Evaluating hybrid production systems in the
food processing industry: A multiple case study
Master Thesis, MSc Technology & Operations Management University of Groningen, Faculty of Economics and Business
Kees van Herpen S2031418
k.a.f.van.herpen@student.rug.nl
First supervisor: Prof. Dr. D.P. van Donk Co-assessor: Dr. O.A. Kilic
Word count; 12,058
1
Abstract
Purpose – The goal of this thesis is to explore characteristics that influence the organization of hybrid production systems in the food processing industry. While previous research has reviewed numerous food processing industry characteristics, their impact on hybrid production systems is not fully understood. We explore the impact of these characteristics that are expected to influence the organization of hybrid production systems: expensive single purpose factory, long sequence-dependent set-up time, variable supply, variable, quality, variable price, variable demand, perishability of goods, variable yield and processing time, one process has homogenous products, different labor intensities among different production stages, production rate is determined by capacity, a divergent product structure, and several different recipes available for one product.
Design/Methodology/Approach – An explorative multiple case study is conducted in which information from interviews and several other sources is used. Both the context in which production systems are employed in the food processing industry and the relative impact of food processing characteristics on the organization of the production systems are reviewed.
Findings – The main finding of this research is the importance of perishability of goods and the existence of long sequent-dependent set-up times when reviewing the applicability of production systems. It was found that these characteristics have a larger influence on the way a production system is organized than other investigated characteristics. Furthermore, the impact of an unavailable/unskilled workforce was also found to have an impact on the execution of the production system. On the other hand, expected characteristics such as variable supply, quality, and production yield have limited to no influence on the applied production system.
2
Preface
3
Table of Contents
Abstract... 1 Preface ... 2 1. Introduction ... 5 2. Background ... 72.1 The food processing industry ... 7
2.2 Planning and control in hybrid production ... 8
2.3 Linking FPI characteristics to hybrid production... 11
3. Methodology ... 14
3.1 Research method ... 14
3.2 Case selection & description ... 15
3.3 Data collection ... 16
3.4 Data analyses ... 17
4. Results ... 18
4.1 MTO/MTS decision ... 18
4.2 Capacity coordination ... 22
4.3 Scheduling & Control ... 30
4
Table of Figures
Figure 1: Basic food processing structure (Akkerman and Van Donk, 2009: p.3) ... 8
Figure 2: Hierarchical production planning framework for MTO–MTS production (Soman et al., 2004)... 9
Figure 3: Conceptual model ... 12
Figure 4 Production system cases... 19
Figure 5: Perishability vs production system ... 20
Figure 6: Production driver vs production system... 22
Figure 7: Sequence-dependent set-up time vs production system ... 25
Figure 8: Variable yield vs production system... 26
Figure 9: Variable quality vs production system ... 28
Figure 10: Variable supply vs production system ... 29
Figure 11: Sequence-dependent set-up times vs sequencing methods ... 32
Table of Tables Table 1: FPI Characteristics (Van Donk, 2001) ... 7
Table 2: Research quality criteria ... 15
Table 3: Short overview cases used ... 16
Table 4: Data collection details/overview ... 17
Table 5: Strategic (HPP) level decisions ... 18
Table 6: Tactical (HPP) level decisions... 23
Table 7: Operational (HPP) level decisions ... 31
Table 8: Case A coding scheme ... 32
Table 9: Case B coding scheme... 34
Table 10: Case C coding scheme ... 36
Table 11: Case D coding scheme... 37
Table 12: Case E coding scheme ... 39
Table 13: Case F coding scheme ... 42
Table 14: Inductive coding example... 44
5
1. Introduction
Research on multi-product production planning and control systems has expanded from ‘pure’ make-to-order (MTO) and make-to-stock (MTS) systems to hybrid systems. Hybrid systems are production systems where a group of products are produced MTO while others are produced MTS. Hybrid systems are often employed when there are long sequence-dependent set-up times, when there is perishability of goods (Van Donk, 2001), or to better utilize available capacity (Beemsterboer, Land & Teunter, 2016). The relationship between MTO and MTS products in a single production system creates a difficult trade-off between capacity usage, inventory and lead-time (Stevenson & Found, 2016). In such systems the aim is to avoid stock-outs for MTS products while, at the same time, maintaining short lead-times for MTO products (Soman, Van Donk & Gaalman, 2006). Due to these different performance indicators the combination is problematic when incorporated in a production planning. Operating a hybrid system involves complexities across different production planning levels that should be thoroughly evaluated and incorporated in the production system (Romsdal, Arica, Strandhagen & Dreyer, 2013; O’Reilly, Kumar & Adam, 2015). Soman, Van Donk & Gaalman (2004) emphasize that this calls for adopting tailor made production and inventory strategies to cope with such hybrid systems.
6
Soman et al. (2004) propose a hierarchical production planning (HPP) framework that aids decision making for the design or re-design of a hybrid production system. The framework separates relevant production decisions and places them on the appropriate strategic, tactical or operational level. The decisions that are relevant on each level are respectively; the separation of MTO/MTS product groups, capacity coordination of the groups, and the scheduling and control of the productions system. Without clear constraints at higher planning levels ad hoc decision making will likely occur at the operational level (O’Reilly et al., 2015). The framework has been around for more than a decade and has proven its ability to aid in decision making with regard to hybrid models (Soman et al., 2006; Soman et al., 2007; O’Reilly et al., 2015). While it is known that FPI characteristics such as perishability and long set-up times are of influence on the production system (Van Donk, 2001) there has not been a study that explores the relative influence of these factors on the organization of hybrid production systems.
The focus of this thesis will therefore be on how hybrid production systems are organized in the FPI and what characteristics influence the way they are organized. Due to the existence of various important characteristics of which the impact is not fully understood, a multiple case study is proposed. The combination of challenges that this industry faces, such as, but not limited to, perishability of the goods, seasonality of demand and supply, and the increasing demands of customers (Claassen et al., 2016) emphases the need for this. These characteristics are expected to influence production and inventory strategies (Soman et al., 2004). We propose the following research question;
How are hybrid production planning systems organized in the food processing industry?
7
2. Background
In this section the theoretical background of the research is set. We will start by defining the expected influential FPI characteristics that have been found in previous research. In the following section we provide an overview of the HPP framework and literature on hybrid production systems. We end the background section with a discussion on the findings and a conceptual model of the expected relationship of the different FPI characteristics and the HPP framework.
