Supply chain performance evaluation with data envelopment analysis and balanced scorecard approach

Accepted Manuscript Supply Chain Performance Evaluation with Data Envelopment Analysis and Balanced Scorecard Approach M. Shafiee, F. Hosseinzadeh Lotfi, H. Saleh PII: DOI: Reference:

S0307-904X(14)00126-7 http://dx.doi.org/10.1016/j.apm.2014.03.023 APM 9911

To appear in:

Appl. Math. Modelling

Received Date: Revised Date: Accepted Date:

16 December 2012 9 March 2014 18 March 2014

Please cite this article as: M. Shafiee, F. Hosseinzadeh Lotfi, H. Saleh, Supply Chain Performance Evaluation with Data Envelopment Analysis and Balanced Scorecard Approach, Appl. Math. Modelling (2014), doi: http:// dx.doi.org/10.1016/j.apm.2014.03.023

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Supply Chain Performance Evaluation with Data Envelopment Analysis and Balanced Scorecard Approach M.shafieea+, F.Hosseinzadeh Lotfib, H.Salehb a b

Department of Industrial Management, Shiraz Branch, Islamic Azad University, Shiraz, Iran

Department of Mathematics, Science and Research Branch, Islamic Azad University, Tehran, Iran

Abstract. One of the most complicated decision making problems for managers is the evaluation of supply chain (SC) performance which involves various criteria. Though vast studies have been recorded on supply chain efficiency evaluation via balanced scorecard (BSC) approach, these studies do not focus on the relationships between the four perspectives of BSC approach. The present paper is an attempt focusing on these relationships, especially the returnable ones. To do so, at first, all relationships between the four perspectives of BSC were determined and then the DEMATEL approach was employed to obtain a network structure. This network structure was then used to create a network DEA model. Since it was not possible to calculate the efficiency evaluation score by BSC, the data envelopment analysis (DEA) model was used for such an evaluation. Moreover, after reviewing different tools to evaluate the performance of supply chain, a new approach, relying on network DEA with BSC approach, was generated. Finally, this model was applied in the Iranian food industry to evaluate its supply chains efficiency and the results proved the high efficiency of the model designed. The findings could be used in various evaluation processes in different industries.

Keywords: Supply Chain Management, Balanced Scorecard (BSC), DEMATEL, Data Envelopment Analysis (DEA), Network DEA

1. Introduction Market globalization has made supply chain management an interesting topic to be discussed: An efficient supply chain can lead to a range of benefits including reduced cost, increased market share and sales, and sustainable customer relationships (Ferqusen, 2000). It has also been cited that evaluation of supply chain performance can improve the overall performance of the organization (Chen & Paulraj, 2004). Efficiency of the supply chain is the result of integration of the performance of all members. As such, managing the overall supply chain efficiency is a challenging task. In 1900 William Durant-the founder of General Motors-claimed that profit is the outcome of a cost stream that spreads throughout the supply chain, not the result of an accounting exercise. Since then, the principle of identifying profit and controlling cash flow has been used to evaluate organizational performance. Generally, the efficiency of the supply chain, which is usually managed as a series of simple +

Corresponding author. Department of Industrial Management, Shiraz Branch, Islamic Azad University, Shiraz, IRAN Tel.:(+ 989179248672); fax: (+987116470060) E-mail address: ([email protected]) 1

business functions, is measured by taking the ratio of revenue over the total supply chain operational cost. However, since the demands for quick order fulfillment and fast delivery are increasing, new trends have emerged. In addition to the usual financial measures, other specific indicators such as customer satisfaction should be taken into consideration. The emergence of multiple performance measures has made the efficiency measurement a difficult and sophisticated task. Hence, the means utilized to measure the performance should provide not only the quantitative reasoning, but also the qualitative perspective to remain aligned with the strategic goals of the organization. Since "Performance" and "predetermined parameters", can be defined as "measurement" and "the ability to monitor events and activities in a meaningful way", respectively, performance measurement can be defined as the process of quantifying the effectiveness and efficiency of action (Neely et al., 1995). A number of approaches have been mentioned to measure performance. Some of most prominent ones frequently refereed to are: balanced scorecard (Kaplan, Norton, 1992), the performance measurement matrix (Keagan et al., 1989), performance measurement questionnaire (Dixon et al., 1990), criteria for measurement system design (Globerson, 1985), and computer aided manufacturing approaches. However, in practice these approaches have suffered some shortcomings including lack of strategic focus, forcing managers to encourage local optimization rather than seeking the continuous improvement and, being incapable of providing adequate information about competitors. For the application of performance measurement, it is essential that companies’ tangible and intangible targets be defined the way that is more appropriate to the requirements and objectives of these targets, and that its strategy is more extensively operationalized, quantified, and linked in a mutually supplementing way (Chi-Sun, 2010). As emphasized by Ghalayini and Noble (1996), the literature concerning performance measurement has two phases: In the first phase, which went on until the 1980s, the center of attention was performance measurement based on the financial criteria supplied by the management accounting system. The second phase which started in the late 1980s, and is still proceeding, brought about many changes within performance measurement, and interest in this field has increased tremendously. In the late 1980s, the problems of the traditional way of measuring performance were clearly known and researchers started to discuss introducing new performance measures such as shareholder value, economic profit, customer satisfaction, internal operations performance, and intellectual capital and intangible assets (Neely & Bourne, 2000). One of the approaches for this purpose was Balanced Scorecard (BSC) proposed by Kaplan and Norton (1996). They argued that BSC provides managers with the means they need to navigate future competitive success. It included more non-financial measures derived specifically from the organization’s strategy. BSC is one of the most comprehensive and simple performance measurement means which emphasizes both aspects of financial and non-financial, long-term and short-term strategies as well as internal and external business measures. Several studies have focused on the evaluating performance of supply chain based on BSC (Bhagwat &Sharma, 2007; Bigliardi & Bottani, 2010; Chia, Goh & Hum, 2009; Park, Lee & Yoo, 2005; Sharma & Bhagwat, 2007; Varma, Wadhwa & Deshmukh, 2007). The strongest point of BSC is its ability to illustrate 2

the cause and effect relations between strategies and processes through the four BSC’s perspectives of financial perspective, customer perspective, internal business process perspective, as well as learning and growth perspective. In other words, the relationships between the four perspectives of BSC are very important for performance evaluation. There have been quite a few studies conducted on these relations (Amado et al., 2011; Asosheh et al., 2010; Chen et al., 2007; Chen et al., 2008; Eilat et al., 2005; Eilatet al., 2007; Min et al., 2008; Richards, 2003; Valderrama et al., 2009; Wang et al., 2006). These studies determine the efficiency and performance in organizations and R&D projects. In these studies, the researchers do not focus on all possible relationships between the four perspectives of BSC. The fact is that they study these relationships in a simple manner as in their models the learning and growth perspective influences the internal business process perspective, and the internal business process influences the customer perspective, and the customer perspective influences the financial perspective. It must be mentioned that other important relationships between these perspectives may exist. Considering these possible relationships, particularly the returnable ones, is exactly what this paper tries to take into account, based on the network structure for supply chain performance evaluation. Since it is not possible to determine the efficiency evaluation score by applying BSC, we used the data envelopment analysis model (DEA) to calculate the efficiency score of supply chain performance. DEA is a non-parametric method to analyze efficiency, proposed by Charnes et al (1978) to produce the efficiency frontier based on the concept of Pareto optimum. DEA is also a powerful means in evaluating organizations with multiple inputs and outputs and takes the qualitative and quantitative measures into account. There have been conducted some studies on the combination of DEA and BSC (Amado et al., 2011; Asosheh et al., 2010; Chen et al., 2007; Chen et al., 2008; Eilat et al., 2005; Eilatet al., 2007; Min et al., 2008; Richards, 2003; Valderrama et al., 2009; Wang et al., 2006). In these studies the classic DEA model of efficiency evaluation has been used and the researchers have not focused on all possible relationships between the four perspectives of BSC. In several studies, the researchers have determined all possible relationships between the four perspectives of BSC (Teseng, 2010; Wu et al., 2011), however, they have not been able to determine the efficiency score for the decision making unit. To meet the purpose mentioned above, determining the efficiency score for the decision making unit, the present authors created a network structure by determining these relations and then created a new multi stage DEA model for efficiency evaluation of the supply chain. The proposed model involves all possible relationships between the perspectives, especially the returnable ones. This proposed approach is designed to solve the problems of measuring supply chain performance, in particular: -

