Understanding the influence of company-level factors affecting the MTO/MTS decision in hybrid food processing organizations.

(1)

1

Understanding the influence of company-level

factors affecting the MTO/MTS decision in hybrid

food processing organizations.

Master Thesis

Msc. Technology and Operations Management University of Groningen

Faculty of Economics and Business

Quinten Olde Riekerink q.p.olde.riekerink@student.rug.nl

S2765535

First assessor: Prof. Dr. D.P. van Donk Co-assessor: Dr. O.A. Kilic

(2)

2

Abstract

Hybrid production systems combine the benefits of Make-to-Order and Make-To-Stock systems, but they require a structured process for their MTO/MTS product classification to reap the full benefits of a hybrid system. In this classification, company-level factors are often neglected. This study looks at how hybrid food processing organizations partition their products, and how capacity considerations, and other company-level factors, influence the partitioning decision. Using semi-structured interviews and observational data, four Dutch and four Greek cases are analyzed to better understand the partitioning process. On a product-level, risk of obsolescence and demand patterns greatly influence the production strategy preference. On a company-level, contractual agreements, supply patterns, capacity constraints and financial resources determine the final MTO/MTS capacity distribution. This study proposes a hierarchical approach that aids the partitioning process, combining the most important product- and company-level factors.

Key words: hybrid production system, make-to-order (MTO), make-to-stock (MTS), food processing industry (FPI), capacity constraints, multiple case study research.

Preface

The paper that lies before you is the final project for my Master degree in Technology and Operations Management. Finalizing this study at the University of Groningen would not have been possible without help from others. Therefore, I want to express my gratefulness to my supervisors prof. dr. D.P. Van Donk and dr. O.A. Kilic for guiding me during the writing of this paper with feedback and suggestions. Additionally, my student colleague Nikolas Dimitrakopoulos fostered the discussion on a lot of aspects in this paper, for which I am grateful as well.

Moreover, this research would not have been possible without the options provided by the participating companies that were willing to spend their time and put in effort to provide data. This way I could learn a lot about the food processing industry and its planning and scheduling issues. Seeing how these different companies go by in their everyday tasks and hearing how problems described in the literature are tackled on a daily basis has been an educational experience. Lastly, I would like to thank everyone that helped motivate me during the plentiful hours in the library and at home. Specifically, my parents and girlfriend have helped me get through the times I was struggling, for which I want to say thanks.

(3)

3

List of Tables and Figures

1. Figure 2.1: Simplified HPP framework for hybrid systems (Soman et al., 2004, p. 232) 7 2. Table 2.1: Food Process Industry characteristics (Van Donk, 2001, p.300) 9 3. Table 2.2: Partitioning criteria and the FPI specific influences 10 4. Figure 2.2: Proposed model for the partitioning process in the FPI 11

5. Table 3.1: Case company details 13

6. Table 3.2: Interview details 14

7. Table 3.3: Overview of case company characteristics 15

8. Table 3.4: Coding classes and examples 16

(5)

5

1. Introduction

Following market demands, food processing organizations have increased their range of both products and packages (Claassen et al., 2016). The desire to produce more customized products (Zandieh & Motallebi, 2018) and technological developments (Federgruen & Katalan, 1999) led to producing make-to-stock (MTS) and make-to-order (MTO) products in one hybrid system (Williams, 1984; Beemsterboer et al, 2016), combining the benefits of both systems (Beemsterboer et al., 2017). There are however no well-established methods towards planning and control of hybrid systems (Beemsterboer et al., 2017; Zandieh & Motallebi, 2018). It is clear (intuitively at least) that planning and control of hybrid systems becomes more straightforward when a significant part of capacity is allocated to either MTO or MTS products, as planning and control activities could then simply be designed around the dominant production strategy. However, the benefit of hybrid production systems is much larger when capacity is distributed evenly between MTO and MTS products (Beemsterboer et al., 2016; 2017). This underlines the importance of the distribution of capacity in MTO/MTS partitioning, as the interaction between products and capacity is critical in determining the production strategy (Williams, 1984; Beemsterboer et al., 2016). Interestingly, this has mostly been neglected in the literature (Zaerpour et al., 2009).

Setting production strategies is of strategic importance (Soman et al., 2004) this is however generally viewed as a product-by-product activity through evaluating a set of characteristics, as for instance in Van Donk (2001). Company-level factors, such as capacity distribution, are not taken into account. Academics’ focus has been on quantitatively solving models to determine a products’ production strategy (Beemsterboer et al., 2016). However, qualitative research can identify new factors and relations that can improve these models (Perona et al., 2009), therefore this study explores the influence of company-level factors on partitioning.

The research will be set in the food processing industry (FPI), as industry specific issues and characteristics make combined MTO/MTS production systems quite common (Soman et al., 2004), however, partitioning is more challenging as well (Van Donk et al., 2005). To understand how the FPI specific issues and characteristics influence the MTO/MTS decision, and what role company-level factors play in this process, a multiple case study will be conducted to answer the following research question:

‘How do company-level factors influence the partitioning of hybrid food processing organizations?’

(6)

6

rely on common sense and experience instead (Perona et al., 2009). This study contributes to making tacit decision-making processes explicit, focusing on the influence of company-level factors. Providing a more detailed list of factors aids future research aiming to developing decision making tools. Moreover, understanding the influence of company-level factors on partitioning leads to better informed decision making. Additionally, logic-based rather than computer-based models are needed to aid smaller firms in their MTO/MTS decision.

(7)

7

2. Theoretical background

In this section literature on hybrid production systems will be discussed, starting with the general research directions, followed by the partitioning decision. Next, the characteristics of the food processing industry are discussed to highlight their influence on the partitioning problem.

2.1 Hybrid production systems

Originally, researchers investigating production systems focused on either pure MTO or MTS production systems (Soman et al., 2004; Garn & Aitken, 2015), or start from the perspective of a pure system and modify it to deal with hybrid system issues (Beemsterboer et al., 2016; Iravani et al., 2012). Lately, hybrid production systems attracted academics’ attention (Rafiei & Rabbani, 2012), showing its potential value in for instance Barton (2013) and Yousefnejad et al. (2019). However, research in this field is mainly focused on quantitatively solving partitioning models (Beemsterboer et al., 2016). The most notable qualitative development on this topic has been the HPP framework by Soman et al. (2004). Their framework, presented in Figure 2.1, started the discussion on hybrid production systems in the FPI, however it is mainly conceptual and lacks shop-floor considerations.

Figure 2.1: Simplified HPP framework for hybrid systems (Soman et al., 2004, p. 232)

(8)

8

maintaining service levels (Beemsterboer et al., 2016). This research investigates if adding company-level factors to the HPP framework presented in Soman et al., (2004) can help reap the rewards of hybrid production systems.