2.1 The food processing industry
When reviewing the various food processing characteristics Van Donk (2001) shows that various FPI characteristics can severely impact the applicability of production systems. The author compiles several factors from literature that can be present in this industry. The author states that each of the factors should be taken into account for planning and scheduling purpose s. An orientation to use capacity as much as possible and the existence of high set-up times, cause planning of long production runs and high stocks of end-products (Van Donk, 2001). The characteristics compiled by Van Donk are separated in plant, product and production process characteristics and are given in table 1.
Table 1: FPI Characteristics (Van Donk, 2001)
Plant characteristics Product characteristics Production process characteristics Expensive a nd single-purpose
ca pa city coupled with s mall product va ri ety a nd high vol umes. Usually, the fa ctory s hows a flow s hop ori ented design
The nature and source of raw material i n food processing i ndustry often i mplies a va riable supply, quality, and pri ce due to unstable yield of farmers
Proces ses ha ve a va ri a ble yi eld and processing time.
There a re long (sequence-dependent) s et-up ti mes between di fferent product types
Vol ume or wei ghts a re us ed rather tha n distinct i tems.
At l east one of the processes deals with homogeneous products.
Ra w ma terial, s emi-manufactured products, a nd end products are peri shable.
The proces sing s ta ges a re not labor i ntensive.
Producti on ra te i s ma i nly determined by ca pa city.
Food i ndustries ha ve a di vergent product s tructure, especially in the packaging s tage. Fa ctori es tha t produce consumer goods can ha ve an extensive, labor-intensive packaging pha se.
Due to uncertainty i n pri cing, quality, and s upply of ra w ma terial, s everal recipes are a va ilable for a product.
8
which implies that the production rate is mainly determined by a high usage of capacity (Soman et al., 2004).
Akkerman and Van Donk (2009) compiled a basic food processing structure in which the process is separated into two stages: a processing stage during which the food products are manufactured, and a packaging stage in which the products are packaged. Pool, Wijngaard & Van der Zee (2011) further define these production steps and differentiate the steps as process industry and discrete industry respectively. The food processing is characterized as a process with long production runs, small number of products, and sequence restrictions while the packaging is characterized by short production runs and a large number of stock keeping units (SKUs). This is an important differentiation to consider as it can determine when a product becomes customer specific, which can affect the flexibility that a company has.
Figure 1: Basic food processing structure (Akkerman and Van Donk, 2009: p.3)
2.2 Planning and control in hybrid production
9
defined. First the MTO/MTS product segregation is determined. Based on this decision the capacity coordination is set, which includes lot-sizes and safety stocks for MTS products and due date policies for MTO products is set. Finally, the detailed scheduling decisions are made that include the production sequences and weekly production volumes (Soman et al., 2004). For the sake of completeness, the full framework is shown in figure 2.
Figure 2: Hierarchical production planning framework for MTO–MTS production (Soman et al., 2004) MTO/MTS decision
10
grouped in a standard, repetitive group and a custom group. Afterwards it was advised to manage the groups based on efficiency (MTS) where short lead-times are considered given and effectiveness (MTO) with respect to design quality, conformity and delivery time respectively. Kerkkänen (2007) performed a case study of a steel mill that switched from MTO to hybrid production. In this study, the author finds that the partition of products in MTO/MTS groups cannot be made purely based on demand analysis as technical aspects can play a role as well. The author found that when a large amount of the information needed for the decision-making is in an intangible and qualitative form, organizational factors will dominate the decision-making process. This can be problematic as it can result in a lack of information when decisions need to be made. When considering the MTO/MTS decision in the FPI this can be problematic when taking into consideration e.g. perishability (Soman et al., 2007). While perishability and demand/volume variability have been mentioned as possible factors that influence the MTO/MTS decision it is still unclear if there are possible other factors that can play an important role.
Capacity coordination
11
at the operational level. The impact of the various uncertainties e.g. the perishability of products, variable yield, variable qualities are not fully understood (Romsdal et al., 2013).
Scheduling and Control
Rafiei et al. (2013) connect the operational level to the tactical level of Rafiei et al. (2012). In their research they apply a priority rule for MTO, MTO/MTS and MTS products in a static job shop. In this research sequence dependent set-up times, pre-emption, and preventive maintenance activities are included. The authors stress the importance of sequence at the operational level: MTS and MTO production must be sequenced in a way that prevents excessive MTS production from increasing holding cost while limited production does not create shortages. On the other hand, on-time delivery and short cycle-times are important in MTO production. Rafiei et al. (2013) conclude that products must be sequenced in a way that the sum of earliness and tardiness time is minimized. On the other hand, Perona et al. (2009) state that planners tend to base their decisions on experience and common sense, which calls for approaches that are simple and illustrative without giving up making in a rational way. Clear rational decision-making is important to ensure that a company does not shift to an ad hoc decision-decision-making policy as that increases production costs in terms of poor production plan adherence and high inventory levels (O’Reilly et al., 2015). This leads to difficulties in maintaining the scheduling of the company and results in a decrease in efficiency and effectiveness. Stevenson and Found (2016) institute the fundamental importance of sequence and its impact on capacity utilization, service, and waste. In their simulation research they concluded that commonly accepted discrete industry MTO mechanisms that rely on quick changeovers and small batches lack applicability in the process industry.
2.3 Linking FPI characteristics to hybrid production
12
production cycle stability (Pool et al., 2011). When this is put in line with the earlier findings, we see that there several importance contextual factors that determine the best production system strategy in the FPI (Van Donk, 2001; Soman et al., 2004; Soman et al., 2006; Soman et al., 2007; O’Reilly et al., 2015) . While several characteristics, such as perishability, have been acknowledged, the characteristics as described by Van Donk (2001) are still not fully understood in context of hybrid production systems.
Concluding, the FPI faces multiple difficulties on several aspects of production and planning. To cope with these difficult aspects hybrid production systems are commonly employed and present interesting challenges such as dividing products to MTO and MTS, capacity co-ordination and schedule control (Soman et al., 2004). While there is a renewed interest in hybrid production systems, this is mainly in the form of highly mathematical scheduling approaches (Beemsterboer et al., 2016; Beemsterboer et al., 2017; Claassen, et al., 2016; Shahvari & Logendran, 2016). It is still not clear how companies organize their hybrid systems in different contexts. In this thesis we focus on this aspect of the model of Soman et al. (2004) and review how companies in the FPI make this decision.
Figure 3: Conceptual model
13
14
3. Methodology
In this section the methodology of the research will be elaborated upon. First the research method will be discussed. Afterwards the case selection and description, the data collection and data analysis will be reviewed.