The identification of the performance structure in which the parameters and necessary metrics for calculating the performance are identified.

-

The identification of the links between the required parameters and the metrics in relation to the achieved objectives.

In general, this paper is organized as follows: first, a brief description of some traditional tools of measuring supply chain performance is given. After that, DEA and Multi stage DEA and their applications 3

and concepts associated with the supply chain are reviewed. This is followed by an explanation of the methodology and Multi stage DEA by BSC approach models developed to measure supply chain efficiency and their application in the Iranian food industry supply chains. The last section of the article provides the readers with some conclusions and implications of the study.

2. Background 2.1 Traditional Methods to Measure Supply Chain Efficiency In the process of evaluating the performance of the supply chain, choosing performance measures is an important task because the action of management and solution for improvement are derived from them. Obviously, these measures vary from field to field. Reviewing literature shows that early performance measures usually focused on cost and because the metrics of cost were easy to understand, managers used to welcome it more (Ballou et al, 2000; Ellram, 2002). But inflexibility and lack of integration with strategic focus made researchers look for better measures containing quantitative as well as qualitative measures in the supply chain. In 1999, Beamon identified three types of measures which involved resources, output, and flexibility. Extending these measures leads to providing a new framework for supply chain evaluation that measures the strategic, tactical, and operational level of performance. Pittiglio, Rabin, Todd, and McGrath generated the first universal performance measures known as PRTM. This was the first comprehensive method providing a world class supply chain measurement. In PRTM, the keys for excellence of the supply chain are identified as: delivery performance, flexibility and responsiveness, logistics and cost, and asset management. The concept of PRTM was extended and the supply chain council proposed the supply chain operations reference (SCOR) model (Stewart, 1995) which became the first cross-industry framework for evaluating supply chain performance. SCOR is structured in 4 levels based on a plan, source, make, and deliver framework and the metrics used in SCOR include a broad range such as delivery performance, order fulfillment, production flexibility, and cash-to-cash cycle time. Although the study of performance measurement was enriched by different researchers and findings, some gaps still exist in certain aspects. Lack of valid measurement criteria and inadequate methodologies to aggregate different performance measures into a single index is one of them. Most methodologies were unable to consider the relative importance of measures which varies among the firms. In addition, there was no aggregate measure of overall supply chain performance that could be utilized to compare performance with other industry members. Despite the numerous advantages of SCOR, Samuel et al. (2004) showed that utilization of SCOR seems to be rather rigid and needs further enhancement. Since networks of the supply chain are becoming more complex, the SCOR model needs to be more dynamic and should be able to provide an adequate platform to measure these complex features. Even though SCOR provides deterministic performance metrics that are controllable by managers and administrators, it should be more dynamic to be able to synchronize

4

different elements. Also, by the review of literature, it can be concluded that past works had failed to address the collaborative relationship in the areas that involve joint decision making.

2.2 Efficiency Measurement of Supply Chain Efficiency evaluation is an important activity for the survival and growth of any firm. As the old adage goes: ‘you cannot improve what you cannot measure’, organizations may need to carry out efficiency measurement for different purposes such as: identifying success, identifying whether they are meeting customer requirements, helping them understand their processes, identifying where problems bottleneck, waste, etc., exist, where improvement is necessary, ensuring that decisions are objective rather than subjective, and showing if improvement planned actually happened (Parker, 2000). It is worth mentioning that the main purpose of efficiency measurement is to evaluate, control and improve operation processes (Ghalayini & Noble, 1996). One of the most significant paradigm shifts of modern business management is that individual businesses no longer compete as solely autonomous entities, but rather as supply chains (Lambert & Cooper, 2000). Supply Chain Management (SCM) is the practice of coordinating the flow of goods, services, information and finances that move from raw material to wholesaler to retailer to consumer (Russell, 2001). SCM is being heralded as a value driver because it has such a wide ranging effect on business success or failure (Farris & Hutchison, 2002); meanwhile, the main reason for poor performance of the supply chain is the lack of a measurement system (Morphy, 1999). This measurement system provides management with a set of actions that can be taken in improving performance and planning competitiveness enhancing efforts (Hoek, 1998). Organizations need to measure not only the final output, but also the processes involved in reaching the final output in order to locate the problem which is causing the variance between the target and the actual specification of the final product (Varma et al., 2008). A good number of studies have concentrated on the efficiency and performance evaluation of supply chain. In this section the most important ones are mentioned: The SCOR model (Supply Chain Council, 2006) was introduced in 1996. Beamon (1998) divided performance measures into two groups of quantitative and qualitative to discuss customer satisfaction and responsiveness, flexibility, supplier performance, cost and other elements of supply chain efficiency modeling. He identified three types of performance measure in 1999. These measures are vital components for the supply chain performance measurement system that includes resource, output and flexibility. Gunasekaran et al. (2001) proposed a framework for determining the performance of supply chain according to the strategic, tactical and operational levels in the supply chain. This framework deals with supplier delivery, customer service, inventory and logistic cost. Husman (2002) determined the performance of supply chain by three metrics including service, asset and speed. Fleix et al. (2003) built a model for determining the overall performance of complex supply chain by applying a systematic process-based approach. Agarwal et al. (2006) applied a framework in which the market sensitiveness, process integration, information driver and flexibility are used for determining the performance of the supply chain. Soni and Kodali (2010) used the facilities, transportation, information, inventory, sourcing and pricing categories of measurement to compare the performance of various supply 5

chains and point out the poor performance functions. There are several studies which use mathematical operations such as maximum, minimum, sum, statistical approach, AHP, ANP, and so on, to integrate their model (Agarwal et al., 2006; Angerhofer & Angelides, 2006; Berrah & Cliville, 2007; Bhagwat & Sharma, 2007; Chang et al., 2007; Sharma et al., 2007). The interactions among various parameters however have not been considered in the aforementioned research. Although the excellent overview of performance measurement provided by Neely et al. (1995) has been widely cited in recent research on supply chain performance measurement systems and metrics (Beamon, 1999; Beamon and Chen, 2001; Gunasekaran et al,, 2001, 2004), the need for new measurement systems and metrics is felt ad reported in the research conducted recently; meanwhile each study focuses on different kinds of measurements. The present authors however have tried to collect all of these measurements and studies and divide them based on several concepts such as: -

Whether they are qualitative or quantitative (Beamon 1999; Chan, 2003).

