MANAGING SUPPLY CHAIN COMPLEXITY: A MULTIPLE CASE STUDY IN THE FOOD PROCESSING INDUSTRY

(1)

MANAGING SUPPLY CHAIN COMPLEXITY: A

MULTIPLE CASE STUDY IN THE FOOD PROCESSING

INDUSTRY

Master thesis, MSc Supply Chain Management

University of Groningen, Faculty of Economics and Business

(2)

ABSTRACT

Because of the negative effects of supply chain complexity, identifying strategies to manage it is crucial. The inherent characteristics of food processing industry make the supply chain more complex compare to other contexts. Thus, it is urgent to identify strategies to manage supply chain complexity in food processing industry. By means of multiple case studies, this study identifies strategies to manage supply chain complexity in food processing industry and influences of food processing industry characteristics on the applications of management strategies. As a result, ten management strategies are identified and the influences are further explained. This provides opportunities for companies in food processing industry to evaluate their level of characteristics and select suitable strategies to manage supply chain complexity.

(3)

1. INTRODUCTION

Despite many scholars studied strategies (e.g. keeping inventory and using information systems) to manage supply chain (SC) complexity in various contexts, none of these studies focus on food processing industry (FPI). The inherent characteristics of FPI in terms of process and product (e.g. perishable nature and divergent product structure) make the SC more complex compared to other context and influence the applications of management strategies. Therefore, it is urgent to study what strategies are applied by companies in FPI to manage SC complexity and how does the FPI characteristics influence the application of these strategies.

SC complexity causes long lead-times, difficulties in supply chains integration, lower supply chain performance and inflexibility (De Leeuw, Grotenhuis, & van Goor, 2013; Bozarth, Warsing, Flynn, & Flynn, 2009; Blecker, Kersten, & Meyer, 2005; Koudal & Engel, 2007; Vachon & Klassen, 2002). These negative impacts make it essential to cope with SC

complexity. Perona and Miragliotta (2004) pointed out that companies which focus on dealing with their SC complexity are more likely to acquire superior results in terms of efficiency (cost) and effectiveness (service).

Because of the perishable nature, high demand fluctuation and the increasing food safety concerns, the SC management of FPI is more complex compare to other industries (Shukla & Jharkharia, 2013). Although literatures identified some strategies to manage SC complexity (Manuj & Sahin, 2011; De Leeuw et al., 2013; Perona & Miragliotta, 2004), it is indicated that the perishable nature hinders the use of keeping inventory (Norhayati, Rasma, & Mohd, 2013), which is frequently used in other context to manage SC complexity. Therefore, there is no guarantee that the strategies applied in other context also work well in FPI. Where none of the previous researches studied the strategies in FPI, the novelty of this research is that it identifies the strategies used by companies in the PFI to manage SC complexity and studies the impacts of FPI characteristics on the application of management strategies. By doing so, a guideline is provided for the companies in the FPI to manage their complexity. Hence, the research aims to fill the aforementioned gap by answering the following research questions:

Research question 1: What management strategies are used to manage dynamic complexity of supply chain in food processing industry?

(5)

application of management strategies?

To answer these questions, a multiple case study was conducted in FPI. In contrast to survey study, multiple case studies provide rich data that help to explore former theories and get more insights. 11 Companies in FPI were selected and semi-structured interviews were designed to conduct the case studies. This study filled in the gap in theory by identifying the strategies to manage SC complexity in the FPI and the influences of FPI characteristics on the application of management strategies. From a practical aspect, companies in the FPI can evaluate their level of FPI characteristics and select the suitable strategies to manage their SC complexity based on the findings of influences.

The reminder of the paper is organized as follow: Section 2 provides a review of the relevant literatures followed by a discussion of the research methodology in Section 3 which includes the process of case selection, data collection and data analysis. Section 4 presents the results followed by a discussion of the main findings in Section 5. Finally, Section 6 discusses research limitations, contributions and future research directions.

2. THEORETICAL BACKGROUND

In this section the theoretical foundation of this study will be discussed. First, there is a short review about existing theories on SC complexity and the complexity driver. Second, past studies on strategies to manage dynamic complexity of SC are reviewed. Third, perspectives are extended to the context of FPI. Finally, a research framework is built based on the theories discussed.

2.1 Supply chain complexity

Upstream and downstream flows of products and service, accompanied by related finances and information constitute SC (Beamon, 1998; Lambert, Cooper, & Pagh, 1998). A variety of definitions of complexity existed in past literatures. Arteta and Giachetti (2004: 497) stated “complexity arises from not only the size of the system but also the interrelationships of the system components and the emergent behavior that cannot be predicted from the individual system components.” In accordance with Wood (1986) and Campbell’s (1988) task

(6)

unpredictable environment of SC (Manuj & Sahin, 2011). This critical perspective offers basis for the following complexity classification.

SC complexity is distinguished into two types in terms of different forms and origins. Static complexity represents the variety of structure in the static system, which includes the number and interaction between the sub-systems in SC, and dynamic complexity indicates uncertainty of the operational behavior of the system, which includes the aspects of time and randomness (Sivadasan, Efstathiou, Frizelle, Shirazi, & Calinescu, 2002; Serdarasan, 2013). To deal with SC complexity, three generic approaches are observed: complexity reduction, complexity management, and complexity prevention (Hoole, 2005; Perona, & Miradliotta, 2004;

Serdarasan, 2013). Serdarasan (2013) further stated that company tends to reduce complexity when facing static complexity and manage complexity when dealing with dynamic

complexity. As mentioned above, the strategies to manage dynamic complexity is the main focus of this study.