3.1 Research method
The goal of this research is to explore how hybrid production systems are organized in the FPI and to address in what context this is applied. Karlssen (2016) stated that case research is good at researching “how’’ questions. Hence, a case study research is used to review the production systems in multiple cases. This explorative paper uncovers areas for research; it explores hybrid production systems in a specific context, the FPI. This multiple case study will be undertaken as this typically provides a stronger base for exploration than single case studies (Yin, 2009). Furthermore, case research is suitable as there are a high number of variables of interest and the knowledge of context is important.
15
Table 2: Research quality criteria
Quality Criteria Research design
Construct validity The use of multiple sources of evidence by using 6 different cases. A chain of evidence is provided by using coding.
The interviewees checked the transcripts of the interviews to mitigate any mistakes.
Internal validity All interviews were conducted by two interviewers.
When contradicting answers were given investigative questions were asked. External validity Due to the 6 different cases the findings are generalizable to a certain level. Reliability A semi-structured case study protocol is used during the interviews.
A database was developed in a cloud that includes all interviews and other information that was gathered.
There are three main phases in this exploratory case study research: case selection & description, data collection and data analysis. These three main phases are designed based on the case study guidelines and procedures provided by Karlsson (2016).
3.2 Case selection & description
This phase describes the case selection and a description of the cases. It is important to set boundaries in the selection phase to get a representative sample and useful variation on the dimensions of theoretical interest (Seawright & Gerring, 2008; Karlsson, 2016). This case study research includes the criteria that are of relevance to answer the research question. The research question at hand focusses on the hybrid production systems of companies in the FPI. More specifically, we focus on the model made by Soman et al. (2004). To ensure that the research would provide information that is relevant for the research question the chosen unit of analysis is the ‘hybrid production system of a FPI company.’ The next criterion is that all companies are food process related and should be located in the Netherlands to increase similarities between cases and to avoid cultural influences.
16
cases, as suggested by Eisenhardt (1989). We exploit the variation in the case sample for the explorative nature of this research.
Table 3: Short overview cases used
Characteristics\Cases A B C D E F
Main products Flour and Granulates
Starch Fries and potatoes Flour and Food Coatings Coffee and Tea Sauces Yearly production volume (tons) 6,600 - 8,400 600,000 100,000 23,000 14,500 235,000 Employees 22 1,000 350 120 350 n/a # of end-products 18 1,000 150 1,000 200 495
3.3 Data collection
The data is collected by means of semi-structured interviews combined with the collection of relevant documents, like production schedules and company websites. Furthermore, observations from the production facility are used to corroborate the data gathered by means of interviews. Multiple sources of collecting data allow for triangulation of data (Karlsson, 2016) and ensure the construct validity of this research (Voss, 2009). The data was collected during November and December of 2017 by two researchers. A case study protocol based on the planning levels of Soman et al. (2004) was used to ensure consistency in the data collection phase and will secure the reliability of the study (Yin, 2009). The case study protocol was divided in a general section, product grouping section, production planning section and detailed scheduling section (Appendix A). These sections were created to understand if the companies had a HPP approach, and what the decisions were at each planning level. This procedure helped gaining understanding of the production system the different companies had.
17
Table 4: Data collection details/overview
Cases Number of interviewees
Job title of interviewee(s) Information gathered Length of interview
A 1 Supply Chain Manager Interview, detailed
scheduling template, company tour
60 minutes
B 2 S&OP Planner & S&OP planner
Interview 70 minutes
C 2 Forecast Manager &
Planning Coördinator
Interview, detailed scheduling template, company tour
80 minutes
D 1 Supply Chain Manager Interview, detailed
scheduling template
70 minutes
E 2 Manager Planning &
Logistics & Team Lead Planning
Interview, detailed scheduling template
70 minutes
F 2 Continuous Improvement
Manager & Supply Planning Coordinator
Interview 70 minutes
3.4 Data analyses
18
4. Results
The three levels of the HPP were evaluated to review the production systems of the cases. For each section we provide an overview of the main decision the companies made in line with the framework created by Soman et al. (2004). We establish what factors influence the choices on each level of the HPP. We focus on the FPI characteristics as mentioned by Van Donk (2001), but explore other factors that are revealed during the interviews as well. In each section the main factors that influence decisions in the HPP framework will be elaborated upon by quotes and figures to provide context to the similarities and differences of the cases. Characteristics that were not forthcoming in interviews are not discussed and characteristics that are interrelated are discussed in succession after which the interrelatedness is stipulated.
4.1 MTO/MTS decision
The cases define their product groups mostly on product characteristics. No formal MTO/MTS decision that was based on demand variation and volume was found in the different cases. Cases with a hybrid production system defined products MTO due to special customer requests or were MTO after reviewing received orders and available inventory. Furthermore, cases C and F had a MTS production system, while case E had a MTO production system. Table 5, provides main MTO/MTS partition drivers, main characteristics by which cases formulated groups, and main production drivers. On this level it was found that perishability had the most effect on the MTO/MTS decision. Furthermore we found that the planning drivers also have an effect on this decision.
Table 5: Strategic (HPP) level decisions
HPP\Cases A B C D E F MTO/MTS Partition Inventory & received orders Customer wishes MTS Inventory & received orders MTO MTS
Product Groups Product characteristic Product characteristic Product characteristic Product characteristic Product characteristic Product characteristic Demand/supply driven Order driven Supply driven Forecast driven Forecast driven Order driven Forecast driven Production systems
19
The cases are set on the line based on the information that was provided during the interviews. To illustrate the difference in production systems, we provide the most relevant quotes below.
Figure 4 Production system cases
Most products that are MTO were due to special customer requests or received orders and available inventory. One Interviewee from case B said that most of their products are not made for a specific customer. ‘We have the largest share of make-to-stock and we produce and supply that to various
customers. Sometimes we have to change a pallet. In a number of cases we actually produce it on order.’
Case D distinguishes the importance of the customer for MTO products. ‘It also happens that you do not
have it in stock. Then you first have to bake, grind, pack. That can be done relatively quickly […]. That is not something you want, but you can. That only happens for the special customers.’ Case A mostly determines
the MTO/MTS decision on the received orders and inventory. ‘What we do is; when the orders arrive, we
collect all the orders for that product. Then we start production and fill cloth silos. That is then bagged and what is left then we put in stock.’ Cases C and F only produce MTS products. (Case F) ‘So we make a generic product and we sell that to those who want it.' Case C does produce for external customers, however only
on stock. (Case C) ‘We only produce on stock, but for [customer 1 and customer 2] we produce specifically
for them on order. However, this is also based on the forecast of these customers. So you have a safety stock, no unique special products.’ Case E is the only company that makes their products only to order. ‘We are focused on a low stock and therefore fully focused on the actual demand. We do not want to produce the wrong product.’
