Effective food supply chains : generating, modelling and evaluating supply chain scenarios

Logistical co-ordination in FMCG supply chains

The overall objectives of the research described in this thesis were to obtain insight into the applicability of the concept Supply Chain Management (SCM) in food supply chains (SCs) from a logistical point of view, and to find an efficient and effective method to analyse and redesign the SC to improve SC performance.

Background

The background and rationale of this thesis are discussed in Chapter 1. Interest in SCM has been spurred by recent socio-economic developments. Because of demographic and socio-economic developments (e.g. strong increase of the ageing population, more double-income families) there is a growing demand for fresher products and products with higher added values. Furthermore, the effects of globalisation, the market entrance of new competitors, and stricter governmental requirements for food safety and environment-friendly production place increasing demands on management. These developments have resulted in a change in performance requirements for food SCs as a whole and, consequently, for all stages in the SC. Managers realised that sub-optimisation occurs if each organisation in a SC attempts to optimise its own results rather than to integrate its goals and activities with other organisations to optimise the results of the entire chain. This holds true especially in food SCs where particular actors in the SC can damage all the efforts taken in another stage to preserve high product quality. There has been growing recognition that it is through logistics and SCM that the twin goals of cost reduction and service enhancement can be achieved. The recent developments in Information and Communication Technology facilitate this process.

Research objective and questions

The main questions food companies face are whether , how, and with whom they should start SCM activities. They should be able to analyse what SCM can do for them and find out what the consequences might be if a SC view is taken together with one or more supplier(s) and/or customer(s). An extensive literature research did not reveal any integral method to generate, analyse and evaluate SC redesigns, i.e. SC scenarios (Chapters 2 and 4). Our research objective was therefore to contribute to the body of knowledge on SCM by developing a step-by-step approach that could generate, model and evaluate SC scenarios in specific food SCs. That is, we aimed to develop:

1. a research method to analyse a food SC and to generate a number of SC scenarios that are estimated to improve the current SC performance;
2. a research method to assess the impact of different SC scenarios for a particular food SC on SC performance and to identify a 'best practice' SC scenario.

A 'best practice' SC scenario refers to a feasible SC scenario that achieves the best possible outcome for the whole system with respect to predefined SC performance indicators. Both methods should assist managers of food companies in evaluating their current position in a food SC and in deciding whether and how they should redesign the SC.

In Chapter 1 we developed a proposition to guide our research. Because of rapid changes in markets, products, technology, and competitors, managers must make decisions on shorter notice, with less information, and with higher penalty costs than in the past. Decision making uncertainty has increased regarding what developments managers should react to and what impact possible actions may have. By breaking down the walls that are present between successive SC stages, decision making uncertainties may decline, since more information and control possibilities will become available to the decision-makers in each stage. This led us to the following proposition: To identify effective SC scenarios one should focus on the identification and management of the sources of uncertainties in SC decision making processes.

On the basis of this proposition the following three research questions were formulated:

1. What is the relationship between uncertainty in SC decision making processes and SC performance in food SCs?
2. How can we identify potentially effective SC scenarios for a particular food SC? (validation of the proposition)
3. How can SC scenarios be assessed with regard to SC performance and the individual performance of the SC participants?

Research design

Considering the research objectives and type of research questions to be answered, we used the multiple-embedded case study design. This research followed the inductive/deductive research cycle, in which literature and case studies were used to devise a research method on the generation, modelling and evaluation of SC scenarios. Three case studies were selected. Case I was conducted in a fresh food SC, comprising growers, auctions, an exporter of vegetables and fruits, and foreign retailers. Case II comprised a salad producer and a retail organisation (made up of a retail distribution centre and retail outlets). Finally, case III comprised two suppliers of desserts, a cheese producer and a retail organisation. We used the chosen case studies in two ways:

• All three exploratory case studies were used for theory building, addressing (1) the relationship between uncertainty and performance and (2) the identification of potentially effective SC scenarios.
• The two latter case studies were further elaborated upon to explore the area of (3) assessing the impact of SC scenarios on SC performance, in order to identify a 'best practice' SC scenario for that particular food SC.

Main definitions

Chapter 2 showed that there is no generally accepted definition of a SC and SCM. The dispute mainly focuses on the level of analysis. Based on the findings in literature, we defined a supply chain as the series of (physical and decision making) activities connected by material and information flows that cross organisational boundaries. SCM was defined as follows:

Supply Chain Management is the integrated planning, co-ordination and control of all logistical business processes and activities in the SC to deliver superior consumer value at less cost to the SC as a whole whilst satisfying the requirements of other stakeholders in the SC.