2.2 The partitioning process

Research on hybrid production systems was initiated by Williams, (1984) that discusses the different impact of ‘special’ and make-to-stock products. Later, Hoekstra and Romme, (1992) introduced the concept of the customer order decoupling point (CODP), which identifies the point in a production process from which an order becomes customer specific. Research dedicated to the CODP, alternatively mentioned as the order penetration point (OPP) in Olhager (2003), identifies characteristics that push a product either downwards (towards MTS) or upwards (towards MTO). Additionally, a model or tool is developed that, given the characteristics, determines the production strategy, as for instance in Van Donk (2001). Determining which products are MTO and MTS is of vital importance for planning and control activities (Hemmati & Rabbani, 2010). The influence of the CODP on production strategies is explained both in general e.g. Olhager (2003) and the FPI e.g. Soman et al. (2004), Van Donk et al. (2005) and Zandieh and Motallebi (2018). However, these studies fail to address properly the influence of company-level factors on partitioning and system performance. Zaerpour et al. (2009) fill a pre-determined amount of capacity for each production strategy, however, it is assumed organizations have specified capacity distribution in advance and is rather straightforward (Rafiei & Rabbani, 2014).

It is clear (intuitively at least) that different capacity distributions result in different performance levels for different products. Allocating more capacity to MTS production will decrease the chances of stock-outs for MTS products, reducing cost of lost sales. However, this can result into an increase in delivery lead-time for MTO products, increasing cost of lateness (Van Donk et al., 2005; Beemsterboer et al., 2016). However, how capacity is distributed, and how company-level factors influence this decision is not known.

2.3 The MTO/MTS decision in the FPI

(9)

9

Table 2.1: Food Process Industry characteristics (Van Donk, 2001, p.300)

Later, for instance Hemmati and Rabbani (2010) have extended this list for the general partitioning problem, adding ‘supplier-related criteria’ to the market-, product- and process-related criteria, however, they view the partitioning problem only on a problem-level. Zaerpour et al., (2008; 2009) propose a list of ‘product-related criteria’ and ‘firm and process related criteria’, where customer and supplier commitment influence partitioning on a ‘system-level’. Extending the list of criteria was required to cover a broader range of industries, and to create more reliable decision-making tools (Hemmati & Rabbani, 2010). As a starting point for this research, the FPI characteristics introduced by Van Donk (2001) are integrated with the partitioning criteria from Hemmati and Rabbani (2010), as presented in Table 2.2.

Organizations will mostly determine their production strategy based on market and product characteristics (Van Donk et al., 2005; Rafiei & Rabbani, 2011, 2014; Garn & Aitken, 2015). For instance, the combination of the risk of obsolescence (product) and delivery lead time requirements (market) of a product has a large impact on partitioning (Hemmati & Rabbani, 2010). However, in some cases the MTO/MTS decision is not so straightforward, for instance when product variety is high and market conditions become more dominant (Kerkkänen, 2007). Furthermore, SME’s unstructured decision-making processes requires a more distinct, easy-to-use classification of factors and an explanation on how they should be used (O’Reilley, 2015). However, process related criteria like long sequence-dependent setups push the CODP downwards in a product-level (Van Donk, 2001) but might influence company-level capacity distribution as well, complicating the classification process.

Plant characteristics Product characteristics Production process characteristics Capital intensive, single

purpose capacity.

Variable supply, quality and price of input materials

At least one of the processes consists of homogenous goods Shop floor-oriented design Volume or weights is used

instead of number of products

Packaging is labor intensive when producing for end -consumer

Small product variety with high volumes

Products are perishable throughout the process

Processing is capital intensive; capacity determines production rate

Long sequence dependent

setups Divergent product structure

Multiple recipes for same product due to variability in supply

(10)

10

Criteria Explanation Integration with FPI

Produ

ct

Risk of obsolescence

Product value reduction over time

Perishability of goods affects required delivery time and profit margins Product type Generic products allow for

creating inventory whereas specifics are made to order

Few intermediate products with a

divergent product portfolio structure in the packaging stage

Cost of each item Product value increases holding cost and therefore should be stocked less

Low cost items and small profit margins with potentially expensive additives for specialized products Mar ke t Predictability of demand

Predictable demand allows usage of MTS production

The competitive nature and dynamic environment of the FPI make demand less predictable

Product customization

The level of customization required by customers

Market pressure to increase product and packaging variety

Required delivery reliability

Service level requirements set by the customer

Consumer switching cost are often low in the FPI due to variety of alternatives at point of sale

Required delivery time

Responsiveness

requirements needs stock buffers

Customers require a certain shelf life when the product arrives, resulting in less room for stock

Proces

s

Production lead time High production lead time requires buffers to deal with delivery time

Capacity determines the production rate, perishable nature reduces stock levels and allowed production lead time

Process and human flexibility

The sharing of resources for different products

Long sequence dependent setups in processing and labor-intensive packaging limits flexibility

Holding Costs Inventory levels and costs tie down capital

Perishability is a large determinant of inventory levels Su pp li er Supplier commitment

The constant availability of raw- and packaging materials

Large variability in quality and quantity of raw materials due to dependence on e.g. weather conditions

Pricing policy of retailers

The power of customers to determine the price

Often many points of sale, but seasonality influences pricing

(11)

11

As is clear from the literature discussed above, there is no clear distinction between what factors influence the CODP on a product-level, and what factors influence capacity distribution on a company-level. However, it is believed that the combination of the product-level CODP preference and the company-level capacity distribution will determine the final MTO/MTS decision, as illustrated in Figure 2.2. In this study capacity distribution is addressed as the percentage of total capacity allocated to either MTO or MTS production. The CODP preference is the preferred production strategy for a single product if it were the only one being produced by the system. This study attempts to identify dominance of one criteria over the other, both on a product and a company level.

(12)

12

3. Methodology

In this section, the chosen research method, the selection of case companies, and the collection and analysis of data is explained.

3.1 Research method

The aim of this study is to investigate how company-level factors influence the partitioning decision in hybrid production systems, adopting a qualitative approach. An explorative approach is appropriate due to the limited amount of explicit knowledge on company-level factors (Karlsson, 2016). In-depth analysis of FPI specific characteristics is required to understand the complex interactions between company-level factors and partitioning (Yilmaz, 2013). To do so, multiple cases of hybrid production systems are reviewed to find commonalities and patterns in the decision-making process. Multiple cases are considered to allow for a more meaningful outcome for the research area (Karlsson, 2016). Furthermore, it allows for a more direct comparison of the partitioning processes observed in different cases (McCutcheon & Meredith, 1993).