Now that the different production systems in the case study have been established, we review the context in which they are applied. We found that the perishability of the raw materials and processed g oods are most influential with regard to the MTO/MTS decision as well as the planning drivers of the different cases. Perishability
20
MTS due to the perishability of the end-products; ‘If you cannot make the most optimal production run,
but you expect that some order will receive soon, I produce these orders in the same batch. This overproduction is most often the fast movers, since the products can only stay for 3 months in our stock.’
Case D mentions a similar problem with the food coating products: ‘We have 4 months shelf life. Basically,
everything in stock that exceeds four months has to be reprocessed or removed.’ The cases that have
relatively low perishable products are cases B, C and F. (Case B) ‘Some products also have a limited shelf
life. That varies from six months to five years. So that does not have a big impact for us.’ (Case C) ‘We do not include the shelf life in this provision of product groups. Our fries can be kept for 24 months.’ (Case F)
‘Furthermore, we do not include perishability or profit margin.’ Figure 5 shows, the relation between the production system and perishability. The perishability is shown relative to the other cases; high perishability in this research is three to four months shelf life, while low perishability is from 6 months onwards.
Figure 5: Perishability vs production system
What was found is that perishability was the most influential factor in this MTO/MTS decision. Cases that have a relative high perishability such as cases A and E employed a MTO focused production system while cases with a lower perishability, B, C and F were MTS oriented. Case D is the exception here, this can be explained by their limited intermediate storage space; ‘Not because we want to produce 100 tons, but that
is linked to the cell contents of the next silo.’ The flour is not only an end-product but also a raw material
21
production, making MTO infeasible for the flour production. Based on this finding we expected that planning drivers are also of influence to the MTO/MTS decision.
Planning drivers
We distinguish cases that are supply driven from cases that are demand dri ven. Case B is supplied by their suppliers regardless of demand. (Case B) ‘The first limitation is the total available starch. We cannot
produce more than we have starch. We receive a harvest update every month, which we then include in our calculations.’ The other cases are demand driven which means that supplies are ordered based on
expected demand. Here a differentiation can be made between cases with a capacity utilization focus, cases C, D and F and cases that are order driven, cases A and D. (case C) ‘Our service level is 99.5%. But in
this summer we did not have enough capacity to produce everything for the catering industry, then we had to worry about the stock. This can have consequences for the future, for example, that customers walk away or that the sales man predicts extra demand to ensure that it is produced.’ (Case F) ‘The company mainly makes a lot of [main sauce] because we have a potential capacity problem there in the long term. So it is more conservative that we are already in 4 shifts.’ Furthermore, case D mentions that the
profitability of their flour group has a relatively low profit margin. Hence, they produce as efficiently as possible. (Case D) ‘By producing the high-volume products as efficiently as possible.’ On the other hand, cases A & E are more order focused. (Case A) ‘all the products are allocated to a customer when I choose
to start producing a product. I would describe inventory as inventory that I can allocate to different customers, but that is only a small amount of the total inventory. (Case E) ‘as a factory we are very flexible when we talk about responding to customer demand.’ Figure 6, provides an overview of how production
22
Figure 6: Production driver vs production system
From figure 6 we can conclude that case B is the only supply driven company in this case study. Case B is owned by the farmers that supply the company which creates a different dynamic than the other cases that are not owned by their suppliers, resulting in a different approach to the demand and supply matching dynamic than the other companies. However, case B applies a similar MTS oriented approach as cases C, D and F. The difference that is found is that case B reviews the available supply more often than the other cases. However, we found that the production planning drivers of cases C, D and F are similar to that of case B. The focus is on capacity utilization rather than reacting to customer orders as cases A and E. We see there that cases A and E are more focused on flexibility and had a MTO, or MTO oriented production system while cases B, C, D and F were more focused on capacity utilization and had a MTS, or MTS oriented production system. When we compare this to the relative perishability of the materials we can conclude that this is the main driver of the MTO/MTS decision as case B is not affected by being supply driven.
4.2 Capacity coordination
23
Table 6: Tactical (HPP) level decisions
Capacity coordination planning
We make the distinction between the short and mid-term capacity coordination planning based on the longest forecasting period the cases use. This provides a clear overview of the formal forecasting horizon of the different companies, which we can link to the way capacity is coordinated. Cases B, C, E and F have a periodic update in their capacity coordination which is updated for the full forecasting period. Case B has a monthly S&OP planning in which they allocate capacity based on demand, available raw materials, and amount of capacity. ‘You can think of outsourcing or deploying other production lines, cancel purchasing
or sales. It may also be that you have excess capacity and then you can see if you can do a summer break or a planned stop.’ Cases E, C and F all have a formal capacity planning in which future problems and
demands are discussed. Furthermore, minimal lot sizes are determined to minimize downtime (cases C and F) and to create a standard flow (case E). Case E starts with the demand of the clients while case F has a focus on production of their main product. (Case E) ‘It starts with the customer's demand, that
determines what is needed in the next period.’ (Case F) ‘we always produce [main sauce], that will be sold anyway.’
HPP\Cases A B C D E F
Number of FTE 22 1,000 350 120 350 n/a
Capacity coordination
Short-term Mid-term Mid-term Short-term Mid-term Mid-term Longest
forecast
3 weeks 2 years 2 years 3 weeks 1 year 2 years
Who are involved Supply chain planner, team lead and operators Account managers, S&OP planners Account managers & line planner Planners S&OP planning S&OP planning Forecast based on Historic data & experience Historic data and forecast Historic data and forecast Experience Historic data and forecast Historic data and forecast (in)Formal capacity planning
Informal Formal Formal Informal Formal Formal
Formal MTO acceptance/due dates
No Yes No MTO No Yes No MTO
24
Cases A and D have a forecasting period based on experience that exceeds their longest formal forecasting period, which is about three weeks. However, as this is based on the experience of the planner rather than a standard procedure we examine this differently. (Case A) Basically, I can do two things. I can look at
historical data and see that a customer does not have a contract but orders a certain amount of kilos every month. Suppose there comes another customer with an order before then I first look at what I expect more so that I can expand the production run. The other way is to look at the contracts. (Case D) ‘For a few customers we have a forecast of a few weeks ahead that the customer determines. The rest is based purely on experience.’
