expedition
End product Production order
Customer order
Figure 3: Three stages of the order cycle
The three different stages are displayed in the figure 3 and will be used to structure this report.
§ 1.4 Customer order decoupling point
levels and irregular market demand pushes the decoupling point upstream. High delivery reliability and short delivery times have a downstream effect. (van Donk 2001). At the top of figure 4 the general production processes of ALG are displayed. Looking at the two order cycles of ALG it can be said that the decoupling point of the configurable products is defined as make-to-order (MTO and DP 3). The decoupling point of the non-configurable products is defined engineer-to-order (ETO and DP 4).
Supplier D.P. 4 D.P 3 D.P. 2 D.P. 1 Customer
Production based on customer oder Production based on forcasts
Upstream Downstream DP 4: Engineer-to-order (ETO) DP 3: Make-to-order (MTO) DP 2: Assembly-to-order (ATO) DP 1: Make-to-Stock (MTS) Engineering Component
manufacturing Assembly Distribution
ALFA LAVAL GRONINGEN
Figure 4: Customer order decoupling point concept (source: Hoekstra en Romme 1985)
Both decoupling points are positioned upstream. A decoupling point almost at the beginning of the order cycles means that most activities are performed on customer order. Although almost all activities are performed on customer order the process is accompanied with a lot of uncertainty. This uncertainty is a consequence of the drawing-process which gradually generates the material requirements. This is the reason why some activities can not be performed before certain previous activities have taken place. A clear example is the purchasing process, in order to maintain a certain stock level and prevent ALG from obsolete inventory this process can only be performed once engineering is finished. Next to this the customer is in certain situations permitted to change their requirements freely till the point where the order enters the physical production process.
§ 1.5 Conclusion
CH 2 RESEARCH DESIGN
This chapter discusses the problem of ALG and clarifies the objective(s) of this research. Both research components are mainly based on conversations with the manager operations and several internal reports. Next to that, this chapter also addresses the research model and conceptual model of this research. The research model discusses how this research is going to be performed. The conceptual model explains what exactly is going to be researched. After this explanation the theories used to form the conceptual model are clarified. This chapter is concluded with a brief summary, discussing the main facts dealt within this chapter.
§ 2.1 Problem statement
Manufacturing companies are increasingly forced to become more customer-oriented in order to fulfil higher market demands (being responsive) while remaining efficient (Welker, G.A., 2004). Fulfilling high market demands at Alfa Laval Groningen BV expresses itself in delivering an almost infinite product range, based on 6 family types, at reasonable delivery times. In order to remain efficient ALG primarily strives, amongst other things, for reducing the stock levels. This efficiency goal is the direct cause for the closure of a privately owned component warehouse and a better inventory turnover rate. The elimination of these inventories (wastes) mainly can be attributed to the change to lean manufacturing. Leanness operates at best in a low demand variation and high volume environment. Not an ideal situation considering the fact that ALG’s customers are permitted to make changes on standard designs. The uncertainty causes high variation in demand. This high variation in demand combined with inflexible, but efficient, production processes causes negative influences on the gauges of responsiveness. Delivery reliability at ALG is an important gauge of responsiveness. Due to constant measuring, ALG discovered that the delivery reliability keeps fluctuating between 60% in busy periods and 100% in calm periods. This is a problem because unreliable delivery of orders leads to unsatisfied customers and different kind of control measures. These control measures can be seen as forms of waste of the production processes and are amongst others:
Control activities between sales and production. Extra transport cost due to split delivery
Rush-orders in order to make up lost time of the previous production processes disturb the production line.
§ 2.2 Research objective
Taking the above into consideration it can be said that by improving the delivery reliability and keeping it at a steady level, ALG can improve responsiveness. At the same time ALG can also decrease the control measures taken to manage the arising problems due to the present delivery unreliability. In order to improve the delivery reliability the following research objective has been defined:
Explore the causes of the fluctuating delivery reliability at Alfa Laval Groningen to improve the delivery reliability, while processing the order efficiently through the entire organization.
In the next paragraph this research objective will be transformed into a main research question.
§ 2.3 Research question
To address the research objective the following main research question is developed:
What are the main causes of the fluctuating delivery reliability and how can these causes be overcome?
To answer this main research question the following boundaries have to be taken into consideration. These boundaries determine the scope of this research:
This research will be concentrated on the Groningen facility only. Solutions need to address the entire order cycle without sub
optimization and without decrease in efficiency.
This research is performed by order of the manager operations and logistics of Alfa Laval Groningen B.V. Recommendations need to be in his field of authority.