Serdarasan (2013) mentioned that management strategies should be developed by analyzing the complexity drivers in advance. Previous researches identified a number of complexity drivers (Bozarth, Warsing, Flynn, & Flynn, 2009; De Leeuw et al., 2013; Sivadasan et al., 2002; Perona & Miragliotta, 2004; Isik, 2010). As indicated by Serdarasan (2013), the distinctions of static and dynamic complexity are also valid when classifying SC complexity drivers. Based on the way they are generated, this paper uses uncertainty, variety and speed (see table 1) as dynamic complexity drivers. The uncertainty of SC related to the accuracy of demand forecast and the reliability of performance from both suppliers and customers (De Leeuw et al., 2013). One kind of action leads to different results is one of the outcomes of uncertainty, which results in complexity (Bozarth et al., 2009; De Leeuw et al., 2013). Variability refers to alteration in resources and requirements over time (De Leeuw et al., 2013). If the elements of system change rapidly, the system complexity will increase (Isik, 2010). Speed indicated the required responsiveness across the SC (De Leeuw et al., 2013). The uncertainty and variability consider the unpredictable operational behavior of the SC and the speed takes the aspects of time and randomness into consideration that are in line with the generation of dynamic complexity.

Dynamic complexity driver Definition

Uncertainty The lack of predictability and reliability of demand and of supply chain in process.

(7)

Speed

requirements, and sources in the course of time.

Required responsiveness across the supply chain in terms of throughput times, delivery times and frequencies.

Table 1: Dynamic complexity drivers (Isik, 2010: 3684; De Leeuw et al., 2013: 967)

2.2 Strategies to manage dynamic complexity of supply chain

Existing studies have already identified some strategies to manage dynamic complexity of SC (De Leeuw et al., 2013; Manuj & Sahin, 2011; Perona & Miragliotta, 2004; Serdarasan, 2013). By comparing and combining the strategies in those studies, the most cited strategies are showed in table 2.

Strategies Author (year) Communication and information

exchange

De Leeuw et al. (2013); Manuj and Sahin (2011); Perona and Miragliotta (2004); Serdarasan (2013)

Workforce flexibility Manuj and Sahin (2011); De Leeuw et al. (2013)

Keeping inventory Manuj and Sahin (2011); De Leeuw et al. (2013)

Product modularization De Leeuw et al. (2013); Perona and Miragliotta (2004)

Relationship management Manuj and Sahin (2011); Perona and Miragliotta (2004); Serdarasan (2013) Table 2: Strategies identified in previous literatures

Communication and information exchange contribute to manage the uncertainty of SC by improving the coordination between the SC partners (Costantino, Di Gravio, Shaban, & Tronci, 2015). It is also identified as one of the major means to mitigate bullwhip effect (Skjøtt-Larsen, Mikkola, & Kotzab, 2007). This approach can be applied by using

information systems which help a team to make better, quicker decision and hence improve competitive advantage (Manuj & Sahin, 2011). To be more specific, Perona and Miragliotta (2004) stated that investment in information systems for production planning and control (PP&C) can both diminish the number of workforce required in production planning and enhance production readiness.

(8)

workforce can be achieved by employing more people during the busy period and using temporary employees. From another aspect, Manuj and Sahin (2011) indicated several managers trained employees to have a better understanding of the initiatives of their jobs and to have good multi-tasking skills to gain flexible workforce.

Keeping inventory is mentioned as a major coping mechanism because it can reduce the influences of uncertainty (De Leeuw et al., 2013). Manuj and Sahin (2011) further added that when customers release an unexpected promotion without prior notice, companies need to rely on inventories. However, since keeping inventory enhances costs and requires suitable capabilities to handle inventory, a trade-off need to be made (De Leeuw et al. ,2013).

Lau and Yam (2005) stated that designing a product as a set of sub-assemblies can simplify the complexity. Product modularization is the method to develop sub-assemblies and integrates them in a maximum number of ways (Lau & Yam, 2005). In FPI, product

modulation can be conducted by blending a limited number of intermediate products into end products (Kilic, Akkerman, van Donk, & Grunow, 2013). Product modularization contributes to shorten the lead time (Lau & Yam, 2005) and increase the efficiency performance (Perona, & Miragliotta, 2004).

Perona and Miragliotta (2004) described keeping stable relationships with both suppliers and customers as a win-win solution because it can add value to activities. Manuj and Sahin (2011) further mentioned that it is important to identify the importance of each relationship and build long-term relationships with key partners.

2.3 Food processing industry

Van Donk (2001) distinguished two kinds of companies in FPI. The first category is companies which process raw material to intermediate products and the second category is companies which further produce customer products from these intermediate products (Van Donk, 2001). The cases selected contain the companies of both categories. An overview of the steps involved is showed in figure 1.

(9)

The SC of companies within FPI is more complex compared to other SC because the raw materials and end products are perishable, the demand and price are highly fluctuated and customers’ concerns for food safety are increased (Shukla & Jharkharia, 2013; Jongen, & Meulenberg, 1998; Meulenberg, & Viaene, 1998). Based on a literature review, Van Donk (2001) classified the characteristics of FPI into three categories, which are plant

characteristics, product characteristics and production process characteristics, and enumerated the detailed characteristics for each category (see table 3) (van Dam, 1995; Fransoo & Rutten, 1994; van Dam, Gaalman, & Sierksma, 1993).