-

What they measure: cost and non-cost (De Toni & Tonchia, 2001; Gunasekaran, 2001); quality, cost, delivery and flexibility (Schnetzler et al., 2007); cost, quality, resource utilization, flexibility, visibility, trust and innovativeness (Chan, 2003); resource, output and flexibility (Beamon, 1999); supply chain collaboration efficiency; coordination efficiency and configuration (Chan, 2003); and input, output and composite measures (Chan and Qi, 2003).

-

Their strategic, operational or tactical focus (Gunasekaran et al., 2001).

-

The process in the supply chain they relate to (Chan and Qi, 2003; Huang et al., 2004; Li et al., 2005; Lockamy & McCormack, 2004; Stephens, 2001).

-

their applicability to the five supply chain process defined in the supply chain operations reference (SCOR) model (plan, source, make, deliver and return or customer satisfaction)

-

resource, output and flexibility measure (Beamon, 1999)

-

the Cooper, Lambert and Croxtion model (1997) (GSCF)

There are also some problems in the studies about the performance evaluation of supply chain which go as follows: -

Lack of connection with the strategy (Beamon, 1999; Chan & Qi, 2003; Gunasekaran et al., 2004).

-

Encouragement of short termism

-

Focus on cost to the detriment of non-cost indicators (Beamon, 1999; De Toni & Tonchia, 2001).

-

Lack of a balanced approach (Beamon, 1999; Chan, 2003).

-

Insufficient focus on customer and competitors (Beamon, 1999).

-

Loss of supply chain context, thus encouraging local optimization (Beamon, 1999).

-

Lack of system thinking (Chan, 2003; Chan & Qi, 2003).

-

Failure to provide adequate information on what competitors are doing through benchmarking.

Although recent studies have several problems, they have several strengths as they;

6

-

present methods for benchmarking (Lohman, et al., 2004)

-

consider global as well as local optimization (Lohman, et al., 2004)

-

pay attention to environment dynamic (Lohman et al., 2004; Schnetzler et al., 2007; Vorst, 2001)

-

systemically look to supply chain (Angerhofer et al., 2006; Chan, 2003; Chan et al., 2003; Gunasekaran et al., 2001).

Having analyzed the previous works, the present authors assert that there are a lot studies carried out on performance and the efficiency of supply chain management, but there are still two main needs to study supply chain performance and efficiency evaluation as Gunasekaran et al. (2001, 2004) claimed. These two main needs are: -

Lack of a balanced approach

-

Lack of a clear distinction between metrics and measures at a strategic, tactical and operational level

The present paper has focused on a comprehensive method to study the measures of the supply chain performance and efficiency and the comprehensive supply chain with end-to-end approach. To do so, the balanced scorecard approach was used to study non-financial measures as well as financial measures. Therefore, it is clear that to have effective SCM, measurement goals must consider the overall efficiency and the metrics to be used. It should represent a balanced approach and should be classified as strategic, tactical and operational levels and be financial and non-financial measures as well. Taking into account the above factors, a balanced SCM scorecard has been proposed and developed to discuss several measures and metrics of SCM. A balanced performance evaluation of SCM (BSC) not only helps organizations in faster and wider monitoring of their operations, but can also help them in improving their internal and external function of business such as engineering and design applications, production, quality improvement, material management, quick response, gaining lost market shares, and proper implementation of business strategies. Several studies have researched the efficiency and performance of the supply chain based on the BSC approach (Bigliardi & Bottani, 2010; Bhagwat & Sharma, 2007; Chia, Goh & Hum, 2009; Park, Lee & Yoo, 2005; Sharma & Bhagwat, 2007; Varma, Wadhwa & Deshmukh, 2007). These studies have divided the metrics of supply chain efficiency evaluation into four perspectives of BSC. These are listed in Table 1. These metrics and measures and others metrics identified in the literature review provide the theoretical bases for this paper in order to determine the measures of supply chain efficiency based on BSC approach. Since the metrics employed in supply chain efficiency evaluation are vast, the most important of these metrics are shown in the following table. The internal process perspective measures

The customer perspective measures

Total supply chain cycle time(Bhawat et al., 2007);(Sharma et al., 2007) Total cash flow time(Bhawat et al., 2007);(Sharma et al., 2007) Flexibility of service systems to meet particular customer needs(Bhawat et

Customer query time(Bhawat et al., 2007);(Sharma et al., 2007); (Bigliardi et al.,

al., 2007);(Sharma et al., 2007)

et al., 2007)

Supplier lead time against industry norms(Bhawat et al., 2007);(Sharma et al., 2007)

Range of products and services(Bhawat et al., 2007);(Sharma et al., 2007) Order lead time(Bhawat et al., 2007);(Sharma et al., 2007); (Bigliardi et al.,

Level of supplier’s defect free deliveries(Bhawat et al., 2007);(Sharma et al.,

2010)

2007)

Flexibility of service systems to meet particular customer needs(Bhawat et

Accuracy of forecasting techniques(Bhawat et al., 2007);(Sharma et al., 2007);

al., 2007);(Sharma et al., 2007)

(Bigliardi et al., 2010)

et al., 2010)

Buyer–supplier partnership level(Bhawat et al., 2007);(Sharma et al., 2007) Delivery lead time(Bhawat et al., 2007);(Sharma et al., 2007) Delivery performance(Bhawat et al., 2007);(Sharma et al., 2007) Effectiveness of delivery invoice methods(Bhawat et al., 2007);(Sharma et

Planned process cycle time(Bhawat et al., 2007);(Sharma et al., 2007);

al., 2007)

Product development cycle time(Bhawat et al., 2007);(Sharma et al., 2007) Purchase order cycle time(Bhawat et al., 2007);(Sharma et al., 2007); (Bigliardi

7

2010)

Level of customer perceived value of product(Bhawat et al., 2007);(Sharma

(Bigliardi et al., 2010)

Effectiveness of master production schedule(Bhawat et al., 2007);(Sharma et al., 2007); (Bigliardi et al., 2010) Capacity utilization(Bhawat et al., 2007);(Sharma et al., 2007) Total inventory cost as:(Bigliardi et al., 2010); (Bigliardi et al., 2010) Incoming stock level(Bhawat et al., 2007);(Sharma et al., 2007) Work-in-progress(Bhawat et al., 2007);(Sharma et al., 2007) Scrap value(Bhawat et al., 2007);(Sharma et al., 2007) Finished goods in transit(Bhawat et al., 2007);(Sharma et al., 2007) Supplier rejection rate(Bhawat et al., 2007);(Sharma et al., 2007); (Bigliardi et al., 2010) Efficiency of purchase order cycle time(Bhawat et al., 2007);(Sharma et al., 2007) Frequency of delivery(Bhawat et al., 2007);(Sharma et al., 2007); (Bigliardi et al., 2010) Manufacturing lead time(Park et al., 2005) Yield(Park et al., 2005) Perished inventory (Park et al., 2005) Obsolete inventory (Park et al., 2005) Inventory accuracy (Park et al., 2005) Material inventory (Park et al., 2005) Material stock-out(Park et al., 2005) Delivery flexibility(Park et al., 2005) Truck cube utilization(Park et al., 2005) Responsiveness to urgent order(Park et al., 2005) Adherence to schedule(Park et al., 2005) Forecast accuracy(Park et al., 2005) Volume flexibility(Park et al., 2005) Mix flexibility(Park et al., 2005) New product time to market(Park et al., 2005) Percentage of sales from new products(Park et al., 2005) Steady supply of raw material(Varma et al., 2008) Transportation costs(Varma et al., 2008) Inventory costs(Varma et al., 2008) Integration with supply chain partners(Varma et al., 2008) Optimization of enterprise(Varma et al., 2008) Volume flexibility(Varma et al., 2008) Quality of purchased goods(Park et al., 2005) Procurement administration cost(Park et al., 2005) Price of purchased cost(Park et al., 2005) Time for successful bids(Park et al., 2005) Material return rate(Park et al., 2005) Supplier on-time delivery(Park et al., 2005) Inventory information sharing(Park et al., 2005) Order information sharing(Park et al., 2005) Forecast information sharing(Park et al., 2005) Trust with partners(Park et al., 2005) Percentage of online purchase(Park et al., 2005) Order processing(Park et al., 2005) Purchase order fill rate (Park et al., 2005) Quality services(Chia et al., 2009) New services implemented per year(Chia et al., 2009) On time delivery(Chia et al., 2009) Waste reduction(Chia et al., 2009) ****** to be continued********