The research will be performed within 5 months.
The research will be performed according to the prescriptions of the Faculty of Management and Organization at the University of Groningen.
§ 2.4 Research model
Fluctuating delivery reliability
Due date
assigning Internal control
Insight in Capacity ordercycles Insight in Component Availability Insight in Capacity orderprocessing Insight in Capacity production process Internal control production process Internal control expedition Step 1 Concretization Step 2 Diagnosing Step 3 Performance Step 4 Recommendations
+
Internal control orderprocessingFigure 5: Research model with magnified conceptual framework
The point of departure for the OM-scan is the operational performance of the order cycle. Generally this performance is concretized with the help of the following performance objectives: Quality, Dependability, Speed, Costs and Flexibility. The first step of the OM-scan serves as a first acquaintance with these performance objectives and arranges them at relative importance for a specific organization. This step is generally used to set a direction for further research. By emphasizing delivery reliability ALG already set a direction for this research (see also § 2.1). Nevertheless it is still useful to perform this first step. During the preliminary research, several definitions of delivery reliability were found throughout the organization. To be able to draw conclusions and propose recommendations on the delivery reliability of ALG, delivery reliability has to be defined.
The second step of the scan represents the actual diagnosis. In this diagnosis the order cycle stages and the methods, used to achieve a certain delivery reliability are analyzed. The third step makes the performance objectives operational and brings forward possible causes of the bad performance. The fourth and added step tries to overcome the causes discovered in the third step through improvements of the processes analyzed in the second step. With the addition of this last step the OM-scan becomes a continuous improvement circle.
§ 2.5 Conceptual framework
framework should contribute to the possibility of investigation (de Leeuw, 2001). The conceptual framework is, together with the research model depicted in figure 5. The conceptual model, magnified in the separate box in figure 5, needs to be interpreted as follows. Due to constant measuring, ALG discovered that the delivery reliability keeps fluctuating between 60% in busy periods and 100% in calm periods, where a stable 95% is desired. According to Bertrand et al (1998) both due date assigning rules and the internal departmental control determine the delivery reliability of an organization. By taking both variables separately into consideration possible causes of delivery unreliability should become visible.
Due date assigning
Different authors (Rudberg and Wikner 2004, Hill 1989) posit that a high delivery performance involves considerations of capacity scheduling and inventory work in process and finished goods. With the consideration of these variables they aim at the assignment of a reliable due date. Assigning a reliable due date has a positive effect on the delivery reliability. Companies, like ALG, which do not build an inventory of end products, generally lack the ability to provide reliable completion dates to customers that are achievable, tight and computed in real time (i.e. a few seconds) for dynamic order arrivals. Instead, they can only offer rough estimated lead times (Moses et al 2004). Nevertheless, accurate order promising for new or changed demands is one of the most important issues for ALG. This places the organization in a delicate situation.
In this thesis the due date assigning process will be analysed with the help of a framework developed by Rudberg and Wikner. According to Rudberg and Wikner (2004) the due date assigning process must be based upon the planning constraints faced by a company. Planning constraints are constraints based on either material or capacity (or both). Given the fact that ALG practically holds no inventory, it is important to pay attention to both constraints. Without knowledge of the presumable delivery lead times of components or knowledge of the available capacity an assigned due date is almost worthless. Both constraints will be analyzed in this research in order to explore possible causes of the delivery unreliability. In this analysis the capacity constraint is split up into the three stages of the order-cycle.
Internal department control
ALG is controlled. After this analysis it should be possible to appoint causes of the delivery unreliability.
Although, due date assigning and internal control are both important factors which contribute to a high delivery reliability it are certainly not the only factors. It however seems fair enough to concentrate on only these two factors considering the available time. Where necessary other factors will be mentioned, it is however not the intention to give a full description of these factors.
§ 2.6 Sub-questions
Breaking down the conceptual model in several pieces results in the following sub-questions. Together, these questions should give an answer to the main research question, presented in paragraph 2.3.
1. What is the relative importance of delivery reliability as performance objective?
Due date assigning:
2. How is the due date assigning process arranged at Alfa Laval Groningen?
3. Which factors do influence the due date assigning at Alfa Laval Groningen?
Internal control:
4. In what way are the processes internally controlled at Alfa Laval? 5. Which factors do influence the internal control of the processes at Alfa
Laval Groningen.
§ 2.7 Theoretical framework
This paragraph clarifies the theories used to find answers to the sub questions. After reading this paragraph the reader should have an idea of how the sub-questions are going to be answered. It depicts the relationship between the different theories that will be used to answer the sub- and eventually the main question.