Plant characteristics (a) Expensive and single-purpose capacity coupled with small product variety and high volumes. Usually, the factory shows a flow shop oriented design.

(b) There are long (sequence-dependent) set-up times between different product types.

Product characteristics (a) The nature and source of raw material in food processing industry often implies a variable supply, quality, and price due to unstable yield of farmers.

(b) In contrast with discrete manufacturing, volume or weights are used.

(c) Raw material, semi-manufactured products, and end products are perishable.

Production process characteristics

(a) Processes have a variable yield and processing time.

(b) At least one of the processes deals with homogeneous products. (c) The processing stages are not labor intensive.

(d) Production rate is mainly determined by capacity.

(e) Food industries have a divergent product structure, especially in the packaging stage.

(f) Factories that produce consumer goods can have an extensive, labor-intensive packaging phase.

(g) Due to uncertainty in pricing, quality, and supply of raw material, several recipes are available for a product.

Table 3: Characteristics of food processing industry (Van Donk, 2001: 300)

2.4 Conceptual model

(10)

strategies (Communication and information exchange, Workforce flexibility, Keeping inventory, Product modularization and Relationship management) that are identified by literatures to manage dynamic complexity of SC. Section 2.3 indicated the characteristics of FPI. Based on the concepts discussed above, the following line of reasoning provides the bases for the relations described in the research framework (figure 2).

Figure 2: Conceptual model

Based on the suggestion of Serdarasan (2013) that analyzing complexity drivers contributes to identify the management strategies, the five management strategies are linked with

complexity drivers in section 2. To conclude, communication and keeping inventory are used to manage the uncertainty. Relation management and workforce flexibility cope with

fluctuation of elements in SC, hence manage the variability. Product modularization contributes to satisfy the required responsiveness to manage the speed. Besides, additional strategies are expected to emerge from the case studies.

Although previous studies rarely studied the influences of industry characteristics on strategies application, some lines of reasoning could still be found. Van Wezel et al. (2006) stated that some typical characteristics of processing industry (batch processes, flow

production, sequence dependent clean time) make planning requirement oppose to the market requirement, which makes it hard to use the ERP system or advanced planning systems to increase the flexibility of planning practices. In addition, raw materials, semi-structured products and end products in FPI may become unfit for consumption after a certain time because of the perishable nature (Nahmias, 1982). If companies want to keep inventory to manage dynamic complexity, complicated inventory policy model is required to cope with perishability, which is time-consuming and costly (Nahmias, 1982). Thus, it is expected that several characteristics have influences on the application of strategies.

(11)

3. METHODOLOGY

The aim of this research is to identify the strategies to manage dynamic complexity of SC and gain insights into the influences of the FPI characteristics on the application of management strategies. A case study approach was taken for this research. Case studies are recommended for exploratory research (Voss, Tsikriktsis, & Frohlich, 2002). One of the significant

advantages of case study is rich data, which can provide strong base for theory explanation compare to survey approach (Karlsson, 2009). The second reason to choose case study approach is that it is a suitable approach to answer exploratory question (Karlsson, 2009). By conducting multiple case studies, observer bias can be minimized and the generalizability of the conclusion can be improved (Voss et al., 2002). In addition, it is especially crucial to pay attention to reliability and validity when doing case study (Karlsson, 2009). Based on the suggestion of Yin (1994), several tactics were used to increase the reliability and validity of this research. These tactics were further explained in the following sections.

3.1 Sample selection

The unit of analysis is the plants within FPI. Both literal logic and theoretical logic were used to guide the case selection and data analysis. It is literal logic because we are expecting similar findings within the groups of plants, which experience same level of FPI

characteristics, for the application of strategies. It also is theoretical logic because we are expecting to find different results between two groups of plants that experience different level of FPI characteristics. Therefore, some of the cases selected are anticipated to experience the same level of FPI characteristics, compared to the rest, to form a group. The use of replication logic promotes the external validity. Based on the information provided on the website, a list of companies in FPI was selected as the initial options. A telephone call or e-mail was

conducted to ask the company to participate the research. Those companies that are interested to join were further informed with some additional information through e-mail. 11 companies were finally selected to conduct the interview.

(12)

are more based on customers than suppliers and upstream indicates they are based more on suppliers. Eight of the eleven cases are more focus on customers. Because of the diversity in market and due to the comparable parameters mentioned above, it is expected that the participating 11 companies experience different level of FPI characteristics and use different strategies to manage dynamic complexity of SC. To gather all the information required, one or more interviewees were selected from each company. These interviewees together are required to cover all the knowledge about the company in terms of operation, marketing, logistics and production.

Table 4: Research samples

3.2 Development interview protocol

As mentioned above, a semi-structured interview is the foundation of this research. The interview questions (see appendix A) contain three parts: the first part is some general questions about the companies and the interviewees. These questions gave insights into the main products and process characteristics, information and geographical dispersion about the main suppliers and customers, and the annually turnover in terms of volume and value. This part helps interviewers have a better understanding of the samples.

(13)

The third part asks questions about how the companies experience the FPI characteristics and what is the influence on SC complexity. The questions about the PFI characteristics are based on the study of Van Donk (2001) in terms of plant, product and production process

characteristics. Combining the questions from the second and third part, we can analyse the influence of FPI characteristics on the applications of management strategies.

In addition, a questionnaire (see appendix B) is designed to measure the level of FPI characteristics of each sample. The questionnaire contains 10 dimensions that are adapted from the study of Van Donk (2001). The results of each dimension must be placed on a scale from one to five, where one stands for low level and five for high level. The questionnaire provides quantitative data and backs up the results of interview.