The financial perspective measures

Delivery reliability(Bhawat et al., 2007);(Sharma et al., 2007) Responsiveness to urgent deliveries(Bhawat et al., 2007);(Sharma et al., 2007) Effectiveness of distribution planning schedule(Bhawat et al., 2007);(Sharma et al., 2007)

Information carrying cost(Bhawat et al., 2007);(Sharma et al., 2007) Quality of delivery documentation(Bhawat et al., 2007);(Sharma et al., 2007) Driver reliability for performance(Bhawat et al., 2007);(Sharma et al., 2007) Quality of delivered goods(Bhawat et al., 2007);(Sharma et al., 2007) Achievement of defect free deliveries(Bhawat et al., 2007);(Sharma et al., 2007)

Product quality(Park et al., 2005) Product price(Park et al., 2005) Range of products and services(Park et al., 2005) Customer’s product return rate(Park et al., 2005) Customer response time (Park et al., 2005) On-time delivery (Park et al., 2005) Finished goods inventory (Park et al., 2005) Finished goods stock-out (Park et al., 2005) Repeat vs new customer sales (Park et al., 2005) Order fill rate (Park et al., 2005) Order tracking performance (Park et al., 2005) Percentage of resolving customer’s first call(Park et al., 2005) Image(Park et al., 2005) Reputation(Park et al., 2005) Purity of product(Varma et al., 2008) Steady supply of finished product(Varma et al., 2008) Distribution lead time(Bigliardi et al., 2010) Distribution performance(Bigliardi et al., 2010) Delivery reliability(Bigliardi et al., 2010) Effectiveness of distribution planning schedule(Bigliardi et al., 2010) Quality of delivery goods(Bigliardi et al., 2010) Customer perceived value of product(Bigliardi et al., 2010) Flexibility of service system to meet particular customer needs(Bigliardi et al., 2010)

Responsiveness to urgent delivery (Bigliardi et al., 2010) Market share (Chia et al., 2009) Number of customer retained(Chia et al., 2009) Customer satisfaction(Chia et al., 2009) ****** to be continued********

The learning and growth perspective measures

Net profit vs. productivity ratio(Bhawat et al., 2007);(Sharma et al., 2007) Rate of return on investment(Bhawat et al., 2007);(Sharma et al., 2007); (Chia

Supplier assistance in solving technical problems(Bhawat et al.,

et al., 2009)

Supplier ability to respond to quality problems(Bhawat et al., 2007);(Sharma

Variations against budget(Bhawat et al., 2007);(Sharma et al., 2007); (Bigliardi

et al., 2007); (Bigliardi et al., 2010) Supplier cost saving initiatives(Bhawat et al., 2007);(Sharma et al., 2007) Supplier’s booking in procedures(Bhawat et al., 2007);(Sharma et al., 2007) Capacity utilization(Bhawat et al., 2007);(Sharma et al., 2007) Order entry methods(Bhawat et al., 2007);(Sharma et al., 2007); (Bigliardi et al., 2010) Accuracy of forecasting techniques(Bhawat et al., 2007);(Sharma et al., 2007) Product development cycle time(Bhawat et al., 2007);(Sharma et al., 2007) Flexibility of service systems to meet particular customer needs(Bhawat et al., 2007);(Sharma et al., 2007) Buyer–supplier partnership level(Bhawat et al., 2007);(Sharma et al., 2007) Range of products and services(Bhawat et al., 2007);(Sharma et al., 2007) Level of customer perceived value of product(Bhawat et al., 2007);(Sharma et al., 2007) Human capital(Park et al., 2005) Information capital(Park et al., 2005) Organizational capital(Park et al., 2005) Use of IT (Varma et al., 2008) Postponement (Varma et al., 2008) Level of information sharing (Bigliardi et al., 2010)

et al., 2010)

Buyer–supplier partnership level(Bhawat et al., 2007);(Sharma et al., 2007) Delivery performance(Bhawat et al., 2007);(Sharma et al., 2007) Supplier cost saving initiatives(Bhawat et al., 2007);(Sharma et al., 2007) (Bigliardi et al., 2010) Delivery reliability(Bhawat et al., 2007);(Sharma et al., 2007) Cost per operation hour(Bhawat et al., 2007);(Sharma et al., 2007); (Bigliardi et al., 2010) Information carrying cost(Bhawat et al., 2007);(Sharma et al., 2007) Supplier rejection rate(Bhawat et al., 2007);(Sharma et al., 2007) Total Profit(Park et al., 2005) Total revenue(Park et al., 2005) Sales growth(Park et al., 2005) Total cost(Park et al., 2005) Cost per unit produced(Park et al., 2005) Inventory carrying cost(Park et al., 2005) Delivery cost(Park et al., 2005) Setup/change-over cost(Park et al., 2005) Cash flows(Park et al., 2005)

8

2007);(Sharma et al., 2007); (Bigliardi et al., 2010)

Customer query time; (Sharma et al., 2007) Raw material prices(Varma et al., 2008) Length of supply chain(Varma et al., 2008) Physical risks(Varma et al., 2008) Market share(Varma et al., 2008) Information carrying cost(Bigliardi et al., 2010) Supplier cost saving activities(Bigliardi et al., 2010) Variations against budget(Bigliardi et al., 2010) Cost per operation hour(Bigliardi et al., 2010) Return on investment(Bigliardi et al., 2010) Gross revenue (Chia et al., 2009) Profit before tax(Chia et al., 2009) Cost reduction(Chia et al., 2009)

Buyer-supplier collaboration in problem solving(Bigliardi et al., 2010) Employee satisfaction (Chia et al., 2009) Employee turnover per year (Chia et al., 2009) Number of suggestions implemented per employee yearly (Chia et al., 2009)

Money invested in employee training yearly (Chia et al., 2009) ****** to be continued********

****** to be continued******

Table 1. Measures of the supply chain performance evaluation based on BSC approach