In this study, the field-based approach of the case study method is important, as it allows for investigating the partitioning process in its natural setting (Campbell et al., 2013; Voss et al., 2002). Understanding the reasons behind the MTO/MTS decisions leads to theory building based on primary data (Karlsson, 2016), which is desirable as theory on partitioning is still in its infancy (Beemsterboer et al., 2016).

3.2 Case selection

The most important consideration in the selection of cases is a mixture of both MTO and MTS products within one production facility. Furthermore, a variety in MTO/MTS capacity distributions is necessary to understand how company-level factors influence capacity distribution (McCutcheon & Meredith, 1993). Differences in cases allows for direct comparison of reasons leading to either a high MTO or a high MTS portion. Additionally, an effort was made to incorporate companies of different sizes and from different sectors within the FPI, in an attempt to make the sample as representative as possible. To increase validity, cases were selected using the replication logic (Karlsson, 2016). Only selecting hybrid food processing organizations shows the use of literal replication, whereas selecting cases from different sectors, sizes (indicated as number of employees and products), and different capacity distributions (indicated as % of capacity allocated to either MTO or MTS production) shows the use of theoretical replication (Karlsson, 2016).

(13)

13

driven production. Moreover, they were asked to give an indication of the percentage of stock and order driven production in order to get an even spread of capacity distributions (Karlsson, 2016). After initial contact, e-mails were sent to elaborate on what was expected from the company and the researcher. Table 3.1 provides an overview of the company details.

Case Sector Number of

employees Number of products Weekly production hours %MTO %MTS A Alcoholic beverages 15 8 120 80 20 B Flour processing 282 750 168 40 60 C Wheat processing 127 200 120 65 35 D Wheat processing 264 180 168 70 30 E Alcoholic beverages 30 250 40 10 90 F Meal preparation 1000 40.000 66 75 25 G Milk processing 3000 300 168 5 95 H Meat processing 350 125 92,5 70 30

Table 3.1: Company details

3.3 Data collection

(14)

14 Case

Inter-viewees Function

Duration

(min) Data gathered Nationality A 1 Production & quality manager 55 Interview & tour Greek

B 1 Plant manager 75 Interview & tour Greek

C 3 VP-CEO / quality + / sales manager 70 Interview & tour Greek D 2 Production + / logistics manager 70 Interview & tour Greek E 2 Financial+ / production manager 65 Interview & tour Dutch F 2 Production manager + planner 45 Interview & tour Dutch G 2 Production + / supply chain manager 65 Interview & tour Dutch

H 1 Planner 50 Interview & tour Dutch

Table 3.2: Interview details

Data has been collected by two researchers, one in Greece, one in the Netherlands, each operating in their country of origin to reduce the potential pitfall of language barriers. All interviews were combined with a tour through the production facility, in order to triangulate data and put it into perspective. In addition to the spoken descriptions, cases A-D also provided a printed version of their production process.

3.4 Data analysis

(15)

15

Table 3.3: Coding tree example

3rd

order 2nd

order

1st_order _{Example quotes}

Pro d u ct -lev el cr iter ia Pro d u ct relate d c riter

ia risk of obsolescence H: In the freezer we give them a year shelf life. | Basically, the MAP is packing with a certain combination of gases, including Co2. So

that it lasts longer. In the vacuum It is 13 days, after that it is no longer good.

product type C: Yes, of course this is very important when making the MTO/MTS decision and is normally also related with the cost of the raw

materials and whether more expensive raw materials are used that are typical in more specialized products.

variability of raw materials

A: Of course. It is vital, this is linked directly with the weather conditions though, which vary from year to year. So, the grapes come out better depending on whether the year is dry and things go well. There are however years that it can rain constantly a

cost of each item G: We also communicate with commerce, what is the price of the [product] in the market.| If the milk price is now low, but the

commerce thinks that it is going to rise soon then we are going to make a lot now. Then we make a lot of stock

Ma rk et relate d cr iter ia

predictability of demand B: The most important criteria that relates to the MTO/MTS decision is the predictability of demand as it defines whether a product can be produced with smaller stock or needs higher stock or without any stock if it is a specialized product.

product customization requirements (PL)

C: Everyone wants more customized product that fits their needs and smaller lead times ...we have seen this shift in the market … increased product codes that are requested from us, especially from supermarkets who want to include everything they can.

delivery time & reliability

C: Delivery reliability and delivery time as terms are like brother and sister, they are interrelated... the most important aspect regarding our service level is that we have reached a level of organization … that allows us to be very responsive

Pro ce ss -r elate d cr iter ia

production lead time D: The number one criterion we look at when deciding on MTO/MTS is the processing time and this is logical as the processing time

might be a restricting factor towards MTO or MTS.

holding costs B: Yes, this is also an important factor, but for us it isn’t that relevant because we have high inventory turns. If we see it as a quantity

as a whole and not according to products, I would say 3 times per week. I mean stock of the semi-finished products.

process and human flexibility

F: The customer could ask for 1.2 kilos or 2.4 or 5.6 or 3.8. We made everything. | In terms of volumes we are on I think 70-75% of the max. But we are free to stretch our working hours.

C o m p an y -lev el cr iter ia Su p p lier relate d cr iter

ia Supplier commitment A: I try to create an informal contract with the producers of grapes in order to ensure that at least a large portion of the production _{will arrive ... although I am constantly trying to also find some new as our production capacity increases.}

pricing policy of suppliers

H: Yes, if the prices go down for our breast caps. Then our cost price goes down, and then we reduce the prices too. And then we hope that the customers will order more because the prices are lower.

Pro d u ctio n p ro ce ss Customer order decoupling point

B: There is no clear point, every product is different. Because an order might refer to a quantity of a typical product code and another might refer to something very specific ... If you work on the basis of stock, you make sure to replenish the stock.

Product families E: I think we have 55 products in stock. But if you look at all the labels, you’ll soon get to 250 products we make. This has to do with

herbal bitters that we make for the smallest wearers. | Every month we spend 1.5 to 2 days for the small things

Stockouts/ unexpected orders

D: When we see that backlog might occur, we might be behind schedule and stocks are not enough to cover the demand then we communicate with sales in order to make arrangements and try and find a compromising solution with our customers.

Pro d u ctio n p ro ce ss co n str ain ts Processing time constraints

B: The production process has a specific capacity … Given that orders are not evenly distributed in time in their quantity, it is always possible for a bottleneck to occur.