Relative importance delivery reliability
factors are factors that directly and significantly contribute to winning
business.
Due date assigning in mass customization
In order to analyse the due date assigning process the framework of Rudberg and Wikner is used. According to Rudberg and Wikner (2004) the due date assigning process must be based upon the planning constraints faced by a company, constraints based on either material or capacity (or both). Material constraints are those related to the limited availability of materials (raw material, modules, finished products, etc.). Capacity constraints are caused by limited production capacities. Rudberg and Wikner posit that these constraints occur in two different dimensions. The engineering dimension and the production dimension. The material and capacity constraints in both the engineering- and production dimensions form together the four cornerstones in the order promising process in mass customization environments. These cornerstones are depicted in figure 6. According to the authors these four cornerstones can help companies design a production planning and control system in order to support mass customization. The four cornerstones are now clarified.
Figure 6: Four cornerstones in designing an order promise process. (source: Rudberg & Wikner 2004) Available-to-promise of the production dimension (ATP pd)
ATP pd takes the pre-produced products and inventory of raw material into consideration. It is that portion of the on hand inventory that is not already allocated to specific demands and still available to satisfy new demands or demand configuration changes. This check is typically performed by a planner. Together with the company’s purchasers he is able to keep the overview via the company’s information system.
Capable-to-promise of the production dimension (CTP pd)
Besides the ATP pd check the CTP pd is the second check which should be performed. It checks whether there is enough available capacity to produce a specific order. It is that portion of the “on hand” capacity that is not already allocated to specific demands.
(Rudberg and Wikner 2004). The ATP ed checks whether a drawing of a specific order is already available. The CTP ed check is performed to check whether there is enough drawing capacity. Once the checks have been performed the earliest completion and availability dates should be known. With this information the organization should be capable to promise a reliable due date to the customers. Notice that standard orders usually do not require a check of the engineering dimension.
Internal control
According to Bertrand and Wijngaard (1985) production of products has evolved over time in the direction of specializations. Because of this, organizations have become entities with different production departments and complex good flows between those production departments. Each of these departments has its own objectives while for the delivery of an end product several departments need to work together. In order to realize customer requirements on delivery flexibility and delivery reliability a company needs to coordinate its production activities. Taking the above as starting point Bertrand et al (1998) developed a framework in order to analyze the internal coordination activities of an organization. The framework exists of six general operational key decisions which play an important role, according to Bertrand et al (1998) in determining the internal control of the company. These operational decisions take place on three different levels inside the organization. The framework is visualized in figure 7.
Capacity planning Occupation planning
Order acceptation Due date assigning
Work order release
Work order detailplanning
Capacity assignment Capacity variation Priority dispatching Logistical parameters Company level Factory level Department level
Figure 7: Operational internal control decisions Bertrand et al (1998)
decisions mentioned in the framework, in practice seem to be standard. In this research it is analysed whether the different stages of an order cycle are in control. For this analysis only the four departmental decisions are used because the analysis is performed on departmental level. Next to this the order acceptation and due date assigning is already treated in the left side of the conceptual model. For that reason only these decisions are clarified below. Work order acceptance/release
Work order acceptance and release are two different decisions. “The acceptance of order results in a number of orders to be produced in a certain period” (Soepenberg et al, 2007).
The release decisions releases customer order to a certain production department. The aim of work order release is to control the workload level in the department in relation to the available capacity in order to control the throughput time (Bertrand et al, 1998). The release decision is one of the most important decisions of the order cycle. Once this decision is taken incorrectly, the entire process will be stressed.
Work order detail planning
Every order needs a plan according to which activities are being performed. In a work order detail plan each process at a department has a start time and an end time between which the activities need to be performed. Working according to this plan results in a controlled and clear department.
Capacity assignment
According to Bertrand et al the department must have flexibility in its capacities in order to accept different compositions of configurable and non-configurable orders. Flexibility can be expressed on two different ways. Firstly, volume-flexibility is the ability of a department to vary it total capacity. Secondly, mix-flexibility is the ability of a department to vary the separate capacity output in one department with a constant rate of total capacity.
Priority dispatching
Priority dispatching is a control decision which arranges the sequence of the work orders inside a production unit. This sequencing decision is a control decision which is taken within a production unit after work is released. It influences the throughput time performance.
entry pool release queue process Acceptance due date Release Priority dispatching Med. Term capacity control Short term capacity control Daily cap. adjustments Input control Output Control
Figure 8: figure of due date control decisions (source: Land & Gaalman 1996)
Figure 8 clearly shows the point in time on which these three decisions are taken. More recently Soepenberg et al (2007) tried to link those three decisions to performance indicators by describing the way they affect the delivery performance of a company. They argued that process control and therefore delivery reliability contexts with the average lateness and variance in lateness of orders. This relationship is depicted in figure 9.