2.3 Tools Used in Supply Chain Evaluation Basically, the tools of measurement can be categorized into 2 types. These two categories utilize different tools to evaluate supply chain. In parametric analysis, gap-based techniques are usually used for performance measurement. "SPIDER" and "RADAR" diagram and the "Z" chart are just some examples. These tools are highly graphical in nature which makes them easy to understand. However, they are not useful when analysts need to integrate different elements into one complete picture. Another parametric method which has been used in different areas is the "Ratio Method". This method makes it is easy to compute and calculate the relative efficiency of output versus input. However, different ratios provide different interpretations and it is difficult to integrate the entire set of ratios into a single judgment. Analytical hierarchy process is also used to analyze data in performance measurement. It utilizes personal views of experts to convert various weighted scores into a single score. Though this method provides managerial insights in quantifying measures, it is subjugated to a high degree of subjectivity. Statistical methods such as regression are also useful tools to some extent as they are able to provide meaningful relationships for decision makers, but regression can only analyze one single output at a time. Moreover, by adding another criterion, the approach has to be repeated and it can only consider average values which probably do not occur in the real world. On the other hand, it assumes that all firms have the same performance in combining their input factors. One of the most commonly used tools that can provide a comprehensive framework in performance measurement is balanced scorecard (BSC). Some critical areas that BSC considers are product, process, customer, and market development (Kaplan & Norton, 1992). In addition, it has four main perspectives which are traditional financial indicators, customers, internal business processes, and innovation and learning. BSC can link strategic objectives of a firm to a comprehensive set of measures and this feature distinguishes BSC from other tools. However, it cannot provide any mathematical logic for relationships among different criteria. Hence, utilizing BSC causes some difficulties in comparing the internal and intra performance of the firm. That’s why utilizing BSC to evaluate performances is not good enough and parametric method is proposed to arrive at some judgments, instead. Another non-parametric tool in evaluation is Data Envelopment Analysis (DEA) which considers qualitative as well as quantitative measures, which enables managers to provide reasonable judgment on the efficiency of the resource usage. It uses the concept of efficient frontier which was suggested by Farrel 9

(1957). DEA can also compute efficiency for multiple inputs and outputs by dividing the weighted sum of outputs by a weighted sum of inputs. All the efficiencies would lay between 0 and 1, in which 1 represents the most efficient DMU and 0 shows the inefficient DMU. Although DEA can cover some shortages of the other tools it has some problems too. For example, it is obvious that availability of data in order to make DEA results meaningful is really vital. All inputs and outputs of a DMU are needed to reach reasonable findings. However, sometimes it is difficult to access to some of these data and in some cases firms are not interested in sharing their data. The number of DMUs being compared cannot exceed a certain lower limit. If so, the number of efficient DMUs would increase. DEA assumes that all of DMUs in the firm have the same strategic goals and objectives. Hence, DMUs which are different in these aspects cannot be comparable by DEA and interpretation of the results of DEA is very critical. Although DEA provides reasonable rankings for efficient DMUs, it seldom addresses the reasons of inefficiencies or solutions to reduce or eliminate it. The set of profits and shortages of each tool makes it difficult to select the appropriate one. Nevertheless, DEA has some advantages that distinguish it from other tools.

2.4 Data Envelopment Analysis (DEA) DEA is a linear programming based methodology that can evaluate DMUs qualitatively as well as quantitatively and also calculates multiple inputs and outputs. The term DMU stands for Decision Making Unit and can be used either for comparing different firms or evaluating the efficiency of one firm over time. DEA was first proposed by Charnes, Cooper and Rhodes (CCR) in 1978. In 1984 Banker et al. suggested the evolutionary form of the CCR model named BCC. In subsequent years, DEA received greater attention and a large number of researchers studied it and developed various models to run evaluation performance. In general, these models differ in orientation, disposability, diversification, and returns to scale and types of measures. The underlying concept of measuring efficiency in DEA is efficient frontier function. So a set of efficient and inefficient units have emerged. The analysis of inefficient units has two aspects. First, it can show the maximum input level in order to attain a given amount of outputs. Second, it can show the highest output level attained for a given amount of input. These approaches are called "minimal principle of efficiency" and "maximum principle of efficiency", respectively.

2.5 Motivations to use DEA in Supply Chain A tool which is used for evaluation should carry some characteristics in order to make it useful and suitable. Many researchers believe that simplicity and ease of use should be considered when selecting the tool (Detoro, 1995; Sheridan, 1993). It must also be reliable and output results must be realistic enough to support the process of decision making.

10

DEA is a robust, standardized and transparent methodology that can fulfill all the requirements mentioned above. It also inherits some additional features which make it suitable to be selected as the supply chain efficiency tool some of which are as follows: • Ability in processing multiple elements (Charnes & Cooper, 1978). • There is no need to specify the relationships among the performance measures. • The concept of efficient frontier which is used in DEA serves appropriately as an empirical standard of excellence. • DEA can analyze qualitative measures as well as quantitative measures simultaneously. • In the approach of utilizing DEA, there is no need to assume priority estimates. This feature increases the acceptability of its results. • DEA provides information about inefficient DMUs as well as efficient DMUs. • DEA is highly flexible, and can mold easily to other analytical methods such as statistical analysis and other multi criteria decision making techniques (Golany, 1988; Spronk & Post, 1999; Zhu et al. 2004).

2.6 Balanced Scorecard (BSC) Kaplan and Norton (1996) argued that BSC provides managers with the means they need to navigate future competitive success. It includes more non-financial measures derived specifically from the organization’s strategy. BSC is one of the most comprehensive and simple performance measurement tools that emphasizes both the aspects of the financial and non-financial, long-term and short-term strategies, and emphasizes internal and external business measures. The strongest point of BSC is its ability to illustrate the cause and effect relations between strategies and processes through the four perspectives of: “Financial”; “Customer”; “Internal business process”; and “Learning and growth". Based on this reasoning, to achieve its financial benefits, an organization has to take its customers’ needs and expectations into account, initially. To do this, organizations should take on a process approach when developing and implementing a quality management system. The contents of four perspectives of BSC are described as follows: 1. Financial Perspective A company’s past operating performance containing setting up a financial goal and the implementation of executing strategy achievements can be shown in this perspective as organizations gain growth, return and risk control from operating strategies can all be checked in this perspective. The appraisal indices usually include a return on investment operating income, operating costs, net profit rate, cash flows, etc…. 2. Customer Perspective In order to focus on the customer market segmentation, organizations should use their intrinsic advantages and resources to show their differences in comparison with their competitors since the main

11

measurements contain customer continuation, customer satisfaction, customer acquirement, share ratio, and customer profitability. 3. Internal Process Perspective This perspective refers to the internal organizations’ operating process which should be followed, the operating strategy plans presented, as well as the attempts made to accomplish the customers’ and shareholders’ expectations. The total process is commenced by understanding customer requirements, it is followed by after-sales services as well as the innovation and operating processes, and finally, it ends in customer requirement achievements. 4. Learning and Growth Perspective In order to have a sustainable operation and development, organizations should rely on continual innovation and growth. Further, Kaplan and Norton (1996) pointed out that ‘‘companies should regard some principles such as employees abilities enhancing, information systems performance, encouragement, authority consistence, etc.’’ In other words, this perspective contains three main basic appraisal criteria which are employee satisfaction, employee continuation, and productivity of employees. On the other hand, organizations should establish performance appraisal indices based on these three criteria.