The existence of formal policy rules influences capacity coordination at the different cases. Case B applies certain cycle times, which they include in their MTO order acceptance. ‘We have a rule that we can produce
an order in 6 weeks. That way we have time to add it to the production plan. It is not that we quickly accept a MTO rush order.’ When we compare this to cases A and D we see that these lean more towards
experience and customers satisfaction. (Case A) ‘The customer order is always leading. That is why the lead
time is so important. The customers know that we have a lead time of max. 15 days. Now, I can consolidate different orders to long production runs. I can promise the customers 2 days lead time, but then I have to change and clean a lot of machines what is inefficient.’ (Case D) ‘(MTO orders) That only happens for the special customers.’ Case E has a formal freeze in their production schedule. ‘On Thursday we transfer the final planning for the coming week to production.’ This is contrary to the approach of cases C and F who
fully utilize their capacity and change their plans during the running week. (Case C) ‘We adjust the schedule
daily. The forecast is the ideal world, but that is never true.’ (Case F) ‘We always plan more than is actually possible. Because when production is faster than expected, operators should have to do something, and by doing it like this, they always can continue with the next order.’
Perishability
From figure 5 we see that cases with a higher perishability tend to be more MTO focused. For cases B, C and F the perishability of the products, both raw materials and processed products , is relatively low. Hence, the focus can be on capacity utilization as there is a low risk of products becoming obsolete. For case B there is a link to the variable supply that is more severe than the other cases. This link will be further explained in the later section on variable supply. Case D is the exception where we see that the perishability is relatively high while the focus is on capacity utilization. This can be explained by the low profit margin of the flour products and that the lead-time of production exceeds the customer lead-time.
25
one and a half days. […] so you cannot deliver the next day.’ Cases A and E focus more on supplying the
right customer orders according to incoming orders. Case C explains this as the need to produce the ‘right product for the customer.’ ‘We do not want to produce the wrong product.’
Sequence-dependent set-up times
In the definition of long sequence-dependent set-up times we include the cleaning times and, for case E, preventive maintenance. Figure 7, shows the existence of long sequence-dependent set-up times to the applied production system. (Case C) ‘2 weeks production, 20 hours of cleaning and then again. […] you also
have change overs which takes time. For example, the knife change or foil change (20 min). The batter change is 30 min.’ (case F) ‘Cleaning takes 8 hours and we want to minimize that so we’ll produce a week later.’ (Case D) ‘..only with the coatings. From color to white, you have 6 hours of cleaning. Switching between the colors is only the product set-up which is 30 to 45 minutes. Cleaning from color to white is 6 hours and from white to color too.’ (Case E) ‘For example, packaging works in 3 shifts from Sunday evening to Friday evening and every Monday from 6 to 12 is autonomous maintenance.’ (Case B) ‘You can sort of clean, but you can also produce some ‘cleaning products.’ (Case A) ‘… on average the cleaning is half an hour, and the total change over takes 1 hour.’
Figure 7: Sequence-dependent set-up time vs production system
26
This is similar to case B who only accepts MTO orders with a lead time of three weeks, even though they use ‘cleaning products’ that reduce their changeover time and increase their capacity utilization.
Variable yield
The variable yield and processing time is closely related to the variable quality of the raw materials. Case A mentions that their capacity coordination is partially based on the smooth running of the production line; ‘Assume that I know that a particular order is coming in a week, and the process for that product is
now running very well, then I decide to plan that product now.’ Case C mentions that their advanced
planning system keeps track of the different production rates. ‘You take into account that 6 tons of one
product lasts longer than 6 tons of other product. It is set up in the line rate that is in the planning tool.’
Case D tells that during certain months the yield is lower due to the quality of the raw material. ‘New
harvest in August always gives a lower extraction. […] You see that after 2 months that it is already much better.’ Furthermore, in case E overproduction is packaged as fast movers. This way they balance their
production. ‘because it is a natural product and your extraction process has a certain yield that you can
extract. So one time it is more and the other time it is less. So that is why we have defined runners that if you produce a little more this week, it will come back next week.’ In figure 8, we compare the variable yield
to the production system.
Figure 8: Variable yield vs production system
27
overproduction from one week to replenish underproduction of other weeks. Furthermore, case E is able to thoroughly analyze their production system and adjust the process where necessary. Case D is a more expected find as it has a relatively high variable yield and is MTS focused. Reason for this is the limited intermediate stocking space and that more profitable products are made from the main product they produce. Hence there should be enough in stock to start production of the more profitable products. Cases B, C and F have the relatively lowest variable yield but nonetheless produce to stock. This can be explained by the low perishability of their products: as the risk of obsolescence is low it is more beneficial to focus on capacity utilization. When reviewing variable yield it is necessary to link it to the quality of the raw materials. This is seen in the multiple cases.
Variable quality
Most of the variable yield is caused by the variable quality of the raw materials. This is dealt with by cases D and E by blending the raw materials. (Case D) ‘It is blended, because you want to deliver an end-product,
with a constant quality.’ (Case E) ‘We blend a number of raw teas. We do this in order to guarantee a certain quality level.’ This was similar to the coffee production. Case F mentions that a constant quality is
ensured by their minimal run length as this ensures the proper mix of their sauces. ‘That is the technical
minimum, which is about 2 or 3 tanks depending on the product. That has to do with the quality you want to achieve.’ We see that for case A and B the variation in quality has a large influence on the capacity
coordination, as the quality limits both end-products that can be produced and the available raw material. We see that this results in a shorter planning horizon and more ad hoc decisions for case A. ‘It might be
that we have problems with the quality of the beans, and that there are exceptions for certain batches. In this case the beans can be used for the one but not for the other batch. […] it makes your planning more difficult. Case B mentions that after knowing the amount of incoming raw materials the amount of
available material for production is still unclear. ‘So if you know how much raw material you have, there is
still uncertainty as you do not yet know the starch percentage. It can vary from 18 to 27%. Case D mentions
the limited possibilities it has to adjust the production process to better the quality. ‘So you actually have
very few moments where you can adjust. So producing 1 ton is possible, but you do not know what comes out. While if you produce 20 tons, you can mix up the first ton that may not be completely top, you can mix with a good product and then you just have a top product again.’ Case C mentions that they are better
than their competitors when the harvest is bad. However, the influence of the quality of raw materials is not mentioned to be a large influence. Moreover, there are some customers who do not want their product to be made during certain months of the year. You have the so-called pre-build products for
28
receives pre-processed raw materials causing a relatively low variation in the materials. ‘We do not use
fresh tomatoes, we use tomato paste. That comes from the supplier with a certain quality value. Depending on the quality of the pasta we have to add pasta or water so that we always work towards a narrow range of amount of solid part of tomato.' In figure 9, we compare the variable quality to the production system.