Reducing the average lateness Reducing the variance of lateness Lateness Delivery reliability
Figure 9: Relation between the average- and variance of lateness and delivery reliability (source: Soepenberg et al 2007)
Figure 9 represent three distribution functions of lateness. The vertical line indicates zero lateness. The shaded area, at the right of the vertical line, represents the percentage of orders delivered late. Soepenberg et al posit that the percentage of orders delivered late can be reduced by reducing the average lateness and/or by reducing the variance of lateness.
standard order is most of the times shorter than the lead time of a standard on request order (Recall § 1.3 for the product definition)
By combining the approaches of Bertrand et al (1998) and Soepenberg et al (2007) it can be investigated to what extent the departmental control decisions influence the average lateness and variance in lateness. The combination of both approaches will be depicted with the help of the matrix presented in table 2. The first column, of the matrix depicts, the four departmental control decisions of Bertrand et al. The second and third columns depict the type of relationship between a certain control decision and the variance in lateness and the average lateness. The type of relationship can be classified as positive and/or negative. A positive relationship means that the control decision already is used to decrease the average lateness and/or variance in lateness. A negative influence means that a certain control decision is not taken into consideration which automatically increases the average lateness and or variance in lateness. A positive and negative relationship indicates that a certain control decision is taken into consideration for some of the activities performed in an order cycle stage but not for all activities.
+ = Positive influence
+/- = Can have a positive or negative influence depending on the characteristics of the customer demand.
- = Negative influence
Table 2: relation between control decision and the performance indicators
§ 2.8 Overview of thesis
This thesis is structured in five different parts. After this problem introduction and methodology section, the first part is concluded with a small summary of this chapter. Part two contains the first step of the OM-scan, the concretization, in which delivery reliability is described and prioritized. In the third part of this thesis the actual analysis, depicted in the conceptual model, is worked out. This will be done by discussing the interfaces of each order cycle-stage (recall § 1.3) with due date assigning and internal department control in separate chapters. After this part the reader should have a clear picture about the way how ALG’s delivery reliability comes into existence. Part four addresses the main causes of the delivery unreliability. Ultimately, in part five, improvements are presented which try to solve the main causes of the delivery unreliability.
Influence on average lateness
Influence on variance in lateness
§ 2.9 Conclusion
CH 3 DELIVERY RELIABILTY
In this chapter the first step of the OM-scan will be performed. In this step delivery reliability is concretized by addressing the following topics: (1) Making delivery reliability operational, (2) Determine its relative importance and (3) Review the delivery reliability of ALG. By the end of this chapter the reader should be aware of how delivery reliability is defined, its importance and how it is measured.
§ 3.1 Making delivery reliability operational
Throughout the literature definitions of delivery reliability are more or less the same. The only difference between the definitions is the vagueness about the variables to measure. To illustrate this difference the following two definitions are brought forward:
Hill (1989) defined delivery reliability as the ability of the firm to deliver on or before the promised scheduled due date is referred to as delivery reliability.
Schönsleben (2004) defined delivery reliability as the number of products delivered on confirmed delivery date divided by number of confirmed products.
Both authors seem to agree on the fact that the delivery reliability is measured against the promised due date. The definition of Schönsleben however specifically addresses the variable to measure (number of products), whereas Hill remains vague at this point. The second difference between both definitions is that Hill also takes orders/products delivered before the promised scheduled due date into account, whereas Schönsleben on his turn, remains vague at this point. Combining both definitions seems logical and leads to the definition used by the Alfa Laval Group. This definition is stated below.
Number of products delivered on or before promised delivery date divided by the number of confirmed products. Whereby 3 days before delivery day is counted as reliable.
valuable benchmark tool on delivery reliability is very important in the case of ALG. This importance is explained in the next paragraph.
§ 3.2 Relative importance delivery reliability.