2.7 Decision Making Trial and Evaluation Laboratory (DEMATEL) Decision Making Trial and Evaluation Laboratory (DEMATEL), which was developed by the Science and Human affairs Program of the Battelle Memorial Institute of Geneva between 1972 and 1976, was utilized in the research and solving a group of complicated and intertwined problems. DEMATEL approach can recognize the interactions among alternative systems and evaluation criteria, because it can calculate the impacts among criteria successfully. On the other hand, DEMATEL has the potential to separate a set of composite factors into a dispatcher group and receiver group effectively, and also, conversion into an outstanding structural model. By this method of utilization, we can easily extract the mutual relationships of interdependencies among various criteria and the strength of interdependence (Tamura & Akazawa, 2005) which can be arranged briefly by the following steps: Step 1 determines the relations among the factors defined. By brainstorming or doing a literature review, system element understating and the relations between elements will be judged by professionals subjectively using a questionnaire design: A professional questionnaire is formed by comparing criteria of each element pair which is shown by numbers from 0 to 4, each standing for a level from ‘‘no influence’’ to ‘‘very high influence’’. Step 2 sets up a direct-relation matrix, because by the influential degree between one element and another comparison, an

matrix could be generated. The direct-relation matrix is shown with Z, and figures inside the matrix show the influential extent between the elements. Step 3 computes normalized direct-relation matrix.

12

MarkingS =

1   Max n ≤i ≤1  ∑ nj =1 Z ij   

(1)

Later the elements of direct-relation matrix (Z) by S are multiplied, which leads to standardized directrelation matrix (X) as equation 2: (2)

X= S × Z

Step 4: computes total-relation (direct/indirect) matrix. We used T to show a total-relation matrix and I as a unit matrix, where X as a total relation matrix will be used, also X =  xij  , limk →∞ ( X 2 + ... + X k ) means   n×n that it is an indirect matrix. When 0 ≤ X ij < 1, then lim X = 0 k

(3)

k →∞

(

)

(

T = limk →∞ X + X 2 + ... + X k = limk →∞ X 1 + X + X 2 + ... + X k −1 T = X (1 − X )

)

(4)

−1

(5)

T = tij  , i, j ∈ {1, 2,...., n} Step 5 draws causal diagram and result analysis drawing. The total amount of each row is shown by Di and the total amount of each column is shown by Rj.

Di = ∑ nj =1 tij ( i = 1, 2,...., n ) R j = ∑ nj =1tij ( j = 1, 2,...., n )

(6)

(7)

The causal diagram uses (D + R, D - R) as ordered pairs. The horizontal axis (D + R) shows the influential degrees of relations between elements where vertical axis (D - R) shows the influential relation degrees between one element and the others. Therefore, the sophisticated causality elements themselves could be observed as a simple and clear structure by the causal diagram, and the structure could be referred to as a guide of counsel or a strategy against problems made up by decision makers or managers.

3. Proposed Efficiency Evaluation Framework and Analytical Method According to the analysis of the previous literature review, the efficiency evaluation model of the supply chain proposed by this research is shown in Figure 1. The analytical process is divided and carried out in four stages: (1) in the first step we determined the efficiency measurements of supply chain with review literature and expert ideals; (2) in the second step we divided these measurements into the four perspectives of the BSC approach; (3) The DEMATEL method was applied to determine causal relationships and mutual influence among perspectives; (4) an empirical analysis of synthetic performance evaluation of the supply chain was made throughout Network DEA to grade the order among the organizations. The analytical methods, BSC, DEMATEL, and Network DEA employed by this research are introduced in brief as follows:

13

Determining the efficiency metrics of supply chain

BSC Dividing the efficiency measurements of supply chain into the four perspectives of BSC DEMATEL Determining the relationships between the four

perspectives of BSC DEA

Creating the Network DEA model focus on the relationships between the four perspectives of BSC, and raking the supply chain

Fig 1. Proposed performance evaluation model of food industry

Since BSC is based on causal relationships, DEMATEL was used to determine these relationships in the next stage. These relationships organize a network structure. Therefore, the network DEA model was created to determine the efficiency of the supply chain. The supply chain was then ranked with this network DEA method using the BSC approach.

4. EMPERICAL RESULTS In accordance with the proposed performance evaluation model shown in Figure 1, this study conducted an empirical analysis based on 22 supply chains of Iranian food industries. 4.1 Determining the Efficiency Metrics of Supply Chain 4.1.1

Structure of Supply Chain

On the basis of the above mentioned considerations, the present study aimed at developing a network DEA based on the well-known balanced scorecard (BSC) framework suitable for implementation in the food supply chain. The reason to choose this context was that the supply chain of food products has received a great deal of attention in the past decade due to issues related to the public health. It is apparent that, in the near future, the design and operation of food supply chains will be subject to more stringent regulations and closer monitoring, in particular, those products destined for human consumption. This implies that the traditional supply chain practices and the corresponding performance measurement should be subject to revision and change (Ahumada & Villalobos, 2009). The perishable nature of foods makes food supply chain performance evaluation issues a complex multiple criteria decision problem. Thus, food supply chain performance evaluation problems with such features are difficult to adjust for common performance evaluation methods. Also, the tremendous importance of quality and the safety of food products that must be provided to customers have revealed the need to determine and select the performance evaluation criteria proportionate to such supply chain. Several studies have been conducted in supply chain performance 14

evaluation literature, but only a few studies have evaluated the performance of food supply chain. In this paper, the efficiency of 22 supply chains has been evaluated and a new network DEA model with the BSC approach for supply chain efficiency evaluation has been proposed. In order to develop a model to assess the performance of the food supply chain, the following conceptual framework and relationships among supply chain members can be considered (see Figure 2).

Animal husbandry

Supplier of animal food

Supermarkets

Refrigerating Room

Factory

Wholesaler

Final customer

Supermarket

Supplier of others goods

Fig 2. Conceptual framework of food supply chain

The above food supply chain framework can be divided into four levels, namely, suppliers, producer/manufacturer, retailer/distributer, and end customers. The food supply chain and its relations are depicted in the following Figure.

Supplier

Retailer

Final Customer

Manufacturer

Supplier

Retailer

Fig 3. Food supply chain network

4.1.2

Efficiency Metrics of Supply Chain

By reviewing the related backgrounds of efficiency indices on internal and external logistic and supply chain management measures, several efficient indices about the food industry have been acquired. Some of these efficient indices are shown in Table 2 below.

15

Financial Metrics

Efficiency Metrics

Flexibility Metrics

*Profit *Amount of sale *Staff cost *Overhead cost *Transportation cost *Inventory cost *Spoilable products *Maintenance cost *Return on investment *Amount of export

*Amount of the water consumption *Amount of energy consumption *Inventory *The number of failures *Days of storage *Lost sales *Use of wastage *Amount of production *Production time *Amount of fuel consumption *Amount of return on investment *Variety/time *Rate of order completion *Cycle efficiency *Distribution program *New technology *Production time for new good

*Packing *Delivery time *Response to customer *Emergency delivery *Delivery place *Delivery alternate *Amount of production *Combination of production *Operation process *Production formula

Quality and Safety Metrics

Customer and Market Metrics

*Return on production *Condition of storage *The test of the raw material *Transportation conditions *External speculations’ *Cupboard life *Temperature of place *Quality of packing *The test of final production *Environment Hygiene *Possibility of follow up *Hygiene License *Packing *Chemical material *Nutrition

*On time delivery *Number of agency *Number of complaints *Share market *Satisfaction of customer *Constance of customer *brand *Customer response time *Attendance in fairs

Table 2. Some efficient indices of food supply chain

By using the metrics obtained from the previous section, and via making further inquiries, key indices were recognized. In the initial list, fifteen criteria were determined. These indices are shown in Table 3 below. Criteria Cost saving by supplier initiatives