Financial constraints A: the government in recent years has requested that we prepay tax based on our prediction of what we will sell the next year. We are

drained from all sides from our liquidity and we are lucky to be able to operate and survive in this hostile

C o n tr ac tu al ag ree m en

ts _{Strategic partnerships} _{G: [Company G] has an appointment with the xxx, and a joint factory that slices all the cheese for the xxx. And for xxx we also make}

cheese, and which is also made of milk from farmers who are affiliated with xxx. That is a closed chain

Private labelling D: In private labels we have contractual agreements as regards to product specifications of final products and the same holds with

(16)

16

Using the codes presented above, the researchers scored each case company on a set of characteristics, as indicated in Table 3.4. Scores rank from Very Low (1) to Very High (5) and are based on the coded data. Looking at for instance perishability, case F mentioned: ‘Our shelf life does not allow us to produce

in stock… If I have only 5 days of shelf life I can not start producing in stock.’ On numerous occasions

it was mentioned as an important factor they had to deal with in all of their processes, therefore perishability of both inputs and outputs was given the highest relative score (5), MTS production was only made possible through using different cooking techniques. For case G, the inputs are highly perishable, requiring processing within 2 days, and thus a relative score of (5) was allocated to input perishability. However, their final products are barely influenced by perishability; ‘It is a product that

can mature for a year.’ Which indicates the differences between inputs and outputs and highlights the

need to separate the two. Furthermore, E works with non-perishable raw materials, and does not consider perishability of final products to be an influential factor: ‘That's quite long, we don't set a shelf life date.’ This resulted in a Very Low (1) score in comparison to the rest of the sample.

Additional examples of how the scores were established are given in Appendix C. Furthermore, the detailed first order coding for each case is presented in Appendix D. The %MTO/MTS relate to the percentage of total capacity being allocated to each of the production strategies, and the number of total production hours indicates whether or not total capacity could be expanded without investing in new production technologies. Characteristic A B C D E F G H %MTO/MTS 80/20 40/60 65/35 70/30 10/90 80/20 5/95 70/30 Production hours 4 5 4 5 2 2 5 3 Product variety 2 4 3 2 3 4 3 2 Perishability (Inputs) 3 1 1 2 1 5 5 5 Perishability (Outputs) 1 2 1 1 1 5 2 3 Leadtime 1 2 2 2 2 1 3 1 %Private label 5 3 4 4 2 1 4 3

(17)

17

4. Findings

In this chapter the findings will be presented. First, an overview of each case is given to understand the difference between product and company level criteria. Next, the influence of each of the criteria on the MTO/MTS decision is analyzed. Lastly, the combined effects of both product and company level criteria are analyzed, looking at how they determine the MTO/MTS decision.

4.1 Identifying product and company-level factors

Case A (80% MTO): Company A produces alcoholic beverages, which are non-perishable and few in number (variety): It is white, red and rose wine as well as [product]. Also, there is a categorization of

dry wine, semi-sweet … In [product], there is a categorization with anise and without anise.’ However,

changeover costs are still a problem: ‘I mean we have to make countless switchovers that require

significant time.’ Furthermore, demand patterns are stabilized through a strategic partnership that buys

their main product, which takes up approximately 80% of total capacity. A stable demand pattern, low risk of obsolescence and product variety suggest MTS production on product-level. However on a company-level, contractual agreements significantly increased MTO production, resulting from a continuous order flow, which reduced A’s holding cost.

Case B (40% MTO): Working with flour means B is barely influenced by perishability; ‘flour can live

for 6 months’ and holding cost are reduced through high inventory turns; ‘I would say 3 times per week. I mean stock of the semi-finished products.’ This results in a low risk of obsolescence. Furthermore,

predictability of demand is determined by customer segment and product type. Supermarkets buy smaller bags as either private label (30% and MTO) or the companies’ brand (70% and MTS). Secondly, craft industries and bakeries buy larger sacks (70% and MTS), and order more customized products (30% and MTO). Thirdly, industrial clients order complete silos or tanks, which are produced primarily MTO (90%) as otherwise holding costs become too high: ‘You cannot have tanks or silos of stock. We

don’t hold serious stock for these we normally have a stock of 2 days.’ Retailers’ demand is rather

unpredictable: ‘The actual customer orders we normally learn them in a very small time frame and thus

we cannot know it beforehand.’ Whereas demand patterns for the other groups are more predictable: ‘if we know that client C obtains every week 250 tons, even if they haven’t informed us today, for next week we know that the requested quantity will be more or less similar.’ In this case, the predictability of

demand and holding costs determine the product level CODP. The standard companies’ brand products are unpredictable and thus produced MTS. Private labels, which have more predictable demand and a higher maximum delivery time, are produced MTO. Product level-criteria are more dominant in this case, however partitioning is also influenced by contractual agreements for private labels and industrial customers.

(18)

18

MTO/MTS: ‘To give an example, we can have [product] of 400 grams flavored and be producing and

packaging it for 6 or 7 private labels.’ Partitioning is therefore very dependent on the type of packaging;

private label is MTO and own brand is MTS: ‘a product that has stable demand can also be produced

more to order and have less stock, whereas a product that has variability in demand should have more stock to account for this uncertainty.’ On a product-level, predictability of demand and production costs,

which are associated with product type, determine the CODP: ‘The most important aspect in the process

related criteria is the cost per unit production.’ On a company-level however, capacity distribution

varies as a consequence of seasonality: ‘So our most produced and traditional product in those 70

years…, has a certain seasonality and as such forces us to have stock.’ In the months leading up to

Easter they dedicate up to 50% of their capacity to satisfy peak demand, after which they scale down to reduce holding costs: ‘This is in line with our effort to operate with minimum stock and be lean, because

stock is money that sits around doing nothing. We want to have the necessary stock quantities only for the equivalent time period.’ Additionally, market pressure has shifted capacity distribution towards

MTO: Everyone wants more customized product that fits their needs and smaller lead times … we have

seen this shift in the market in the form of increased product codes that are requested from us, especially from supermarkets who want to include everything they can.’

Case D (70% MTO): On a product-level, risk of obsolescence and product type are important criteria: ‘Unfortunately most exports are more specialized and are created especially for them.’ Furthermore, the demand patterns and production technology are important in setting the CODP: ‘On our standard

products we need to have stock because they are produced in large quantities and have high demand.’

On a company-level, D attempts to reduce inventory, as they face financial issues and wish to limit holding costs: ‘However, in general it is company policy to operate in a ‘lean manner’ and maintain

minimum stock in order to use our capital more effectively and not tie it down.’ Moreover, this idea is

adopted throughout the supply chain. ‘In general we see a trend … with more frequent deliveries. And

that is because stock is not held by them, and essentially decreases the need for energy costs… as well as less capital tied down.’ Furthermore, storage capacity limits their potential for keeping stock: ‘In general regarding inventory, we try to maintain a balance between raw materials and packaging and the physical space available for storage.’ Additionally, product variety is an important factor, as it

congests the labor intensive packaging stage: ‘where the process becomes more congested…the

packaging stage, which is more labor intensive… due to the fact that we have many different product codes, shapes, sizes and weights and at the same time we produce for many different private labels.’