Until now the importance of the delivery reliability improvement was mostly underlined with internal opportunities to decrease waste (inside-out). However the actual need for better delivery reliability can also be found in the environment of ALG (outside-in). As most production companies in high wage countries ALG also faces fierce competition out of low wage countries. Because ALG is not able and does not want to compete with factories in these countries they made a clear choice to follow a differentiation strategy. “A differentiation strategy places specific demands on the operations function, though their exact nature will depend on how differentiation is to be achieved” (Slack et al 2002). At ALG this strategy is given proof by producing customized products with high quality standards. Because direct competitors of ALG (GEA/Goedhart and Güntner) also operate in the differentiation segment, ALG needs to perform better, in comparison with their direct competitors, on certain performance objectives. A useful way to determine on which objectives ALG needs to perform better than its competitors, we distinguish order-qualifiers and order-winners. In the case of ALG, quality, delivery reliability, product flexibility and volume flexibility are typical order winners. Customers select ALG as their supplier for its customized high quality products with a reliable delivery time. Costs and delivery time are typical order-qualifiers. With only a fraction (15%) of the offers becoming a final-order, delivery reliability seems even more important as an order winner.
§ 3.3 Delivery reliability ALG
ALG commits itself at a delivery reliability of 95%. Considering the importance of the delivery reliability, mentioned in the previous paragraph, the height of this figure seems logical. The delivery performance, as depicted in figure 10, is calculated by the production leader.
Delivery reliability Industrial 2007 (wk 12-26): Goal > 95%
50% 60% 70% 80% 90% 100% 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Week# S co re
Weekscore progressive average 15 wks Average last year Goal
The graph shows several drops in delivery reliability. Especially week 18, 19, 21, 22 and 24 show dramatic falls. The causes of these falls are investigated and discussed below:
- Week 18 and 19
Due to a sudden rising order intake compared to previous years the inventory aluminium was consumed faster than predicted. Although ALG reacted fast and timely placed new orders for aluminium, both their aluminium suppliers were not able to deliver on time. Aluminium is a main component in many heat exchangers, fabricated by ALG, this clarifies the high delivery unreliability in week 18 and 19.
- Week 21 and 22
Although the fall of the delivery reliability in these weeks was not as dramatic as in the previous weeks it is still interesting to indicate the causes of this descend. The fall was caused by the production of a special. The shop floor faced the problems of manufacturing a special never build before. The special needed some special handlings which caused delay of several shop orders behind this order.
- Week 24
Causes for the fall in week 24 are: (1) Delay previous order(s), (2) production error, (3) machine failure, (4) absence through illness and (5) norm time deviation. It therefore can be concluded that it was a concurrence of circumstances causing the great fall in delivery reliability. Besides the causes of the great falls discussed above there were also found some other, not described, causes of delivery unreliability. The most frequent causes are presented in the figure 11.
Causes of delivery unreliability (week 19 till 26)
23 19 6 4 3 3 1 0 5 10 15 20 25 Com pone nt s hort age Del ay p revi ous orde r Prod uctio n er ror Mac hine failu re Nor m ti me devi atio n Abs ence illn ess Rus h-or der Frequency
Figure 11: Measured causes of the delivery unreliability form week 19 till 26
reliability. Both causes however, are in close contact with planning and control activities performed during the order processing stage. It therefore seems justified to pay extra attention to the order processing. Before doing so, some general terms, used within ALG, to calculate the delivery reliability need to be clarified.
ALG distinguishes different order due dates throughout the order processing. Each specific due date clarifies something about the status and progress of the order and is communicated on different moments to the customer. These different due dates are:
o The requested due date is the desired due date for a specific order. It is the result of a negotiation agreement between the customer and the sales department. However, ALG has a lot of influence on the determination of this due date.
o The promised due date is the assured due date, communicated back to the customer, after some regular availability checks have been made. It is an internal agreement to communicate this due date back to the customer within 2 working days.
o The planned due date is the due date which is assigned to the customer order after the production planning has been made. This date represents the date on which the order comes out of production.
The precise point of determination and the reliability of this due date will be discussed during the diagnosis of the different order-cycle stages.
The production leader calculates the delivery reliability as follow: Every time an order exceeds the planned due date of the production planning the order is counted as unreliable. Virtually he divides the number of too late orders according to the production planning through the total number of planned orders on the production planning. The approach of ALG, calculating the delivery reliability with the planned due date instead of the promised due date, differs fundamentally from the definition stated in paragraph 3.1. The planned due date can considerably differentiate from the promised due date because the moment of determination is totally different. Measuring delivery reliability with the help of the planned due date mostly measures the reliability of the planning. In other words it only measures the internal reliability of production department instead of the external reliability influencing the customer service level. What happens, for example, if the waiting time between production and expedition is too long to arrange on-time transport? In practice it happens someon-times that work instructions for expedition have not arrived yet while the product already is finished. It therefore can be said that the actual delivery reliability is even lower than measured.
§ 3.4 Conclusion
to order environment, delivery reliability is of inestimable value. This justifies the priority and attention paid to the delivery reliability.