Figure 9: Variable quality vs production system
The variable quality and variable yield are closely related, and we see that all cases have similar variable quality and yield apart from case B. Case B has a large variety in the quality of the products. However, once the raw materials are processed the variability in yield is controlled. What we see is that the variability of quality and yield does not severely influence the way the production system is organized. However, we do see that it requires some flexibility in the process to ensure production plan adherence in terms of MTO order acceptance (case B) and a production freeze (case E). Otherwise it results in a short planning horizon in which the planning is adjusted often (case A).
Variable supply
The variable supply has a severe impact on case B, as that company is owned by the farmers that supply the company. Due to the variable harvest the company has a variety in supply which they receive. (Case B) ‘Because the company is driven by the harvest that gives us an extra complexity. That potato is harvested
from August to March. So we also have to process all the potatoes in that period, but we do want to produce the whole year.’ Case E mentions that they have some products that have a long lead time which causes
29
there a considerable complexity there for tea. There are many components in a single tea. […] That must all be planned. ' The other cases do not mention such a difficulty with the variability of supply.
Figure 10: Variable supply vs production system
What becomes evident after reviewing the variable supply, quality and yield and processing time is the interrelatedness of the factors. There are several points of interest when reviewing these characteristics and their relationship to the production system. We see that cases C and F have a lower variability in supply, quality and yield. In case C this is mostly due to their raw materials and their ability to produce at a constant level. Case F ensures a constant quality, supply and yield by using pre -processed raw materials. What is remarkable is that even though this results in the lowest variability, there is still a variable yield in the process. Furthermore, case B is severely impacted by the variable quality and supply volume. However, after the first production step the variation in yield becomes lower, similar to the pre-processed raw materials of case F. Furthermore, the variable supply and quality does not shift their production focus to MTO. Finally, case A and E do apply a short planning time fence to which case E adds a formal production freeze to mitigate variable yield and quality whereas case D focuses on capacity utilization.
Workforce
Additionally, it was found that workforce planning has an impact on capacity coordination in some of the cases. Case E mentions that seasonality in demand changes their workforce needs. By creating synergies between production lines they try to scale their workforce based on their needs. ‘packaging is very close
30
long term or what synergy can I achieve.’ Furthermore, the skillset of available workers causes issues in
certain cases. (Case F) ‘that has to do with having a lot of new and many old machines. This complexity
ensures that you have a mix of skills to people who are not always present to solve a particular problem.’
To mitigate this case F created priority lines; ‘Certain lines are given priority, so the experienced people are
put there. Hence it is possible that certain lines actually run very well, but do not run well on paper because they do not have a high priority.’ Case E mitigated the needed skillset by separating the work of operators
in difficult and easy tasks; ‘In the past, we only looked for operators who could keep a whole machine
running and at a certain point we were able to get operators so badly that we checked if we could not divide the work. For example, we have made a split between difficult and easy work where, for easy work, we can take on someone who does not have the skills of an operator but is deployable. ’
Workforce planning is an additional limitation to capacity coordination for cases E and F. Case E tries to create synergies between different lines and found ways to improve the use of skilled workers to mitigate this. Case F, on the other hand, used line prioritization to mitigate this problem. This is only possible due to the fact that case F has overcapacity on certain lines, making this approach possible.
Concluding, capacity coordination is performed differently among the different cases. What became evident is that the existence of sequence-dependent set-up times impacted focus of the production systems more than variable yield, quality and supply. This can be explained by the low perishability of the products that are produced in a MTS oriented system. However to mitigate the variable yield, quality and supply we do see that cases E and B use formal rules to accept incoming orders in order to react to these characteristics.
4.3 Scheduling & Control
31
Table 7: Operational (HPP) level decisions
HPP\Cases A B C D E F
Fixed production sequences
No No Yes Yes/No Yes Yes
Fixed production volumes No Minimal batch size Minimal batch size No Yes Minimal run length Sequence-dependent set-up time
The sequence-dependent set-up times are elaborated upon in the previous section. The effect of this on the applied fixed or dynamic sequencing is reviewed in this section. We see that cases C, E, F, and partly D have a fixed sequencing. Case C has a fixed bi-weekly cycle in which it sequences from light colored to dark colored products; ‘You start with plain and the batter then gets getting darker. You cannot return because
of allergens and gluten-free etc. This is in a cycle of 14 days, 2 weeks of running, 20 hours of cleaning and then again.’ Case E has a similar light heavy sequencing at their tea department while their coffee
department is clustered based on size of end-products. ‘for tea it is important to have certain flavors at
the end of the week. So that the location has the whole weekend to air so that you can make another product on Monday. [...] (for coffee) within that week we cluster the [End-product sizes] and look at the most efficient order of those [sizes].’ Case F ensure that there is only one cleaning in each production week.
‘One of the agreements is that we do not clean during the week. So we can only make special specialties
once a week, because you cannot run them in sequence without cleaning.’ Lastly, case D has a flour line
and a coating line. What is interesting here is that the flour line can produce without a severe set-up while the coating line has a severe cleaning time. ‘With flour everything is produces without cleaning, except if
you have eco products, than you have to clean first. Then you have lost half an hour to three quarters. [...]with the coatings. In color to white, you have 6 hours of wet cleaning. That not deeded between colors than there’s only half hour to three quarters set-up time. Cleaning from white to color is also 6 hours.’ On
32
Figure 11: Sequence-dependent set-up times vs sequencing methods
What we can take from this is that sequence-dependent set-up times are the main cause of the existence of fixed or dynamic sequencing. It is possible to see that fixed or dynamic sequencing is not necessarily related to the production system in which it is employed. Furthermore, we see that the formality of the capacity coordination does not affect the existence of fixed or dynamic scheduling. What is interesting is that the larger companies have a formal minimal batch size while the smaller companies do not. This is in line with what was found at the MTO/MTS decision, where there was a tendency towards customer wishes. Case E is the only case that used a fixed production volume of one pallet of end-products. Workforce
From the interviews the influence of the workforce was also found on the scheduling & control level. Cases A and D make short-term decisions based on their available workforce. (Case A) ‘I have 3 operators in Unit
A and the unit B with 1 operator. Based on the demand and the available operators I look what we can produce today.’ This is similar to case D, however they have a more creative approach to solving the
problem; ‘Machine capacity is not an issue, available workforce is [...] if an operator has been there for 4
hours, then the planner himself will be behind the machine.’ Another interesting finding is the variability
the cases had with regards to their workforce planning. This especially with regards to the competence different workers have. The competence of workers is mentioned by case D; ‘We have a difference in work
teams. That means that every team works in a different way, with different attitudes. Production is working towards standardization: making everyone work in the same way, so that no product or time is lost. Case
F adds to this the influence workers have on machines; ‘When you change, people have to convert
33
deliver this in expected net production hours and what we see is that a lot is deviated from the expectation in practice. So, for example, they state that they do 120 hours and in practice they often do a net of 110 per week.’