Our system and process view on SCs was presented in Chapter 3, resulting in a definition of a SC scenario:

A SC scenario is an internally consistent view of the settings of all SC redesign variables concerning the managed, managing, and information systems and organisation structure in the SC.

The four descriptive system elements comprising a SC scenario were described in detail in Chapter 3. A SC redesign variable is defined as a management decision variable at strategic, tactical or operational level that determines the setting of one aspect of the SC configuration or management and control. Furthermore, we identified Key Performance Indicators (KPIs) for food SCs that are needed to assess the effectiveness of SC scenarios, i.e. the degree to which the SC objectives are fulfilled. Finally, Chapter 3 concluded with a conceptual model that can be used to describe, analyse and typify a SC in detail to facilitate the SC redesign process.

Approach for SC analysis and redesign

In Chapter 4 we developed a preliminary research method for generating potentially effective SC scenarios. First, a review of literature in several areas (SCM, Logistics Management, Business Process Re-engineering and Operational Research) led us to a generic list of 22 SCM redesign principles that are thought to be able to improve performance on one or more SC KPIs. Each redesign principle refers to alternative settings for one or more of the SC redesign variables, thereby representing various SC scenarios. Second, by linking the list of redesign principles to potential sources of SC uncertainty in a SC, we found a means of identifying potentially effective SC redesign variables for that SC. Those sources of SC uncertainty that impact the SC KPIs are the first candidates for the redesign process. This approach was tested and further elaborated in Chapter 5, in which we discussed the findings of three exploratory case studies.

We concluded that sources of SC uncertainty refer to inherent characteristics of the SC and characteristics of the managed system, managing system, information system and/or organisation structure that are present at a certain point in time and that generate SC uncertainty. Our definition of SC uncertainty is based on the general requirements for effective control presented by De Leeuw (1988):

SC uncertainty refers to decision making situations in the SC in which the decision-maker lacks effective control actions or is unable to accurately predict the impact of possible control actions on system behaviour because of a lack of:

• information (or understanding) of the environment or current SC state;
• a consistent model of the SC presenting the relationships between SC redesign variables and SC performance indicators.

Generating SC scenarios in food supply chains

By applying the framework for SC analysis and redesign developed in Chapter 3 to the three case studies (Chapter 5), a detailed picture emerged of the configuration and operational control of activities in the SC. In this process, Organisation Description Language (ODL) and Event Process Chain (EPC) mapping techniques provided a powerful basis to redesign the SC, because it made the total process transparent and it made it possible for us to illustrate the opportunities for eliminating non-value-adding processes to managers. These techniques facilitated discussions with key employees in the SC and helped in identifying SC uncertainties and, more important, sources of SC uncertainty.

In all three cases the identification of SC uncertainties and especially their sources led to the recognition of potentially effective SC redesign variables. In this process the generic list of SCM redesign principles helped in identifying a complete overview of possible SC redesign strategies. By estimating the impact of each strategy on the SC KPIs, potentially effective SC redesign variables were identified.

In all three case studies an extended list of sources of SC uncertainty was identified. This allowed us to create a generic list of sources of SC uncertainty that may be found in food SCs. By linking this list to the generic list of SCM redesign principles developed in Chapter 4, we generated a valuable tool for SC redesign projects. It lists potential improvement areas in the SC when certain types of SC uncertainty are encountered in an investigated SC.

Modelling and evaluating SC scenarios

After SC scenarios are identified, we need an approach to model and evaluate these scenarios to determine the best practice SC scenario. In Chapter 6 we concluded that the combination of two methods is most effective:

• First, a mathematical model of the SC has to be built that allows for the quantitative assessment of the impact of SC scenarios on the SC KPIs. Because of multiple performance- and time-related process aspects that need to be taken into account when modelling SCs we focussed in our research on simulation instead of analytical modelling.
• Second, a field test experiment has to be conducted in the SC comprising one of the most promising SC scenarios, to provide practical and organisational restrictions and to thus reveal the feasibility of alternative SC designs.

In Chapter 6 we also developed a modelling framework for modelling the dynamic behaviour of food SCs. We believe the modelling framework captures all relevant concepts of the SC system needed to adequately model and simulate SC scenarios. The main modelling components are business processes, business entities, databases, resources, performance indicators and redesign variables. To model the dynamic behaviour of SCs some general assumptions are used as a starting point. For example, we assume system hierarchy, and we view the SC as a network of business processes with precedence relationships that use resources, and as a dynamic system with changing performance characteristics (especially when the time frames considered change). Two simulation tools, ExSpect and Arena, are introduced that were used in the two case studies in Chapter 7 for evaluating SC scenarios.