Criteria

Criteria

Cost of information sharing

Quality of Production process

Average of suggestions implemented per

Volume of qualified delivery

Volume of new goods sold

employee yearly

goods per year

Customer response time

Employee Satisfaction

Return on investment

Total inventory cost

Gross revenue

Production capacity utilization

Average number of on time delivery per Learning cost

year

Profit before tax

Table 3. Initial list of food supply chain indexes

Then, in order to select the ultimate criteria, a food supplier selection criteria questionnaire containing three main parts was prepared, validated, and distributed among 400 food experts as respondents. Since the food supply chain framework is composed of four members, as shown in Figure 3, it was tried to distribute questionnaires among the members according to their order of importance. Table 4 below indicates the questionnaire’s distribution process. Supply Chain Members

Submissions

Returned

Percent of submissions

Percent of returned

Supplier

120

53

30%

0.44

Manufacturer

220

145

55%

0.66

Retailer

40

40

10%

1.00

Final Customer

20

20

5%

1.00

Total

16

400 250 1 Table 4. Questionnaires distribution

0.63

Among the 400 questionnaires distributed to respondents, only 258 questionnaires were returned completely to the researchers and out of the 258 returning questionnaires, 250 questionnaires were accepted and analyzed, and a 63% satisfactory response rate was yielded.

The first part in the questionnaire was YES/NO selection, which asked the respondent to determine the criteria that had been taken into account by the food supply chain managers when assessing performance. The results of the first surveys showed that the most important criterion selected by experts was Return on Investment and the least important was Employee Satisfaction by 97.24 and 34.78 percentages of expert’s votes, respectively (See Figure 4 below).

97.24 90.67 87.89 78.74

Return on Investment Gross Revenue Profit Before Tax Volume of Qualified Delivery Goods Per… Quality of Production Process Average Number of On Time Delivery Per… Total Inventory Cost Employee Satisfaction Average of Suggestions Implemented Per… Cost of Information Sharing Learning Cost Production Capacity Utilization Customer Response Time Volume of New Goods Sold Cost Saving by Supplier Initiatives

48.65 79.56 86.38 34.78 69.34 57.45 69.75 54.32 71.43 84.17 70.32

0

20

40

60

80

100

Fig 4. Percentage of experts vote of food supply chain indexes

According to the experts' votes, the criteria with the significance value lower than 50 percent can be eliminated. Thus, the selection criteria can be rated as Table 5 shows:

Criteria

Criteria

Criteria

Cost saving by supplier initiatives

Cost of information sharing

Return on investment

Volume of new goods sold

Average of suggestions implemented per employee yearly

Gross revenue

Customer response time

Total inventory cost

Profit before tax

Production capacity utilization Learning cost

Average number of on time delivery per year Volume of qualified delivery goods per year

Table 5. Measures of food supply chain performance evaluation

The second part of the questionnaire asked the respondents to mark the importance of criteria using 7-

point Likert scale ranging from 1, extremely unimportant, to 7, extremely important, respectively. The average and standard deviation values and ranking of selection criteria are listed in Table 6 below:

Important Indexes

Average

Standard Deviation

Ranking

Return on Investment

6.272

1.09

1

Gross Revenue

6.202

1.122

2

Profit before Tax

6.108

1.156

3

Total Inventory Cost

5.898

1.223

4

Volume of New Goods Sold

5.532

1.241

5

Volume of Qualified Delivery Goods Per Year

5.204

1.609

6

Average Number of On Time Delivery Per Year

5.198

1.358

7

Customer Response Time

4.986

1.398

8

Cost Saving by Supplier Initiatives

4.957

1.661

9

Production Capacity Utilization

4.934

1.458

10

Learning Cos

4.837

1.486

11

4.696 1.691 Average of Suggestions Implemented Per Employee Yearly 4.543 1.603 Cost of Information Sharing Table 6. Ranking the food supply chain indexes

12 13

In the previous parts of the questionnaire, the main criteria, significance values of selection criteria, and their ranking in food supply chain performance evaluation were determined. The main contribution of this paper is to incorporate a performance evaluation model with balance scorecard approach. Therefore, it is necessary to assign selection criteria into the four perspectives of BSC approach. As discussed at the beginning of the current section, both financial and non-financial criteria were considered in the initial selection criteria. So, in the third part of the questionnaire the food experts were asked to determine which criterion belongs to which perspective of BSC, or, to assign criteria into the four perspectives of BSC, namely, financial, customers, internal processes, and learning and growth. The stratified criteria in four BSC perspectives are shown in Table 7 below. The financial perspective measures (F) F1: Return on investment (Rial) (96.56%, 93.45%, 95.67%) F2: Gross revenue(Rial) (97.34%, 93.45%, 96.45%) F3: Profit before tax(Rial) (98.34, 93.45, 96.45%) The internal process perspective measures (P) P1: Production capacity utilization (65.34%, 85.67%, 60.45%) P2: Average number of on time delivery per year (91.34%, 93.26%, 98.56%) P3: Total inventory cost (Rial) (61.45%, 69.34%, 62.45%)

The customer perspective measures (C) C1: Customer response time (89.35, 93.26, 93.45%) C2: Volume of new goods sold (74.45%, 85.76%, 68.45%) C3: Volume of qualified delivery goods per year (65.56%, 79.65%, 60.45%) The learning and growth perspective measures (L) L1: Average of suggestions implemented per employee yearly (89.75%, 94.25%, 78.50%) L2: Cost of information sharing (Rial) (75.45%, 73.20%, 63.25%) L3: Cost saving by supplier initiatives (Rial) (79.87%, 92.25%, 65.25%) L4: Learning cost (Rial) (96.25%, 93.45%, 74.45%)

Table 7. Measures of food supply chain performance evaluation based on BSC approach

18

In order to accurately analyze the discriminating and assigning criteria into the financial, customers, internal processes, and learning and growth perspectives, the percentage of member vote of food supply chain are given in the Table 7. These percentages represent the views of suppliers, manufacturer, and retailer respectively. 4.2 Determining the Relationships Between the Four Perspectives of BSC by Means DEMATEL In this phase, relationships among BSC perspectives are determined by employing the DEMATEL technique. Firstly, 20 checklists were distributed among specialists and food industry managers and they were asked to note the effect of each of the four BSC perspectives on other BSC perspectives. To recognize the relationships among BSC perspectives, it is necessary to take some steps as follows: Step 1: Define elements and determine relations. Specialists and the food industry managers’ ideas are gathered in this stage and proficient indices of each of the BSC perspectives are determined as illustrated in Tables 8, 9, 10 and 11. Then, DEMATEL is used to develop a total-relation matrix of the four evaluations. Step 2: Establish a direct relation matrix X. Ideas gathered from step 1 are abridged in a (4×4) matrix, that is to say, the Z (4×4) matrix as shown in Table 8 below. The numbers inside the matrix demonstrate the influential degrees between one perspective and the others. Financial(F) Customers(C) Learning and growth(L) Internal Process(P)

Financial(F) Customers(C) Learning and growth(L) 0.0000 3.1448 3.2215 2.9815 0.0000 3.4986 2.2596 3.2415 0.0000 3.2512 2.4785 2.9445 Table 8. Direct relation matrix X between perspectives

Internal Process(P) 2.2579 3.1025 3.1248 0.0000

Step 3: Calculate direct normalized relation matrix. By means of equation (2) normalized direct matrix is created as in Table 9. Financial(F) Customers(C) Learning and growth(L) Internal Process(P)