3.3 Data collection

As mentioned above, semi-structured interviews were conducted to collect data. There are five researchers to conduct the interview in total. One or two researchers visited 9 of the 11 companies selected. Due to distance consideration and time limitation, the interviews with two Chinese companies were conducted through Skype. Each interview takes ninety minutes on average across all of the visited companies. Two interviews were first conducted with company E and F to check whether questions from the interview protocol are comprehensive and understandable. The interview protocol was modified according to the feedbacks from the first two interviews. The revised version was used for the rest interviews. By doing so,

reliability of this research can be increased. During the interview, one interviewer asked the questions and the other interviewer took notes and recorded. As Karlsson (2009: 177) pointed out that the use of multiple investigators can “enhance the creative potential of the teams and convergence of observations increases confidence in the findings.” Based on the response of the interviewees, follow-up questions would be asked. To get complete data and to accurately summarize the data, tape record was used with the permission of the interviewees.

In addition, the questionnaire was sent to the interviewees beforehand by e-mail to let them prepare in advance. After the interview, questionnaire was sent back from the interviewees.

3.4 Data documentation and Coding

3.4.1 Documentation

(14)

involved in this phase. Besides, other sources of evidences such as information collected from the Internet and observed by the researchers were also taken into consideration in

documentation. As mentioned by Yin (1994), multiple sources of evidence contribute to the construct validity.

3.4.2 Data reduction and coding

Coding and data reduction methods were utilized to transform the collected data into meaningful and comparable information. The whole data were first split into three sections: the general information of the companies, the qualitative data derived from the open questions of the interview and the quantitative data got from the questionnaire. The data from the first part is summarized in table 4.

The qualitative data were further coded into two parts. The first part contains strategies each case used to manage dynamic complexity in terms of each complexity drivers (see table 5). To gain insight on the application of strategies, the way each case applies specific strategies were also coded in detail (see appendix D). The second part included the data about how each case experience FPI characteristics (see appendix E). By combining the data of 11 cases gathered by 5 researchers, any lacking or faulty data could be identified. Because company J and K lack important information, these two cases were excluded. Therefore, the finding and discussion are based on the rest nine cases.

As mentioned above, quantitative data about FPI characteristics are acquired from the likely scale questionnaire. Because the questionnaire was answered by two interviewees for some cases and by one interviewee for the other cases, the level of characteristics, each case experienced, were calculated as average for those cases with two interviewees (see appendix C).

3.5 Data analysis

To answer research question 1, strategies, which are used to manage dynamic complexity, are concluded in terms of each dynamic complexity driver for each case (see table 5). The coding scheme (see appendix D) is the main input of this process and the qualitative data contribute to give insights on the application of strategies.

(15)

applications of management strategies are obtained by comparing the use of strategies within each group and between each group. It is expected that a higher level of certain characteristics will lead to a higher or lower possibility in applying several strategies. In this process, the coding scheme and qualitative data about the applications of strategies and FPI characteristics are the main input that provide the clues to identify the impacts.

4. RESULTS

This section summarizes the main results of the research, starting with the management strategies, each case used, in terms of each dynamic complexity drivers. Afterwards, the FPI characteristics of each case are shown. At the end, the influences of FPI characteristics on the applications of management strategies are concluded.

4.1 Management strategies

(16)

To answer research question 1, table 5 shows the management strategies each case used to manage dynamic complexity of SC in terms of each dynamic complexity driver. Except case D and I did not mention any strategy to manage required responsiveness within the SC, all the rest cases used one or more strategies to manage each dynamic complexity driver. It is

interesting to notice that some of the strategies are used to manage more than one complexity derivers.

Appendix D gives more insights on how each case applies these strategies. When looking at Appendix D, it appears that the five management strategies that mentioned in section 2.2 are all applied in different levels by these nine cases. In addition, five additional strategies are used by these cases to manage dynamic complexity of SC in the FPI.

(17)

“we seldom use single sourcing. If we do that, we study it beforehand to conclude whether it is wise. And if there is a disaster at the supplier, we are able to move to another supplier.” Therefore, multiple sourcing provides more options in case of supply chain failure to cope with uncertainty. The fifth one is market segmentation. Both case A and F use distinct production and inventory policies for different market segments to manage the uncertainty of SC.

Four strategies contribute to manage the variability of SC within FPI. The first one is multiple sourcing. The head of operations of case A said “It gives us the chances to benefit from their competitive environment as well as the security to have back-ups in case of unexpected events.” The second one is continuous quality control. All the cases mentioned that they continuously measured the quality from the raw materials to the end products. By rejecting the disqualified materials, the quality of the products could be relatively stable. In addition, the quality of the supplier is also evaluated. The sales manager of case B indicated that before they cooperate with a specific supplier, they measured the performance of all the suppliers in the market and selected the most suitable one. After the cooperation, they measure the performance of the supplier once a year and further decide whether to keep a long-term relationship with this supplier. The third strategy is relationship management. Keeping a good relationship with suppliers enables a stable raw material supply in terms of price, quality and delivery. One way to keep a good relationship with suppliers is to provide bonus for the good performance. As the quality manager of case I said, “If you do good, we provide you a little premium. You will be a little better with us than with our competitor. So, this stimulates farmers to do a good job.” Another way is to have long-term contract. The logistics manager of case G, whose main product is chicken, mentioned that “By using long term contracts, we can influence the chain, by together determining how the chicken should be look like before slaughtering, the specifications, the healthy status, what kind of races, etc.” The forth one is workforce flexibility. This strategy is frequently used by the companies within FPI because most company in the FPI experience a seasonal demand. This strategy also can be implied in several ways. Case B cooperates with third party labor companies during the peak demand period. In a similar way, case I collaborate with many school that focused on food processing. These schools provide students to work for the company during the peak demand period and the company offers the working opportunities. The quality manager of case I called this a “win-win strategy”. In addition, case A uses the workforce flexibility by purchasing the production capacity from competitors.