Case E (10% MTO): Not being influenced by perishability, E decided upon their products’ CODP in a straightforward way: ‘In principle, we put our own products in stock. And actually those small orders

are private label customers. And we do have some of that in stock, the bigger one.’ They do so to cope

with the uncertainties in both demand and supply patterns: ‘You actually have two big problems, on the

(19)

19

that is not that simple either. Because we're dealing with discounts.’ On a company-level, MTS capacity

was increased to maximize utilization of capacity: ‘So that we can produce those two days in a row from

the same product… then you have to convert (set-up) much less. They increased batch sizes even beyond

their own storage capacity: ‘What we've seen in recent years is that we're producing the well-running

products extra… We then save them to [external parties’ warehouse], which costs money.’

Case F (80% MTO): Case F produces ready-made meals for institutional customers, where high perishability greatly influences product-level partitioning: ‘By producing as naturally as possible we

keep very short best before dates. And that determines it largely. If I have 5 days of shelf life, I can not start producing in stock.’ Furthermore, there is a large variety of products: ‘Now we are already on 40,000’… ‘Let's say half of it is own production, then that's 20,000 item numbers that we create ourselves. You can never get all of that into stock.’ Therefore, they require flexibility and low lead times:

‘We look at our customer first. We have a very short lead time, produced today, tomorrow to the

customer and the day after tomorrow on the plate.’ However, products that last longer are stocked: ‘We do look at techniques that enable us to do this... Sous-vide, where we can work with very long best before dates, then we can take the step into stock driven production.’ Risk of obsolescence is therefore the most

important factor determining the CODP. On a company-level, F tries to shift to MTS production through using new cooking-techniques and standardization initiatives: ‘The customer could ask for 1.2 kilos or

2.4 or 5.6 or 3.8. … We have been trying to stop that last year. Now they can choose one pound, a kilo, or two kilos.’ Still, shortage of cooling storage limits their stock building possibilities: ‘If a retailer offers the sauerkraut at a discount then there is a sudden increase in demand. We are not able to deal with this because we do not have the cooling capacity.’ Contrary to storage capacity, they have excess

production capacity: ‘In terms of volumes we are on I think 70-75% of the max. But we are free to stretch

our working hours.’

Case G (5% MTO): Company G operates in a market with extreme low margins: ‘Yes, cheese is about

cents. If cheese is half a penny more expensive, it makes a huge impact on your sales margin.’

Furthermore, products vary in terms of milk type (e.g. bio or meadow), maturity, shape, size, and packaging material, increasing scheduling complexity. Additionally, fluctuating prices requires quick responses, the ability to sell high amounts of cheese quickly can generate significant extra revenue. The rise in value of G’s products makes this a unique case in the sample and explains the high MTS preference for the processing stage. In packaging, capacity distribution is affected by contractual agreements: ‘We have a large pallet warehouse where everything is already customer-specific.’ Additionally, on company-level, constant supply of raw material and the production system shift capacity distribution in favor of MTS production: ‘With us, the raw material is quite leading. It keeps

(20)

20

Case H (70% MTO): At company H, customers all have own quality standards for raw materials: ‘They

all need their own hallmark. For *X* we need the *X* holder. For *Y* we need an *Y* certificate.’

Additionally, risk of obsolescence and product type are important factors influencing the CODP: ‘In the

freezer we give them a year shelf life… [product] is packaged with a certain combination of gases … so it lasts longer. In vacuum it is 13 days.’ ‘Depending on what it is, the margins can increase considerably. The frozen products have a lower margin for example.’ This is also related to holding cost: ‘Yes we try to make as few stock as possible, because stock keeping costs money.’ On a company-level, contractual

agreements with very specific wishes favor MTO production: ‘The weight should be good, there are

many specifications we have to meet for *Y* for instance. We also have the *Y* line. *Y* has yet to be marinated.’ Furthermore, the MTO/MTS decision is influenced by the quantity suppliers deliver: ‘If the farmers want to empty the stables, then [raw material] drops in weight, which is the basis to raise the freezer production … If there is a lot in abundance then the price goes down. If it is very scarce then the price goes up… Then we try to make as much fresh products as possible.’ Interestingly, a change in

capacity distribution has greatly influenced profitability: ‘Then we were producing just to be able to

make more and now we got another production director, that said: ‘we should be producing only what we can sell and not just to create maximum output.’ We decreased our kilograms for processing since that time but our profits have gone up. We are working more effectively.’

Criteria A B C D E F G H %MTO/MTS 80/20 40/60 65/35 70/30 10/90 80/20 5/95 70/30 Produ ct Risk of obsolescence 1 2 1 3 1 5 1 3 Product type 1 3 3 2 4 5 2 3 Variability of raw material 5 2 2 1 1 2 3 3

Cost of each item 2 5 4 4 2 3 5 5

Mar ke t Predictability of demand 5 4 3 4 3 5 4 5 Product customization 2 4 4 4 2 5 2 4 Required delivery reliability 3 5 5 5 5 5 4 5 Required delivery time 1 2 3 2 2 2 1 2 Proces

s Production lead _time 4 1 2 2 4 2 5 2

Process and human flexibility 5 4 5 4 3 5 3 4 Holding Costs 1 3 2 2 5 1 5 2 Su pp li er Supplier commitment 5 2 2 3 5 2 5 3 Pricing policy of retailers 1 3 4 4 4 1 5 4

(21)

21

Table 4.1 provides an overview of the criteria introduced in Table 2.2, and the cases’ scores in relation to other cases in the sample. As mentioned before, scores are allocated through discussing the coded data with a fellow student-researcher. For instance, regarding predictability of demand case B mentions that from experience they learned demand patterns are relatively stable (4): ‘However going back to

experience being an important tool, if we know that client C obtains every week 250 tons, even if they haven’t informed us today, for next week we know that the requested quantity will be more or less similar.’ In C, their main product (in terms of output volume) is affected by seasonality, but still they

can cope quite well, resulting in a Medium (3) score.: ‘However even [product] that has seasonal

demand, because we know when it will be we alter our stock for it during that period and then we produce it again more to order and keep less stock.’ Additional examples are shown in Appendix C.