Comparing the definition in paragraph 3.1 with the method of how delivery reliability is measured at ALG brings up the following conclusions:
- Delivery reliability is measured between the production and expedition department. This point of measuring results in unreliable data of delivery reliability.
- Delivery reliability is measured with the help of the planned due date instead of the promised due date. By measuring the delivery reliability with the planned due date the voice of the customer is partly eliminated.
CH 4 ORDER PROCESSING
The order processing is an important stage of the entire order cycle. It translates a customer order into a production order. It is involved in the assignment of a due date and needs to be internally controlled in order to realize this assigned due date. This chapter zooms in on the order processing of ALG. In doing so a distinction is made between the order processing of configurable and non-configurable orders. The non-configurable order flow is more complicated and time consuming compared to the configurable order flow. Both order flows are depicted with the help of an Actor Activity Diagram (AAD) and are described afterwards. “Actor Activity Diagramming is a tool for modelling business processes. By its design AAD is especially useful for communication and discussion about business processes oriented issues like performance or Information Technology enabled redesign”2.
Appendix I depicts the legend which helps to read the presented AAD’s. Paragraph 4.1 concretizes the order processing of configurable orders at ALG. Paragraph 4.2 and 4.3 will pay attention to respectively the due date assigning and internal control at the processing of configurable orders. Paragraph 4.4 discusses the order processing of non-configurable orders. Paragraph 4.5 and 4.6 will respectively discuss the due date assigning and internal control at the processing of non-configurable orders. This chapter will be summarized in paragraph 4.7.
§ 4.1 Order processing configurable orders
Before something can be said about the order processing of configurable orders the process must be concretized. In order to do so the activities performed at the order processing are presented in a AAD. The AAD of configurable products is depicted in figure 12. The presented activities in figure 12 are performed at the processing of so called standards and standard on request product categories (recall §1.3).
Order arrival
An order enters the organization at the sales department per telephone, fax or e-mail. If it is a configurable order the sales employee enters the order in the information system (IFS).
Allocation materials and capacities
Based on the inputs of the salesman, IFS generates the pre-defined process lead-times, required capacity and components etc. The order engineer at the company office checks this generated data and releases the order for the
planner and the purchasing department. A released order automatically gets a shop-floor number on which it can be recognized by the production leader.
Check IFS for standard orders released for planning
Organization Company process
= Alfa Laval Groningen
= Order processing configurable orders
ACTOR ACTIVITY DIAGRAM
Customer Salesman
Order
Engineer IFS
Config. order Order
Enter order in IFS
Check IFS for standard orders Allocation material and capacity Planner Plan order Assign planned delivery date
IFS Check for standard orders
Prepare order Work instructions Production leader Requested due date ATP PD MRP a & MRP b
CTP PD Planned due datePromised & Communicate to
customer Communicate to sales
Figure 12: AAD configurable orders Plan order
Prepare order
Once planned the order gets back to the order engineer he takes care that the work instructions and labels for the different heat exchangers are delivered on time to the production leader.
The AAD forms a great starting point for an analysis after the due date assigning. This analysis will be performed with the help of the framework of Rudberg and Wikner discussed in paragraph 2.7.
§ 4.2 Due date assigning configurable orders
By shoving the framework of Rudberg and Wikner over the due date assigning process of configurable orders it can be checked whether the communicated due date to the customer is reliable or unreliable. According to Rudberg and Wikner a due date can be considered reliable once the relevant planning constraints are taken into account.
As depicted at the right side of figure 12 ALG distinguishes two different due dates at the configurable order stream. The requested due date is established in dialogue between a sales employee and the customer in the pre-sales trajectory. The requested due date is highly uncertain and only serves as a directive for the customer in the negotiation stage. The planned due date however is assigned after the ATP and CTP of the production dimension are checked. These checks respectively consider the component availability and the available production capacity (see also §2.7). The checks of the engineering dimension do not have to be considered, since configurable orders do not need any engineering activities. It therefore can be concluded that the planned due date is as reliable as it can get. This due date however has to be assigned within 2 working days. An analysis of 961 already delivered configurable TR, TC and ZT order lines indicated that 80% of the configurable orderliness were assigned with a planned due date within 2 days. The precise results of this examination are displayed in appendix III. Based on the above it can be concluded that the due date assigning at the configurable order stream has no significant effect on the delivery reliability of ALG. For the sake of clearness it has to be said that this conclusion does not mean that configurable products can not be delivered too late. Delivery unreliability still can be caused by: production errors, delay of other orders, suppliers unreliability etcetera.