(18)

increases the delivery performance. For example, case B, which produces milk powder, divides their market into two part: retail customer, which contains the national key account, and professional baby store. Direct supply model is used for the retail customer to ensure a high service level and distribution model is used for baby store to save cost. The second strategy is product modularization. The application of product modularization in the FPI can be divided in two ways. The first way is producing semi-products in advance and integrate them as different end products based on the demand forecast. The second way is using different types of package to distinguish the end products. Both ways can reduce the reaction times within the SC. The third strategy is to outsource delivery activities. This strategy is applied by cooperating with a third party logistics provider (3PL). Outsourcing delivery activities shifts the fixed costs to variable costs. As the head of purchasing of case H said “we hired a 3PL in order to ensure a smooth process. They have a large automated warehouse where we keep a lot of our raw materials in stock not far away from our production plant.”

4.2 Food processing industry characteristics

The outcomes of the likely scale questionnaires (see appendix C) result in a difference, in terms of each FPI characteristics, for each case. By combining the quantitative data from questionnaire (see appendix C) and the qualitative data from interviews (see appendix E), the FPI characteristics are concluded in table 6 for each case in three levels, low, medium and high.

Table 6: FPI characteristics of each case

When looking at table 6, some interesting commonness can be concluded. First, it is

interesting to notice that all the cases experience a high level of food safety regulations. Cases mentioned that food safety regulations influence the use of preservatives, the labeling of products and the use of pesticides. In addition, regulations vary from country to country and are updated frequently.

(19)

single sourcing causes high risks of raw material supply and the supply chain coordinator of case E stated the variability of raw material quality is high because it depends on the harvest.

Third, case A, B and H experience a low risk of obsolescence in terms of raw materials, semi-manufactured products and ‘end’ products. This is because the shelf life of their products is relatively long and case A mixes their products with preservatives to increase the shelf life.

Fourth, except case A and C, all the cases experience a relatively high level of variability of production in terms of yield and processing time. Because the production is based on demand forecast, the high variability of yield is caused by high fluctuation of customer demand. As can be observed in the coding scheme (see appendix E), almost all the cases indicated that they experience different levels of seasonal demands.

Fifth, the level of divergent product structure of case A, B and C is the highest. This is caused by a large number of final products and different package styles. As mentioned by the head of operation of case C, “In order to retain competitive, we had to enlarge our assortments.”

Sixth, case C and E have a relatively low need of producing at maximum capacity. Both cases indicated that the production is foreseeable and controllable. In contrast, case A, B and H expressed that they need to fully used the production capacity, which leads to a high need of production at maximum capacity.

Table 7: Case grouping

(20)

However, their qualitative data are also been considered to support the findings. In addition, since all the nine cases experience a high level of food safety regulations, this characteristic may influence the application of common strategies used by all cases.

4.3 Influence of FPI characteristics on the application of

management strategies

Figure 3: Influence of FPI characteristics on the application of management strategies

(21)

management strategies mentioned above. Except perishability has a negative influence on the use of keeping inventory and product modulation, the rest characteristics all have a positive influence on the application of strategies. From the management strategies side, the

application of communication and information exchange, regional stock point and outsource delivery activities is not impacted by any of the characteristic according to the results. It is interesting to notice that some of the management strategies are influenced by more than one characteristic and some of the characteristics also impact the use of more than one strategies. These influences are explained in detail in the following paragraphs. Qualitative data are used to support these findings.

Comparing the strategies used by group 1 and group 2, we can conclude that the need of production at maximum capacity has a positive influence on the application of keeping inventory and workforce flexibility. As said before, most companies in the FPI experience a seasonal demand. If the company selects to fully use the production capacity, there will be additional stock in the low demand period. Therefore, Case B choose to stock the additional products to deal with the unexpected demands. Besides, during the peak demand period when production capacity is not enough to satisfy the customer demand, flexible workforce is required.

As mentioned before, all the cases experienced a high level of food safety regulation.

Considering the application of strategies and the qualitative data, food safety regulations have a positive influence on the application of employee training, multiple sourcing, continuously quality control and relationship management. Since there is high requirement on the quality of the products, employee training is required to ensure each employee is familiar with the production processes and sanitary requirements. In addition, continuous quality control guarantees that the quality of products satisfy the rules. As the head of operations of case C said “it is a going process where we continuously need to ensure that the food safety regulations that we refer to are up to date.” If a specific rule changes, companies have to make response in a limited time. For example, the sales manager of case B mentioned, “if the requirements of the proportion of elements changed, we need to not only reorganize the production according to the new rules but also consult with the suppliers about the raw material supply.” Therefore, multiple sourcing can reduce the risk of raw material supply, and a good relationship with supplier helps to make quick response in terms of changed

regulation.

(22)

continuously quality control. As mentioned above, multiple sourcing can provide more options in terms of supply failure, which can cope with the high variability of raw material supply. Besides, continuously quality control contributes to smooth the high variability of raw material quality and hence ensure a stable quality of end product.