34
5. Discussion
MTO/MTS decisionFirst, when reviewing the MTO/MTS decision we found that the companies with a lower perishability level were more likely to have a MTS production system. This is in line with the discussion of Soman et al. (2007) who state that products that are highly perishable are typical MTO products. It was found that this was more likely to be an MTO/MTS indicator than the variation in demand and volume as suggested by Soman et al. (2007) and Perona et al. (2009). However, case D has a relatively high perishability while being MTS focused. This was due to the relatively low margin and the inability to produce smaller batches. The aggregated demand for most of the volume is know which makes MTS focused production applicable (Van Donk, 2001). When we put this in line with the discussion of Romsdal et al. (2013) it is interesting to conclude that the companies with the lowest perishability and demand uncertainty apply a MTS focused system. This is contrary to the expectation of Romsdal et al. (2013) who expected that MTO would be applicable in cases of low perishability and low demand uncertainty. From our results we find the opposite, where we reflect that the reason for this is the expected capacity shortage and seasonality in supply. As a result these companies are more focused on capacity utilization rather than reacting to short-term demand.
Capacity coordination
When comparing cases that focus on capacity utilization with order focused cases, we find that sequence-dependent set-up times are important. Cases C and F are fully committed to utilizing their available capacity. As was concluded by Stevenson and Found (2016) the relation of sequence to capacity utilization is very important in these cases. Due to the high sequent-dependent set-up times and relative low perishability these cases can fully utilize their capacity. On the other hand, cases with lower sequence-dependent set-up times and a relatively higher perishability were more order focused. An unexpected find is that variable yield and processing time was mostly found in MTO focused cases. This is counter intuitive as Stevenson and Found (2016) proposed that it is more likely that this would result in MTS production: inventory buffering could mitigate the risks related to yield and processing time. Rather than that cases A and E use flexibility in their capacity coordination to react to the variation in demand as is suggested by Romsdal et al. (2013). Case A employs a short planning to adjust to this, while case E uses intermediate storage to store overproduction of fast movers.
35
findings of Perona et al. (2009) and O’Reilly et al. (2015). The determination of production sequencing and production volumes was mostly based on experience. Nevertheless, no matter the size of the company, fixed production sequences are applied in situations where long sequence-dependent set-up times are present. This is in line with expectations of Soman et al. (2004) and it is interesting that this can be confirmed. Case B uses a ‘cleaning product’ on some lines to mitigate this long sequence-dependent set-up times. This resolution has not been previously described in literature and is a way to optimize the capacity usage. When we look at the formal production volumes we see that the larger companies have a formal batch size while the smaller companies do not. This can be linked to what was found by O’Reilly et al. (2015) who found that smaller companies have a tendency towards ‘flexibility’ rather than effectiveness. This is caused by large focus on customer service and a lack of a clear decision-making policy. This causes ad hoc decision making in production as described by case D as ‘putting out too many fires.’ Scheduling and control
Another interesting finding is that smaller companies partly base their final schedule on the available workforce. The limitation of skilled and available workforce became evident in this research. The only previous literature where skilled and semi-skilled workers were investigated was by Khakdaman et al. (2015). In their robust optimization planning this limitation is considered in their model. Stevenson and Found (2016) state that cross training of workers is desirable to cultivate a multi skilled workforce to enhance process knowledge and enable accommodation of variation to minimize waste. We see that workforce limitation is dealt with in multiple ways. Case E mentions cross training in order to deal with the variation in orders. On the other hand, cases A and F focus on the smooth runnin g of machines and try to minimize the number of set-ups to avoid human errors. These approaches are not related to a MTO or MTS focused system but rather one of production planning adherence. Case E was the only case that mentioned maintaining a freeze of the production schedule. While cases C, F and B, also mentioned a freeze this was more a guideline rather than a formal freeze as all cases mention adjusting the production schedule quite often.
FPI characteristics vs production systems
36
37
6. Conclusion
This research has analyzed the factors that influence production systems in the FPI by performing a multiple case study. The theoretical contribution of this research is twofold. First, we can confirm the expectation of Soman et al. (2004) that tailor made production and inventory management systems are needed in the FPI. What we found is that due to the fact that not all companies had hybrid production systems, there were different approaches in organizing production systems in relation to the FPI characteristics described by Van Donk (2001). We found that the level of perishability and the existence of sequence-dependent set-up times are the main influencers of the production system organization. However, due other variables such as variable quality, supply and production yield we did not find an easy copy and apply system that could be used across multiple cases. We believe that this is also due to the formality of decision-making in a company as described by O’Reilly et al. (2015). Instead, each of the studied companies is a unique position to determine their own production system based on the discussed influential factors. Even for the cases that were most similar we do not believe that the system can be easily copied.
We believe that the factors with the most severe impact on the applicability of a production system and the decisions to use a production system are the level of perishability of products and the sequence-dependency of the production process. On the other hand, the variation in quality, supply and processing yield were found to have less of an effect on the production system in the cases we reviewed. Several unexpected findings were done that led to this conclusion. First, the MTO focused cases were more likely to encounter a variable yield and quality. The other unexpected finding was the impact of the workforce on capacity coordination as well as scheduling and control. We found that the impact of an unavailable or unskilled workforce can be severe in cases were changeover times and capacity utilization are critical. Companies can mitigate these issues by either prioritizing production lines, reviewing jobs, or using overtime.
38
that includes a freeze schedule. This limits ad hoc decision making and gives some possibility to adjust to unexpected events such as variable yield and quality.