4.2 Product and company factors effect on partitioning

4.2.1. Product level factors

The most important product-level criteria to influence partitioning are risk of obsolescence, product variety and predictability of demand. High perishability alone determined the MTO preference in case F, whereas companies that were not restricted by perishability more often produced MTS. In general, perishability and holding costs impact risk of obsolescence, increasing the number of customer orders. Furthermore, specific raw materials and/or processing steps increased the production and holding cost, as in case D: ‘for more expensive products for example, I mean with some extra raw materials we try to

have as little stock as possible or make them completely to order.’ Additionally, special storage

requirements can increase holding cost even further, increasing the benefits of MTO as highlighted in case H before, which increased profitability through a more effective strategic focus. Product type therefore was an important factor in determining the CODP and stock levels; (B) ‘as it defines whether

a product can be produced with smaller stock or needs higher stock or without any stock if it is a specialized product.’

On the other hand, producing standardized products increases utilization of capacity as the number of setups is often reduced. Most companies show clear signs of delayed differentiation, where the processing stage is standardized, explicitly mentioned in case C: ‘At this point we normally prepare the

product 1-2 months before they obtain it and we have it as semi-finished product ready to be packed as stock.’ The demand of customized or private labelled products was often combined with long-term

contracts, stabilizing demand patterns in for instance case H: ‘In principle, [customer] is a continuous

order, in the beginning of the quarter it is discussed what we have to deliver. Then you know exactly how to do it and when we need to do it.’ In general, predictability of demand was an important factor

influencing partitioning on a product level. In most cases, unpredictable demand pushed the CODP downwards in order to buffer against uncertainty ‘a product that has variability in demand should have

more stock to account for this uncertainty.’ Contrary, company G uses demand planning in stable

(22)

22

of [product] then we pull them out of the warehouse and we can deliver it. Those are packaged MTS… and that's all forecast driven.’ Here, predictability, rather than unpredictability of demand pushed the

CODP downwards. However, this only relates to products that have an acceptable volume, as E mentions: ‘There's no point in storing 25,000 bottles somewhere and you sell 1,000 in the year.’

4.2.2 Company-level factors

The most important company-level factors influencing partitioning were contractual agreements and limited storage or production capacity. First off, contractual agreements were a large driver of allocating capacity to MTO production. Private labelling drastically increased product variety in the packaging stage, as seen in D: ‘There is then a separate forecast by the sales and marketing department, especially

for the private labelling … In private labels we have contractual agreements as regards to product specifications.’ These contracts are a sign of customer commitment, investing in a relationship with the

customer to create common goals. Similar results were seen across the sample, however G has taken it one step further: ‘That is a closed chain, suppose that there is 100 million liters of milk per week, then

all that milk must be processed in [partner] products. If [partner] has less demand, they still have to buy that [product].’ Through a strategic partnership, a significant amount of capacity could be allocated

to MTO production, reducing risk of obsolescence and holding costs through information sharing. Case G therefore did not only invest in customer-commitment but also in supplier commitment, which is key in creating a reliable MTO system. A constant, reliable supply of raw- and packaging materials allows the use of just-in-time principles. Contrary, E was constrained in their possibilities to produce MTO, as their supply was less reliable: ‘You keep a little more stock yourself, both from the empty goods

(packaging), and of the filled goods (end products), that has everything to do with the weird market. And that you can hardly get bottles.’

Another factor that shifted capacity distribution towards MTO is storage capacity. In some cases, like in A, this was very straightforward: ‘I also don’t have the storage space to store it in. So I am operating

with little to no inventory.’ Not having the physical space to store products can limit MTS capacity.

Similar reasons were given at F: ‘We are not able to deal with this because we do not have the cooling

storage capacity.’ This can be overcome through outsourcing warehousing and logistics, like in E,

however A was financially constrained to invest in stock: ‘We are drained from all sides from our

liquidity and we are lucky to be able to operate and survive in this hostile environment.’ This highlights

the benefits MTO production can offer; freeing up capital through minimizing inventory levels. Whereas high utilization of storage capacity favors MTO production, high utilization of production capacity has led to increasing MTS production. Interestingly, only three cases operated anywhere near capacity limits, company E increased MTS capacity through increasing their batch sizes, satisfying demand without increasing operating hours on a weekly basis: ‘We were actually standing against the

boundary of our capacity for a long time, but that's what's been solved in that way.’ However, for case

(23)

23

sells like crazy and as a result I don’t have barely enough time to produce it before it gets sold.’ For A

this was not an issue, as they basically operated as a production facility for their strategic partner. However, production capacity was sometimes seen as an enabling rather than restricting factor. Company H mentioned one of their production lines was not in use, allowing them to be flexible enough to maintain service levels without buffering final products: ‘We are very flexible in what we can do. If

*Y* sends us a message like guys we will need this next week and this as well… Then we can completely throw our schedule around.’ Companies that distributed more than half their capacity to MTO

production all showed lower levels of utilization, related to the need for flexibility. For instance, F mentions: ‘In terms of volumes we are on I think 70-75% of the max. But we are free to stretch our

working hours.’ Interestingly, F is exploring possibilities to shift more towards MTS production,

whereas most companies try to operate in a ‘lean manner’ and minimize holding cost. However, F tries to increase efficiency though standardizing product weight, exploring new cooking techniques (Sous-vide) and investments in IT tools: ‘In the future we want to work in stock, if we will work with M3 (NEW

ERP system). Then it can be done.’ IT has been mentioned as an enabling factor in the MTO/MTS

decision, as data gathering and analyses tools increase predictability of demand, creating opportunities for MTO production.

Figure 4.1 provides an overview of how company-level factors either constrain or enable a company to allocate capacity to MTO or MTS production. It should be interpreted as follows. The influence of each factor on MTO/MTS capacity distribution is shown by the direction of the arrow, e.g. production capacity can constrain the percentage of total capacity being allocated to MTO production, thus shifting the distribution to the right, in favor of MTS. Furthermore, customer commitment enables MTO production through establishing common goals and data sharing, shifting MTO/MTS capacity towards the left (MTO)

(24)

24

4.3 Grouping product and company-level factors

From the data, it was clear that a lot of the factors, both product- and company-level, are interrelated. However, some factors have a larger impact on the MTO/MTS decision than others. In an attempt to structure these findings, Figure 4.2 provides a hierarchical overview of factors impacting the MTO/MTS decision in the FPI. It should be interpreted as follows. The vertical axis displays factors that influence partitioning on a company-level. First and foremost, capital needs to be present to invest in stock and there should be room to store final products, otherwise increasing MTO capacity is recommended. Third, supply of necessary goods should be stable enough, and lastly the production lead time should be short enough (considering utilization) otherwise capacity distribution should favor MTS. These factors should always be taken into account before determining the product-level production strategy. The horizontal axis shows the most important (decreasing from left to right) factors that determine the CODP preference for a single product or product family. Risk of obsolescence is the most important constraining factor in producing MTS, if it is too high, production will have to follow a MTO strategy. If it is deemed acceptable from the decision-makers’ perspective, the second factor is taken into account. Each factor is subsequently assessed until one is believed to be critical enough to determine the CODP. Hybrid products are produced MTS up to an intermediate storage point, after which they are finalized as MTO. It must be noted there are still underlying factors impacting the company- and product-level decision criteria, however the factors presented here were most dominant in the final MTO/MTS decision.