Case B, F, G and I all experience high variability of production and use workforce flexibility and market segmentation as the management strategies. The fluctuation of customer demand results in a high variability of yield. Interviewees of case F and G mentioned segmenting customers into separate groups and using various inventory and distribution model help cope with the high variability of demand and, hence, manage the variability of SC.

Comparing the strategies used by group 1 and group 2, we can observe that perishability has a negative influence on keeping inventory and product modularization. This characteristics impact the inventory and the stock of modular in two ways. First, because of the perishability, there is a high requirement of the stock conditions. Usually, a suitable temperature is required. Some foods require cold storage warehouse. This increases the difficulty of keeping

inventory. Second, it is impossible to keep inventory for products whose shelf life are only few days, such as fresh bread. The products with a relatively long shelf life still experience a high risk of obsolescence because customer usually ask for 2/3 of the lifespan. To cope with that, companies usually use the first-in-first-out (FIFO) strategy and try to keep the turnover rate of inventory as large as possible. In addition, to reduce the losses of obsolescence, the expired products are always sent to other uses. For example, the supply chain coordinator of case E mentioned “there are some product that cannot be send to the customer anymore (past THT) then it is sent to be processed as cattle feed. But it could also happen that it is not even good enough for cattle anymore and then it is send off.”

Strategies used by group 1 and group 3 show that divergent product structure has a positive influence on the application of product modularization. A highly divergent product structure provides the opportunity to produce modules first and assemble them according to the customer demands. Both case A and case B produce intermediated products in advance and assemble them into different final products according to the demand forecast.

5. DISCUSSION

(23)

FPI characteristics with former studies. The third part further elaborates the influences of FPI characteristics on the application of the management strategies. For each part, findings are compared with what was expected based on literature.

5.1 Application of management strategies within FPI

It is not surprising that the five management strategies mentioned in section 2.2 all used by the cases in the FPI. The applications and advantages of these strategies are consistent with what described in previous studies. Furthermore, the study extends the theory by identifying five additional strategies to manage dynamic complexity of SC in FPI. Some lines of reasoning from literatures can back up the use of these strategies.

Findings show that outsourcing delivery activities are used to manage the speed of SC. This finding is consistent with the statement of Skjøtt-Larsen et al.(2007), who indicated that cooperating with 3PL contributes to faster access to new market and distribution channels. This is achieved by providing customers with fixed assets such as warehouse and

transportation network and professional knowledge and skills (Skjøtt-Larsen et al., 2007).

It is interesting to notices that except case C and F, all the rest cases used multiple sourcing for the raw materials of their main products to manage the uncertainty and variability of SC. The head of operations of case C also admitted that the use of single sourcing for their raw material causes risks. As mentioned on the findings, the application of multiple sourcing is mainly because of the nature of raw materials. Since the climate and weather have a great influence on the quality of raw materials in the FPI, there is a large risk of raw materials in terms of quality and supply. This is in line with the theory that multiple sourcing has the advantages of reducing the price by competitive tendering and providing options in case of supply failure (Skjøtt-Larsen et al., 2007; Slack, Chambers, & Johnston, 2004). Liker and Choi (2004) reached similar findings from Japanese car manufacturers. They stated that both Toyota and Honda develop more than two suppliers for every raw material they buy (Liker & Choi, 2004). By doing so, if the performances of one supplier are not adequate, competitor will get the next contract (Liker & Choi, 2004).

(24)

materials. From this point of view, the application of relationship management, continuous quality control and multiple sourcing to manage the variability of SC are interrelated. This is never been mentioned in previous literatures.

In addition, the findings show that some of the strategies can manage two complexity derivers. This point is in line with the finding of De Leeuw et al. (2013). However, they showed that keeping inventory is able to manage not only uncertainty but also speed, by reducing the order delivery time. Our study reveals that keeping inventory can only manage the uncertainty in FPI.

5.2 Food processing industry characteristics

The findings about FPI characteristics are consistent with the theory mentioned in section 2. Although variability of raw materials varies from case to case, interviewees still experience a relatively high variability of raw materials in the FPI. It is also predictable that all the cases experience different level of perishability, variability of production, divergent product structure, sequence dependency production and the need of production at maximum capacity. These findings are in accordance with Van Donk (2001).

It is interesting to remark that all the cases experience a high level of food safety regulations. Buckley (2015) indicated that regulation impacts the experience of food processors in three ways. First, a lot of time, expertise and financial resources are required to implement regulations (Buckley, 2015; Fielding, Ellis, Beveridge, & Peters, 2005; Worosz, Knight, Harris, & Conner, 2008). This is supported by the statement of the sales manager of case B, who indicated “the whole things need to be changed to react to the new safety rules within a limited time period.” Second, Buckley (2015) mentioned that these practical constraints cause a conception of unfairness, which makes small businesses feel that regulations are formulated against them. This sense of unfairness does not show in our cases. Third, studies also showed that regulation may stimulate business owners to improve their practices internally (Buckley, 2015). As the quality manager of case I mentioned, “our production is strictly based on ISO standards. This standard can help us to increase productivity while minimizing errors and waste.”

5.3 Influence of FPI characteristics on the application of

management strategies

Although former studies rarely related specific industry characteristics with the application of strategies, some linkages could still be found. The finding that no influence of FPI

(25)

outsourcing delivery activities could be expected beforehand. Communication and information exchange contributes to mitigate the bullwhip effect, which is a common phenomenon occurs in various industries (Costantino et al., 2015). Besides, whether to outsource delivery activities or not is an organizational selection. Therefore, the applications of both strategies do not rely on the type of industry.