The case studies showed the applicability of the modelling approach. Simulation results, validated by managers of the participating companies (expert validation), showed the trends and order of magnitude of changes in SC performances of different SC scenarios. Both ExSpect and Arena proved to be promising tools for modelling SC scenarios, although each of them had some disadvantages.

The case studies also showed the complementarity of the field test to a simulation study; in one case study the best performing scenario in the simulation study was exchanged for a feasible scenario Y because of the field test results. By using this evaluation approach, SC managers can be supported in deciding whether or not to implement a new SC scenario. In both case studies, SC scenarios were suggested that perform considerably better than the current design.

Finally, an analytical approach was suggested for model validation purposes. It showed that analytical models can be used very successfully for smaller SC problems.

A step-by-step approach to generate, model and evaluate supply chain scenarios

Chapter 8 summarised our complete step-by-step approach for the analysis and redesign of food SCs which aims at identifying a best practice SC scenario for a particular food SC (Figure 1).

The approach starts with the joint definition of the boundaries of the SC to be investigated and the SC objectives ( step 1 ). From these objectives the SC KPIs are identified. By describing the SC processes in detail SC uncertainties and their sources are identified ( step 2 ). Effective SC scenarios are identified by linking the main sources of uncertainty found in the SC ( step 3 ) with the list of SCM redesign principles ( step 4 ). The potential SC scenarios are evaluated quantitatively and qualitatively by, respectively, a model study ( step 5 ) and a field test experiment ( step 6 ). Finally, a 'best practice' scenario can be identified and implemented in practice ( step 7 ).

Figure 1. A step-by-step approach to generate, model and evaluate SC scenarios.

Main conclusions

Finally, Chapter 9 summarised the main findings and answered the three research questions as follows:

1. What is the relationship between uncertainty in SC decision making processes and the SC performance in food SCs?

   The presence of uncertainties in SC decision making situations results in the establishment of safety buffers in time, capacity and/or inventory to prevent a poor SC performance. These safety buffers initiate the existence of several non-value-adding activities that reduce the profitability of the SC. Reducing or even eliminating the SC uncertainties will improve SC performance.

2. How can we identify potentially effective SC scenarios for a particular food SC?

   The use of our step-by-step approach for generating effective SC scenarios will result in a number of important SC redesign variables. The combination of different settings for these SC redesign variables establishes a number of potentially effective SC scenarios. The degree of effectiveness has to be determined in the evaluation phase (research question 3).

3. How can SC scenarios be assessed with regard to SC performance and the individual performance of the SC participants?

   The impact of a SC scenario on SC performance should preferably be evaluated by using both a modelling (simulation) study and a field test experiment. In this way the new SC scenario is evaluated both qualitatively (i.e. considering the behavioural and organisational aspects) and quantitatively. The modelling framework developed in this research provides a good means for capturing the dynamic behaviour of a food SC. The final decision of which SC scenario to implement depends on the trade-off between multiple SC performance indicators for each SC participant and the SC as a whole, and the feasibility of each SC scenario.

We are confident that we have been successful in contributing to the body of knowledge on SCM. Our step-by-step approach (including the suggested tools and techniques) can be used to analyse, redesign and evaluate food SCs more effectively and efficiently, which enhances the competitive advantage of the SCs. It enables managers to assess the impact of SC scenarios on SC performance; it provides insight into SC functioning; and it works as a facilitator to allow managers to rethink current SC business processes.

Opportunities for further research

In our view SCM research should focus on the construction of a toolbox comprising theories, methods and techniques, and working applications to analyse and improve the management of the SC. We acknowledge that our step-by-step approach to generate, model and evaluate SC scenarios should be considered as a first step towards a generic toolbox that can be used to improve SC performances. This thesis concluded by suggesting some interesting areas for further research. The most important, in no particular order, are the following:

• Refinement of the lists of SCM redesign principles and sources of SC uncertainty (by reviewing other literature and doing case studies in other industries).
• Adoption of a SC network perspective within SCM research emphasising the investigation of SC interactions.
• Definition of unique and integrated SC performance measures.
• Further research into the applicability of different types of quantitative methods combined with a typology of SC decision situations.
• Elaboration and refinement of the SC modelling framework and the construction of a library of generic building blocks on the level of individual business processes to facilitate efficient and effective modelling of SCs.
• The construction of a Decision Support System for SC analysis and redesign, using our step-by-step approach and the case study results, which generates SC scenarios automatically.
• Incorporation of requirements of other stakeholders in the SC analysis, such as the government, environmental protection groups, and trade-unions.