(25)

25

5. Discussion

The discussion looks at how the FPI characteristics influence different product and company level factors, and how these factors influence the MTO/MTS decision.

5.1 Customer order decoupling point analyses

It was made clear that FPI specific characteristics influence the partitioning decision, thus it should be researched as a separate entity. In general, risk of obsolescence is an important factor determining the CODP (Hemmati & Rabbani, 2010) and perishability, a common factor in the FPI, increases the risk of obsolescence (Van Donk, 2001; Hemmati & Rabbani, 2010; Rafiei & Rabbani, 2014). As a consequence, customer orders were more frequent but smaller, which has an upstream effect on the CODP (Zaerpour et al., 2009). In some cases, extreme perishability even results in pure MTO production. However, little attention is given to products for which value increases over time. For instance alcoholic beverages or cheese become more valuable if it is stored longer, favoring MTS production. It can be argued products are still work-in-progress, however its downward effect on the CODP has been neglected in literature. Maintaining stock increases holding cost, (Hemmati & Rabbani, 2010; Zaerpour et al., 2008) however, in an industry characterized by capital-intensive production technology (Van Donk, 2001), adding value without having to do anything is a welcome change. Still, in most cases product type and associated holding cost are important factors to produce MTO, it determines inventory levels, where more specialized products have fewer stock, or no stock at all (MTO). As such, empirical evidence is provided to support the CODP theory on perishability and risk of obsolescence, and product type and holding costs.

However, findings on the influence of demand patterns are contradicting. Some findings resemble for instance Van Donk (2001) and Hemmati and Rabbani (2010), which state that in case of unpredictable demand a MTO strategy is preferable. However, most cases seem to do the opposite, building up stock of final products to buffer against uncertainty. This has been related to service levels, not wanting to say no to any customer. Furthermore, in these cases production lead-time was too high, thus MTO was not acceptable, showing the interconnectedness of factors. In line with Van Donk et al. (2005), service requirements were often considered before cost. In today’s competitive market this seems like a must, a company constantly has to meet (and exceed) downward supply chain service requirements only after which they can focus on cost reductions (Soman et al., 2004).

5.2 The MTO/MTS decision in the FPI

Based on the findings, a distinction is necessary between company-level factors that directly or indirectly influence partitioning. For instance, limited storage capacity has a direct effect on a companies’ possibility to keep stock. On the other hand, outsourced storage (E, H) cooled storage (D,

(26)

26

therefore indirectly impact partitioning, as increased holding costs push the CODP upwards (Zaerpour et al., 2008; Hemmati & Rabbani, 2010).

Whereas storage capacity can constrain MTS production, production capacity can constrain MTO production. High utilization on bottleneck processes often increases batch sizes in order to maximize utilization, increasing average stock levels. Efficient planning and scheduling of production is important in the FPI due to capital intensive production systems (Van Donk, 2001; Kilic et al., 2010). Contrary, companies with a capacity distribution favoring MTO production mentioned having spare capacity on many occasions. For some cases, the system was designed with excess capacity to be flexible enough to always fill demand, whereas others increased MTO production in an attempt to become more profitable. Decreased demand or increased profitability for MTO products explain the shift from mainly MTS to mainly MTO production: (H) ‘We decreased our kilograms for processing since but our profits have

gone up.’ This highlights the importance of choosing the right production strategy (Zaerpour et al., 2008;

Garn & Aitken, 2015)

Another important emerging factor was the use of contractual agreements. First, customer commitment can increase predictability of demand through data sharing, exchanging relevant point-of-sale data can reduce the need to keep stock in the rather unpredictable FPI market (Van Donk, 2001). As such, downwards supply chain integrations allows MTO production (Zaerpour et al., 2008, p. 191); Having

long-term relationship with customers and helping customers define their goals and needs are reasons of applying MTO system.’ Second, the divergent product structure present in FPI organizations (Kilic et

(27)

27

6. Conclusion

This research looked at how product- and company-level factors influence the MTO/MTS decision in hybrid production systems. Through the analysis of eight companies operating in the Dutch and Greek food-processing sector, this research investigates how capacity constraints and other company-level criteria influence partitioning. It’s contributions are fourfold: (1) literature has been focusing on quantitatively solving the partitioning problem whereas this research adopted a qualitative method, (2) this led to new findings regarding the influence of company-level partitioning criteria, and (3) empirical evidence to support previous findings regarding product-level partitioning criteria. As a final result, this led to (4) the development of a simplified hierarchical decision framework guiding MTO/MTS partitioning that merges product- and company-level criteria.

6.1 Theoretical implications

Through qualitatively analyzing the planning and scheduling procedures in the FPI this study adds to the knowledge base of decision-making processes and scheduling characteristics of hybrid production systems (Soman et al., 2004). Different reasons for similar capacity distributions were given, however similar arguments were provided to shift from MTO to MTS or back. Therefore, this paper developed a framework to better understand how company-level factors influence partitioning in the FPI, building upon research by Van Donk et al. (2005). A distinction is made between factors that directly, or indirectly influence partitioning. On one hand, customer commitment indirectly affects partitioning, as it stabilizes demand patterns and reduces inventory levels. Following the ideas presented in Akkerman et al. (2010), delayed differentiation can be achieved through contractual agreements, which indirectly influence product variety through private labelling. On the other hand, limited storage capacity or financial resources directly influence the possibility to maintain stock. Additionally, a stable supply pattern is required to implement MTO principles. Future research should investigate how organizations try to use these company-level factors to create an environment in which they can optimally choose between MTO and MTS. As such, company factors that influence partitioning on a higher level are identified, that should be considered before setting the CODP on a product level.