It could also be expected that perishability in terms of raw materials, intermediate products and final products has a negative influence on keeping inventory and stocking modular. As can be obtained from the findings, short shelf life and high requirement on the stocking condition are the main incentives. Nahmias (1982) provide reasons to support this finding. They indicated that if company want to keep inventory for perishable products more complicated inventory model is required, which is time-consuming and costly.

In addition, Kilic et al. (2013) provide lines of reasoning to support that product

modularization is a common strategy used to mitigate the effect of divergent product structure in the FPI. This strategy is also related to postponement. However, they indicated that this strategy is not really active used in the FPI compared to other industry and most of the efforts are focused on the postponement practices at the packaging stages (Kilic et al., 2013). This statement is in accordance with the finding that four out of nine cases used product

modulation, which means it is not commonly used, and conflict with the finding that three cases apply this strategy by blending intermediate products into end products and only one case use it at the packaging stages.

6. CONCLUSION

The aim of this study is to identify the strategies to manage dynamic complexity and the influences of FPI characteristics on the application of management strategies. By doing so, providing companies in the FPI a guideline to manage the dynamic complexity of SC. To answer the research question 1, ten management strategies are identified in FPI and the application of these strategies are further explained. To answer the research question 2, the study links each FPI characteristic with the management strategies. The results show that industry characteristics indeed impact the application of strategies. Besides, except

perishability influences the use of strategy in a negative way, the rest characteristics increased the requirements of management strategies.

(26)

gap by expanding management strategies in FPI. Second, this is the first time that a study links the characteristics of an industry to the application of management strategies, which explores an interesting field concerning the application of management strategies in industry with inherent characteristics.

The main managerial implication of this study is that it provide managers a guideline to manage the dynamic complexity of SC. To be more specific, managers could use the questionnaire to measure their level of FPI characteristics, and by doing so, select the most suitable strategies to manage the dynamic complexity of their SC. For example, if manager wants to fully use their production capacity, workforce flexibility can be selected as a management strategy and if manager intends to provide products with divergent structure, product modularization is a good option according to the results.

A number of limitations of this research can be noted. One of limitation is the fact that for four of the nine cases, the interviews were only held with one employee. Since the

interviewees all have their own function in their respective company, the possibility of bias is increased by only held interview with one interviewee. In addition, the interviews were held by five different researchers. Although we designed the interview protocol together and discussed the results afterwards, it is possible that different researchers understand answers in different ways. Because of the time limitation, we did not send back our transcriptions to the interviewees for check and some lacks of information existed. This impacts the reliability of the results. Besides, because the results of research question 2 are derived from compare of qualitative data, personal bias may influence the results. This increases the difficulties of repeating the operations with the same results and hence reduces the reliability.

The research shows a number of interesting directions for future research. First of all, since the results show that the industry characteristics indeed influence the application of

(27)

REFERENCES

Arteta, B. M., & Giachetti, R. E. (2004). A measure of agility as the complexity of the enterprise system. Robotics and Computer-Integrated Manufacturing, 20(6), 495-503.

Beamon, B. M. (1998). Supply chain design and analysis: Models and methods.

International Journal of Production Economics, 55(3), 281–294.

Blecker, T., Kersten, W., & Meyer, C. (2005). Development of an Approach for Analyzing

Supply Chain Complexity. University Library of Munich, Germany.

Bozarth, C. C., Warsing, D. P., Flynn, B. B., & Flynn, E. J. (2009). The impact of supply chain complexity on manufacturing plant performance. Journal of Operations

Management, 27, 78–93.

Buckley, J. A. (2015). Food safety regulation and small processing: A case study of interactions between processors and inspectors. Food Policy, 51, 74-82.

Campbell, D. J. (1988). Task complexity: A review and analysis. Academy of

management review, 13(1), 40-52.

Costantino, F., Di Gravio, G., Shaban, A., & Tronci, M. (2015). The impact of information sharing on ordering policies to improve supply chain performances. Computers &

Industrial Engineering, 82, 127-142.

Dalton, D. R., Todor, W. D., Spendolini, M. J., Fielding, G. J., & Porter, L. W. (1980). Organization structure and performance: a critical review. Academy of

Management Review, 5(1), 49-64.

Dam, J. P. (1995). Scheduling packaging lines in the process industry. Rijksuniversiteit te Groningen.

Fielding, L. M., Ellis, L., Beveridge, C., & Peters, A. C. (2005). An evaluation of HACCP implementation status in UK small and medium enterprises in food

(28)

Fransoo, J. C., & Rutten, W. G. (1994). A typology of production control situations in process industries. International Journal of Operations & Production Management, 14(12), 47-57.

Hoole, R. (2005). Five ways to simplify your supply chain. Supply Chain Management: An

International Journal, 10(1), 3-6.

Isik, F. (2010). An entropy-based approach for measuring complexity in supply chains.

International Journal of Production Research, 48(12), 3681-3696.

Jongen, W. M., & Meulenberg, M. T. (1998). Innovation of food production systems:

product quality and consumer acceptance. Wageningen Pers.

Karlsson, C. (Ed.). (2009). Researching operations management. Routledge.

Kilic, O. A., Akkerman, R., van Donk, D. P., & Grunow, M. (2013). Intermediate product selection and blending in the food processing industry. International Journal of

Production Research, 51(1), 26-42.