§ 4.3 Internal control order processing at configurable order stream
§ 4.4 Order processing non-configurable products
After having discussed the due date assigning and internal control in connection with the order processing of configurable orders, it is now time to discuss both topics in connection with the non-configurable order stream. The AAD of non-configurable products is depicted in figure 13. The presented activities in the diagram are performed during the processing of so called “customer specific products” and “specials” (recall §1.3). The performed activities are now shortly explained.
Order negotiation
When a non-configurable order is received the sales employee writes down the customer requirements on a so called Prokal form (Prokal: product calculation). This form is passed on to the foreman of the technical draughts men with a request for a specific due date. The requested due date is determined in a dialogue between a sales employee and the customer. In this dialogue the sales employee makes use of a list of delivery times for the different product families provided weekly by the planner. This list is based on CTP ed check and the CTP pd check (drawing and physical production) and serves as a directive. An example of such a list is enclosed in appendix II. Make drawing plan
Once the foreman of the technical draughtsmen has received the Prokal form, he makes a planning for the drawing process. Based on his experience he calculates the approximate lead-time which is needed to draw a specific order. After the foreman made the planning for the drawing table he is responsible for communicating a promised due date back to the sales employee within 2 working days. He calculates the promised due date by adding up a standard production onto the estimated internal due date of the drawing department. This standard production time is dependent on the capacity utilization of that moment and is provided to by the planner.
Reservation materials and capacities
Before the drawing process starts, the Prokal is given to the order-engineer, who performs a pre-reservation. In this pre-reservation the order engineer enters the Prokal in IFS. Based on these imports IFS calculates the approximate production lead-times. These lead-times are used by the production planner to “forecast” upcoming demand. Next to this the order engineer reserves the already known material requirements. At this point only the MRP-a articles are taken into consideration. MRP-a-components are stock components, ordered with the order point technique indicating that demand is uncoupled from a custom order.
Organization Process stage
= Alfa Laval Groningen
= Orderprocessing Non-configurable products
ACTOR ACTIVITY DIAGRAM
Customer Salesman Technical Draughtsman IFS Order Engineer Planner Config. Order Order
Fill-in Prokal form
Prokal form
Promise delivery week
Prokal form
MRP-A component and capacity allocation
Prokal form
Make drawing
Send drawing to customer if requested
Check drawing
Email or fax with approval and/or commments
Redraw If correct
Drawing
Allocate material and capacity
Prokal form
Plan order Communicate Delivery time Prokal form
Prepare order for production Prokal form Material planning To production Requested due date CTP ed / ATP ed
Promised due date
ATP pd MRP a / CTP pd
ATP MRP a & MRP b
Planned due date Make planning drawing dept. Communicate to customer Communicate to sales Communicate to sales Communicate to customer
Make drawing
Once the drawing is finished the technical draughtsman sends the finished Hicat drawing to the customer for order confirmation. Once confirmed, or if no order confirmation is needed, the Prokal form with drawing is passed on to the order engineering. The customer confirmation of an order is a process which is highly dependable on the willingness of the customer to corporate fast.
Allocation and capacity reservation
Once the drawing process is finished it is the task of the order engineer to allocate the material and capacity. The order engineer does this by entering the data into IFS. Based on these inputs IFS calculates the needed production capacity at each workstation more precisely compared to the earlier made pre-reservation. Once the required components are entered into the system, the order-engineer performs regular checks such as cost price, needed production time, etc. If no remarkable things are found the order-engineer releases the order for planning. Once released, IFS generates a shop-order-number with which the planner and the shop-floor recognize the order. Once a shop-order-number has been assigned IFS also makes the material requirements of both MRP-A and MRP-B components visible.
- MRP-A-components are stock components, ordered with the order point technique indicating that demand is not coupled with a customer order.
- MRP-B-components are coupled to a specific customer order. The amount needed for the customer order is ordered at the supplier and no more.
The information generated by the MRP application in IFS, triggers the purchase department to buy the required components. The reason for this late ordering is because some material requirements become gradually known during the drawing process.
Plan order
Once finished the order engineer passes the Prokal form to the planner. Dependent on the ATP pd and CTP pd checks the planner plans the orders at shop floor level according to backward scheduling techniques. Once the planning is made the planner communicates the delivery time to sales. The sales department, on its turn, communicates this planned delivery time to the customer.
Prepare order
The visualization in the actor activity diagram, forms a great starting point for the diagnosis of the due date assigning and internal control of non-configurable orders. The AAD, presented in figure 13, visualizes important communication points at the due date assigning and internal control of the non-configurable order stream. In the next section the reliability of the different due dates assigned to an order are analyzed. After that the internal control of the order processing stage will be analyzed. Both analyses will be performed in accordance with the theories explained in the theoretical framework in paragraph 2.7.