(28)

28

6.2 Managerial implications

Managers of hybrid food processing organizations can use these findings to structurally look at how they are constrained in dealing with the partitioning problem. It can steer investment decision based on the theory of constraints; organizations in unstable environments should invest in long-term contracts with their customers so they can improve supply chain performance. Creating common goals and sharing data with customers is a large enabler of reducing holding costs (Zaerpour et al., 2009). Additionally, contractual agreements are a valuable option for organizations with a bottleneck in their processing stage, as postponement allows increasing efficiency of upstream activities (Van Donk, 2001; Rafiei & Rabbani, 2012). As such, an effort should be made to achieve delayed differentiation, allowing maximizing utilization of capital-intensive processing technologies (Akkerman et al., 2010). In a more direct fashion, managers should consider the balance between production and storage capacity. Capacity constraints in the first favor an increase in MTS production, as it can generate additional output. Whereas constraints on the latter require a shift to MTO production as outsourcing storage (especially cooled storage) can be very costly.

(29)

29

7. Limitations and directions for future research

Some directions for future research have been given in the previous sections, for instance the development of more logic-based decision-making models, that explicitly consider company-level factors. Additionally, this section reflects on the research methods and outcomes to indicate future research possibilities.

Looking at data collection, it can not be guaranteed that a different sample would have generated the same results. Even though an effort was made to select cases from different sectors, varying in size and capacity distribution, different cases might have provided other reasons for their MTO/MTS decision. However, from both the two largest and two smallest companies, one produced primarily MTO and one primarily MTS. Interestingly, the cases that distributed capacity more evenly between MTO and MTS were medium-sized companies, therefore it might be interesting to investigate how size impacts MTO/MTS capacity distribution. Additionally, even though all interviewees were selected based on their relevance to the partitioning process, differences in strategic planning experience has led to different levels of detailed discussions during the interviews. To ensure richness of data, additional questions were asked during the tour, or during follow-up questions in e-mails.

Furthermore, this research has looked at multiple sectors within the FPI, and different results were found between, but also within sectors. It would be interesting to investigate sectors in more detail, incorporating for instance only alcoholic beverage producers into the sample. Subsequently investigating multiple sectors in more detail allows an ex-post analysis of the influences of factors in different sectors to check the generalizability of the framework. Furthermore, the research was of explorative nature, meaning that the factors and relations that emerged from this study require further investigation to understand their explanatory value. Interesting research directions are for instance differences in size or culture, and how this impacts the partitioning process.

(30)

30

8. References

Akkerman, R., Van Der Meer, D., & Van Donk, D. P. (2010). Make to stock and mix to order: Choosing intermediate products in the food-processing industry. International Journal of

Production Research, 48(12), 3475–3492.

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.

Barton, R. R. (2013). Proceedings of the 2013 Winter Simulation Conference R. Pasupathy, S.-H. Kim,

A. Tolk, R. Hill, and M. E. Kuhl, eds. (2010), 342–353.

Beemsterboer, B., Land, M., & Teunter, R. (2016). Hybrid MTO-MTS production planning: An explorative study. European Journal of Operational Research, 248(2), 453–461.

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. Campbell, J. L., Quincy, C., Osserman, J., & Pedersen, O. K. (2013). Coding In-depth Semistructured

Interviews: Problems of Unitization and Intercoder Reliability and Agreement. Sociological

Methods and Research, 42(3), 294–320.

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 andproduct decay. Computers and Operations Research, 76, 147–154.

Federgruen, A., & Katalan, Z. (1999). Impact of adding a make-to-order item to a make-to-stock production system. Management Science, 45(7), 980–994. https://doi.org/10.1287/mnsc.45.7.980 Garn, W., & Aitken, J. (2015). Splitting hybrid Make-To-Order and Make-To-Stock demand profiles.

Retrieved from http://arxiv.org/abs/1504.03594

Hoekstra, S. and Romme, J. (1992). Integral logistic structures: developing customer-oriented goods

flow. London: McGraw-Hill

.

Hemmati, S., & Rabbani, M. (2010). Make-to-order/make-to-stock partitioning decision using the analytic network process. International Journal of Advanced Manufacturing Technology, 48(5– 8), 801–813.

Iravani, S. M. R., Liu, T., & Simchi-Levi, D. (2012). Optimal production and admission policies in make-to-stock/make-to-order manufacturing systems. Production and Operations Management,

21(2), 224–235.

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.

Karlsson, C. (Ed.). (2016). Research Methods for Operations Management. 2nd edition. New York:Routledge.

(31)

31

McCutcheon, D. M., & Meredith, J. R. (1993). Conducting case study research in operations management. Journal of Operations Management, 11(3), 239–256.

O’Reilly, S., Kumar, A., & Adam, F. (2015). The role of hierarchical production planning in food manufacturing SMEs. International Journal of Operations and Production Management, 35(10), 1362–1385.

Olhager, J. (2003). Strategic positioning of the order penetration point. International Journal of

Production Economics, 85(3), 319–329.

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 and Control,

20(7), 559–575.

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. (2011). Order partitioning and Order Penetration Point location in hybrid

Make-To-Stock/Make-To-Order production contexts. Computers and Industrial Engineering,

61(3), 550–560.

Rafiei, H., & Rabbani, M. (2014). Hybrid MTS/MTO order partitioning framework based upon fuzzy analytic network process. Applied Soft Computing Journal, 19, 312–321.

Soman, C. A., Van Donk, D. P., & Gaalman, G. (2004). Combined make-to-order and make-to-stock in a food production system SOM-theme A: Primary processes within firms. Int. J. Production

Economics, 90, 223–235.

van Donk, D. P. (2001). MTS or MTO : The DP in the food processing industry. International Journal

of Production Economics, 69(3), 297–306.

van Donk, D. P., Soman, C., & Gaalman, G. (2005). A decision aid for order and make-to-stock classification in food processing industries. Proceedings EurOM International Conference

on Operations and Global Competitiveness.

Voss, C., Tsikriktsis, N., & Frohlich, M. (2002). Case research in operations management.

International Journal of Operations & Production Management, 22(2), 195–219.

Williams, T. M. (1984). Special products and uncertainty in production/inventory systems. European

Journal of Operational Research, 15(1), 46–54.

Yilmaz, K. (2013). Comparison of quantitative and qualitative research traditions: Epistemological, theoretical, and methodological differences. European Journal of Education, 48(2), 311–325. Yousefnejad, H., Rabbani, M., & Manavizadeh, N. (2019). A Simulation-Optimization Model For

Capacity Coordination In Make To Stock/Make To Order Production Environments. Iranian

Journal Of Management Studies, 12(2), 61–79.

(32)

32

Zaerpour, N., Rabbani, M., Gharehgozli, A. H., & Tavakkoli-Moghaddam, R. (2009). A

comprehensive decision making structure for partitioning of make-to-order, make-to-stock and hybrid products. Soft Computing, 13(11), 1035–1054.

Understanding the influence of company-level factors affecting the MTO/MTS decision in hybrid food processing organizations.