Koudal, P., & Engel, D. A. (2007). Globalization and emerging markets the challenge of continuous global network optimization. Building Supply Chain Excellence in

Emerging Economies, 98, 37–66.

Lau, A. K. W., & Yam, R. C. M. (2005). A case study of product modularization on supply chain design and coordination in Hong Kong and China. Journal of Manufacturing

Technology Management, 16(4), 432-446.

Lambert, D. M., Cooper, M. C., & Pagh, J. D. (1998). Supply chain management: implementation issues and research opportunities. The international journal of

logistics Management, 9(2), 1-20.

Leeuw, S. De, Grotenhuis, R., & Goor, A. R. Van. (2013). Assessing complexity of supply chains: evidence from wholesalers. International Journal of Operations &

Production Management, 33, 960–980.

(29)

review, 82(12), 104-113.

Manuj, I., & Sahin, F. (2011). A model of supply chain and supply chain decision-making complexity. International Journal of Physical Distribution & Logistics

Management (Vol. 41, pp. 511–549).

Matthew B. Miles, & A. Michael Huberman. (1994). Qualitative data analysis: An expanded

sourcebook. Sage.

McIntosh, R. I., Matthews, J., Mullineux, G., & Medland, A. J. (2010). Late customisation: issues of mass customisation in the food industry. International Journal of

Production Research, 48(6), 1557-1574.

Meulenberg, M. T. G., & Viaene, J. (1998). Changing food marketing systems in western countries. Innovation of Food Marketing Systems.

Nahmias, S. (1982). Perishable inventory theory: a review. Operations Research, 30(4), 680– 708.

Norhayati, H., Rasma, S. I., & Mohd, K. A. (2013). Effect of storage conditions on quality of prebiotic dark chocolate. Malaysian journal of nutrition, 19(1), 111.

Perona, M., & Miragliotta, G. (2004). Complexity management and supply chain

performance assessment. A field study and a conceptual framework. International

Journal of Production Economics, 90, 103–115.

Serdarasan, S. (2013). A review of supply chain complexity drivers. Computers and

Industrial Engineering, 66(3), 533–540.

Shukla, M., & Jharkharia, S. (2013). Agri‐ fresh produce supply chain management: a state‐ of‐ the‐ art literature review. International Journal of Operations & Production

Management, 33, 114–158.

(30)

Skjøtt-Larsen, T., Mikkola, J. H., & Kotzab, H. (2007). Managing the global supply

chain. Copenhagen Business School Press.

Slack, N., Chambers, S. & R. Johnston (2004). Operations Management. 4th edition. Harlow: Prentice Hall.

Vachon, S., & Klassen, R. D. (2002). An exploratory investigation of the effects of supply chain complexity on delivery performance. IEEE Transactions on Engineering

Management, 49(3), 218–230.

Van Dam, P., Gaalman, G., & Sierksma, G. (1993). Scheduling of packaging lines in the process industry: An empirical investigation. International Journal of Production

Economics, 30, 579-589.

Van Donk, D. P. (2001). Make to stock or make to order: The decoupling point in the food processing industries. International Journal of Production Economics, 69, 297–306.

Van Wezel, W., Van Donk, D. P., & Gaalman, G. (2006). The planning flexibility bottleneck in food processing industries. Journal of Operations Management, 24, 287–300.

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

management. International journal of operations & production management, 22(2), 195-219.

Wood, R. E. (1986). Task complexity: Definition of the construct.Organizational

behavior and human decision processes, 37(1), 60-82.

Worosz, M. R., Knight, A. J., Harris, C. K., & Conner, D. S. (2008). Barriers to entry into the specialty red meat sector: The role of food safety regulation. Southern

Rural Sociology, 23(1), 170-207.

(31)

APPENDIX A: INTERVIEW PROTOCAL

(a) General Information:

1, Job-title of participant 2, Main products

3, Short description of production process (from the main product) 4, Number of employees (plant and organization)

5, Turnover plant in terms of money and volume 6, Main customers and geographical dispersion 7, Main suppliers and geographical dispersion

8, Position within the supply chain (from far upstream to far downstream) (b) Supply Chain Complexity

Complexity driver Definition

Uncertainty The lack of predictability and reliability of demand and of supply chain

in process

Diversity The number of different elements in a system (products, suppliers,

customers, activities)

Size Relative number or volume of elements (products or activities)

Variability Sudden, large and fluctuating changes in requirements imposed on the

system over time

Structure The interconnections between multiple systems, levels, processes, etc.

within and across elements

Speed Required responsiveness across the supply chain (speed at which

activities must be performed)

1, How do you experience the SC complexity in terms of each complexity driver? (c) Coping Strategies

1, What kind of strategies do you have in place in order to deal with supply chain complexity in terms of the complexity drivers discussed above?

2, What strategies do you have in place to manage your suppliers? 3, What strategies do you use to manage your demand?

4, What strategies are used internally (manufacturing) to cope with complexity? (d) Food Processing Industry

1, For your organisation, what are challenges of doing business within the food processing industry?

(32)

APPENDIX B: QUESTIONNAIRE

Very low Low Neutral High Very high 1, Variability of supply of raw materials



2, Variability of quality of raw materials



3, Variability of price of raw materials



4, Risk of obsolescence of the raw materials, semi-manufactured products and ‘end’ products

MANAGING SUPPLY CHAIN COMPLEXITY: A MULTIPLE CASE STUDY IN THE FOOD PROCESSING INDUSTRY