§ 4.5 Due date assigning non-configurable orders
In order to estimate a due date accurately, ALG distinguishes different order due dates throughout the order processing. Each specific due date clarifies something about the status and progress or the order and is communicated on different moments to the customer. These different due dates are (1) the requested, (2) the promised and (3) the planned due dates. The reliability of these estimated due dates will be investigated by checking whether the company constraints mentioned by Wikner and Rudberg (2004) are taken into consideration.
§ 4.5.1 Requested due date
The requested due date is determined during the negotiation phase between sales and the customer. The due date list depicted in appendix II gives the sales employee an indication of the possible delivery week. This list is based on available capacity in the engineering- and production dimension (respectively CTP ed and CTP pd). It can be concluded that the uncertainty of this requested due date is high because it is assigned without any knowledge of component availability (ATP pd) and/or required lead times. Nevertheless it has a very important function, the requested due date serves as a directive for the customer in negotiation stage and is used to compare lead times between different suppliers of heat exchangers.
§ 4.5.2 Promised due date
inventory, an essential check to assign a due date. In many cases the delivery time of components is longer than the production time. It sometimes happens that the sales employee or technical draughtsman based on their experience notice a scarce component with a long delivery time which should be considered in the due date assigning. They communicate this to the planner who calculates the earliest completion date based on the scarce component. However, there are no formal rules for this kind of situations.
§ 4.5.3 Planned due date
After the drawing is finished and worked out by the order engineer a production order number is created. At this point the MRP-B components become visible in IFS for the purchasing department and are ordered. Based on the delivery time of these components and the available capacity (CTP pd) the planner assigns a planned due date. This planned due date takes the following factors into account: (1) expected lead time, (2) component availability (ATP pd) and (3) production capacity (CTP pd). In other words this due date cannot become more certain than this.
§ 4.5.4 Discussion
Summarizing the characteristics of the due dates treated above, the following table can be put together. The table depicts whether the constraints mentioned by Rudberg and Wikner are taken into consideration during their determination.
COMPONENT CONSTRAINT CAPACITY CONSTRAINT
ATP Check engineering dimension ATP Check production dimension CTP Check Engineering dimension CTP Check Production dimension Requested due date Yes, but roughly
No, not at all Yes, but roughly estimated
Yes, but roughly estimated.
Promised due date
Yes, precise No, only at own initiative
Yes, precise Yes, but roughly estimated
Planned due date
Not relevant anymore
Yes, precise Not relevant anymore
Yes, precise
Table 3: Reliability characteristics of different due dates
within 2 working days after the requested due date is agreed. It can be concluded that ALG is satisfied with only a rough and probably wrong estimate of the expected due date for non-configurable orders. This has a significant effect on the delivery reliability, if it is measured according to the definition stated in paragraph 3.1.
According to the sales employees this discovered significant effect on the delivery reliability is true. They claim, that the promised due date (determined within 2 working days) is not reliable and most of the times drastically differs from the planned due date. For this reason they do not always communicate this date to the customers. They rather wait till the planned due date is assigned and communicate this to the customers. Despite the fact the customer expects a promised due date within a certain time period. According to the sales employees the planned due date is more reliable. The planner, on his turn, is convinced that the promised due date is in 90% of the cases identical to the planned due date. He therefore thinks that this promised due date easily can be communicated to the customer by the sales department. A random sample of 80 Prokals per sales team evenly spread throughout the year (20 Prokals per quarter) proofs that the planner is right. The sample size of 240 Prokals proofs that the planned due date in 93% of the cases is similar to the promised due date. This similarity does not prove that these orders are delivered on time. After the planned due date is assigned it often happens that an order is rescheduled unlimited because of, for example, a component shortage and/or machine failure. The planner communicates the rescheduled orders with their new due dates via an informal e-mail to the sales department. It even happens sometimes that no feedback is given when an order is delayed. These two arguments probably are the causes of the difference in perceptions of the sales department and the planner about the difference between de promised and planned due date and proves the unreliability of the promised due date partly. The difference in perceptions and unlimited changes of due dates via informal e-mails are an important cause of the existing tension between sales and production. Establishing a formal feedback system in which specific appointments are made about the communication of the rescheduled due dates would probably diminish this tension. Next to this it is recommendable to find ways to reduce the uncertainty of the promised due date by integrating the ATP check of the production dimension.
§ 4.6 Internal control non-configurable orders