Limitations and further research
One of the limitations of this research is the limited number of available hybrid production cases. This was not foreseen; however it has turned into an advantage of the research. This research has had the chance to review the differences and similarities of hybrid production systems and ‘pure’ production systems in the FPI. To the authors’ knowledge this has been the first research that has been able to do this. As a result the focus was more on the applicability of production systems in certain contexts of the FPI than comparing hybrid production systems. Another limitation was that there was only one supply driven company, which limits the explorative options of the research for certain FPI characteristics.
39
7. References
Aitken, J., & Garn, W. (2015). Agile factorial production for a si ngle manufacturing line with multiple products products. European Journal of Operational Research, (April).
Akkerman, R., & van Donk, D. P. (2009). Analyzing scheduling in the food-processing industry: Structure and tasks. Cognition, Technology and Work, 11(3), 215–226.
Beemsterboer, B., Land, M., & Teunter, R. (2017). Flexible lot sizing in hybrid make-to-order/make-to-stock production planning. European Journal of Operational Research, 260(3), 1014–1023.
Beemsterboer, B., Land, M., & Teunter, R. (2016). Hybrid MTO-MTS production planning: An explorative study. European Journal of Operational Research, 248(2), 453–461.
Calle, M., Gonzalez-R, P. L., Leon, J. M., Pierreval, H., & Canca, D. (2016). Integrated management of inventory and production systems based on floating decoupling point and real-time information: A simulation based analysis. International Journal of Production Economics, 181, 48–57.
Claassen, G. D. H., Gerdessen, J. C., Hendrix, E. M. T., & van der Vorst, J. G. A. J. (2016). On production planning and scheduling in food processing industry: Modelling non-triangular setups and product decay. Computers and Operations Research, 76, 147–154.
Eisenhardt, K. M. (1989). Building Theories from Case Study Research Published by : Academy of Management Stable URL : http://www.jstor.org/stable/258557 Linked references are available on JSTOR for this article : Building Theories from Case Study Research. The Academy of Management
Review, 14(4), 532–550.
Karlsson, C. (Ed.). (2016). Research Methods for Operations Management. 2nd edition. New York: Routledge.
Khakdaman, M., Wong, K. Y., Zohoori, B., Kumar, M., & Merkert, R. (2015). Tactical production planning in a hybrid Make-to- Stock – Make-to-Order environment under supply , process and demand uncertainties : a robust optimisation model. International Journal of Production Research, 53(5), 1358–1386.
Kerkkänen, A. (2007). Determining semi-finished products to be stocked when changing the MTS-MTO policy: Case of a steel mill. International Journal of Production Economics, 108(1–2), 111–118. O’Reilly, S., Kumar, A., & Adam, F. (2015). The role of hierarchical production planning in food
40
Perona, M., Saccani, N., & Zanoni, S. (2009). Combining make-to-order and make-to-stock inventory policies: an empirical application to a manufacturing SME. Production Planning & Control, 20(7), 559–575.
Pool, A., Wijngaard, J., & van der Zee, D. J. (2011). Lean planning in the semi-process industry, a case study. International Journal of Production Economics, 131(1), 194–203.
Rafiei, H., & Rabbani, M. (2012). Capacity coordination in hybrid make -to-stock/make-to-order production environments. International Journal of Production Research, 50(3), 773–789.
Rafiei, H., Rabbani, M., & Alimardani, M. (2013). Novel bi-level hierarchical production planning in hybrid MTS/MTO production contexts. International Journal of Production Research, 51(5), 1331–1346. Romsdal, A., Arica, E., Strandhagen, J., & Dreyer, H. (2013). Tactical and Operational Issues in a Hybrid
MTO-MTS Production Environment : The Case of Food Production To cite this version : Advances in production management systems: Competitive manufacturing for innovative products and services. Seawright, J., & Gerring, J. (2008). Case Selection Techniques in A Menu of Qualitative and Quantitative
Options. Political Research Quarterly, (1975), 294–308.
Shahvari, O., & Logendran, R. (2016). Hybrid flow shop batching and scheduling with a bi -criteria objective. International Journal of Production Economics, 179(November), 239–258.
Soman, C. A., van Donk, D. P., & Gaalman, G. J. C. (2004). Combined make-to-order and make-to-stock in a food production system. International Journal of Production Economics, 90(2), 223–235.
Soman, C. A., van Donk, D. P., & Gaalman, G. J. C. (2006). Comparison of dynamic scheduling policies for hybrid make-to-order and make-to-stock production systems with stochastic demand. International
Journal of Production Economics, 104(2), 441–453.
Soman, C. A., van Donk, D. P., & Gaalman, G. J. C. (2007). Capacitated planning and scheduling for combined make-to-order and make-to-stock production in the food industry: An illustrative case study. International Journal of Production Economics, 108(1–2), 191–199.
Stevenson, S., & Found, P. (2016). Understanding the lean enterprise: Strategies, methodologies, and principles for a more responsive organization. In Understanding the Lean Enterprise (pp. 153–184). Strijbosch, L. W. G., Heuts, R. M. J., & Luijten, M. L. J. (2002). Cyclical packaging planning at a
pharmaceutical company. International Journal of Operations & Production Management, 22(5), 549–564.
41
Van Donk, D. P., & Van Dam, P. (1996). Structuring complexity in scheduling: a study in a food processing industry. International Journal of Operations & Production Management, 16(5), 54–63.
Van Kampen, T., & Van Donk, D. P. (2014). Coping with product variety in the food processing industry: the effect of form postponement. International Journal of Production Research, 52(2), 353–367. Voss, C. (2009). Case Research in Operations Management. New York: Taylor & Francis.
Yin, R.K. (2009). Case study research: Design and Methods, 4th edition. Thousand Oaks: Sage Publications.
42
Appendix A
Interview: xx-xx-2017
Place: Opening
• Introduction of interviewer and interviewee
• Confidentiality assurance
• Permission to audiotape
General
• Job-title of participant
• Company description:
o Characteristics such as; size, market share and geographical locations
o Number of employees (plant and organization)
o Turnover plant in terms of money and volume
o
Location in entire supply chain• Could you give a rough description of the manufacturing process from raw material to
final products?
• At what point is the product group customer specific?
Product groups/families1. Do you subdivide products in families/groups?
2. On which characteristics are these products grouped?
a. On what specific product characteristics are these products grouped?
- Perishability
- Profit margin
- Variability in quality (raw materials and finished goods)
b. On what specific market characteristics
- Actual customer demand/orders
- Forecastability of orders
faculty of economics and business
43