Financial(F) Customers(C) Learning and growth(L) 0.0000 0.3282 0.3362 0.3111 0.0000 0.3651 0.2358 0.3383 0.0000 0.3393 0.2586 0.3073 Table 9. Normalized direct relation matrix X

Internal Process(P) 0.2356 0.3238 0.3261 0.0000

Step 4: Compute total-relation (direct/indirect) matrix. By introducing normalized matrix T and utilizing equations (3), (4) and (5), total-relation matrix T is obtained as presented in Table 10. Financial(F) Customers(C) Learning and growth(L) Internal Process(P)

Financial(F) Customers(C) Learning and growth(L) 2.7961 3.1598 3.3554 3.2662 3.1480 3.6232 2.9934 3.1649 3.1051 3.0516 3.1174 3.3369 Table 10. Total-relation (direct/indirect) matrix T

Internal Process(P) 3.0117 3.2932 3.0688 2.8166

Note: Numbers in bold are the perspectives which reach the threshold (3.1327) In order to understand the influential relation between the evaluation perspectives, the median (3.1327) is put as the threshold in this research. Should the value reach or exceed the threshold, the perspective is then considered to be more influential than the others. These values are shown in bold in Table 10. The construction of a networked level framework is based on the total-relation matrix T. By using equations (6) 19

and (7) as shown in Table 11, the total amount of each row is presented by Di and the total amount of each column is presented by Rj. The horizontal axis (Di + Rj) presents the influential degrees of relations between elements, but vertical axis (Di - Rj) represents the influential degrees of relations between one element and the other elements in the proposed BSC framework as illustrated in Figure 5. Concerning Table 11, it is observed that the greatest influential value (0.7405) is selected from one of the perspectives toward “Customer’’ that has the highest value of (Di - Rj). RANK

Y-axis  - 

X-axis  + 

RANK

2

0.2157

24.4303

4

Financial(F)

1

0.7405

25.9207

1

Customers(C)

4 3

Learning and growth(L) Internal 3 0.1322 24.5128 Process(P) Table 11. Total-relation matrix with (Di + Ri) and (Di-Ri) -1.0884

25.7528

2

Step 5: Draw a causal diagram. As shown in Table 11, (D + R) is X-axis and (D - R) is Y-axis. Drawing a causal diagram as in Figure 5 makes it is clear that customer and learning and growth perspectives are the main and trivial perspectives, respectively. Based on the above cause & effect diagram, the relation among the four BSC perspectives is as shown in Figure 6. 1.00

DEMATEL

0.50

C

F

0.00 24.00 -0.50

P 24.50

25.00

-1.00

25.50

26.00

L

-1.50 Fig 5. Total-relation matrix with (Di + Ri) and (Di-Ri)

These relationships organize a network structure. This structure is shown in Figure 6. F1 F2 F3 Financial (F)

Customer (C)

C1 C2 C3

Learning and growth (L)

Internal process (P)

Fig 6. Networked evaluation structure

P1 P2 P3

L1 L2 L3 L4

4.3 Network DEA Model for Supply Chain Evaluation In this section an integrated approach to evaluate supply chain has been presented. In the previous sections some methods of evaluating supply chain were explained and it was highlighted that these methods suffer some deficiencies, so using the balanced scorecard approach was proposed. Employing this approach, the researchers studied non-financial as well as financial measures, but the efficiency for each individual supply chain cannot be computed and the analysis of the results is difficult by using BSC. Since the efficiency evaluation score could not be determined by applying BSC, the data envelopment model (DEA) was used to calculate the efficiency score of the supply chain. As DEA has proved to be a helpful tool to deal with the analysis results, it has been combined with BSC (Chen & Yu, 1997). Richard (2003) argued that DEA is suitable for measuring the best practice of the BSC indicator. Eilat et al. (2007) presented a new integrated BSC and DEA model to evaluate the R&D projects. They demonstrated a multi-criteria approach and developed the DEA model. In most cases, the integrated DEA-BSC model addresses three common goals (Baker et al., 1975; Eliat et al. 2007): effectiveness goal, efficiency goal, and balance goal. Based on Figure 6 the relation among the four BSC perspectives is shown in Figure 7 as a network. This network structure is determined by the DEMATEL approach in the previous section. The relationships between the four perspectives of BSC were redesigned based on the DEMATEL approach in Figure 7. It is argued that the basic perspective of the BSC approach is learning and growth (Kaplan & Norton, 1996). This idea was used to design the network as shown in Figure 7. below.

L

P

C

F

Fig 7.Network structure of BSC approach in Iranian food supply chain

In this case, there are four stages and each stage uses its own inputs to produce its own outputs, and there are links between some stages. For example, consider C perspective, link L-C, F-C and C-C indicate that the learning and growth, financial and customer perspectives metrics influence the customer perspective metrics respectively. Therefore, it is possible to say that the learning and growth, financial and customer perspectives metrics are inputs in the Customer stage, and link C-L, C-C, C-F and C-P indicate that the customer perspective metrics influence the learning and growth, customer, financial and internal process perspective metrics. It is also possible to say that the learning and growth, customer, financial and internal process perspective metrics are outputs in the Customer stage. Likewise, the customer perspective metrics influence the customer perspective metrics. On the other hand, the customer perspective metrics are input and output at the same time in this stage; hence, it could be deciphered that customers attract customers. Other relationships between the four perspectives of the BSC approach can be explained with the same method. Inputs and outputs of each stage are shown in Table 12 below: 21

Indexes

Stage 1 inputs

Stage 2

outputs

inputs

L Indexes P Indexes

√

C Indexes

√

F Indexes

√

Stage 3

Stage 4

outputs

inputs

outputs

√

√

√

inputs

outputs √

√ √

√

√

√

√

√

√

√

Table 12. Inputs and outputs of four perspectives of BSC

It must be mentioned that this network structure has no independent inputs and outputs in each stage and it has a cycle structure. Another property is that the network structure obtained by DEMATEL has several returnable relations. These distinguish the network structure presented here form other models presented in the literature. Considering the aforementioned properties, one can come to know that the network structure presented here differs from the ones presented in previous studies. Considering the specific features mentioned above, one can find that the model presented in this paper differs from the previous ones. The present section tries to determine a network DEA model for the efficiency evaluation of this network supply chain structure. Classic DEA models consist of a family with different assumption on inputs and outputs, but every activity should belong to one specific stage and there are no links between stages. Some researchers introduced network structure and modified DEA models to evaluate network structures (Cook et al., 2010; Kao & Hwang, 2008; Tone & Tsutsui, 2009). In the present study, all inputs and outputs for each supply chain were set in one of the four aspects of BSC. In fact, the BSC structure was embedded in DEA models and as a result a powerful measurement tool was developed for practical applications. Now, let’s assume that we have a set of DMUs (supply chains) consisting of DMUj, j=1,…, n, for evaluating the performance of DMUj.

z rjpq is the r-th component (r=1…Spq) of Spq -dimensional for DMUj flowing from stage p and entering to stage q.

u rpq is a multiplier for z rjpq when z rjpq is as the output of stage p.

v rqp is a multiplier for z rjqp when z rjqp is as the input of stage p. Thus, the efficiency ratio for stage p from DMUj is explained as:

θ pj =

∑ ∑ ∑ ∑

urpq zrjpq

q

r

q

qp qp r r rj

(8)

v z

We define the overall efficiency of the network as a convex combination of P stage. Therefore, we have:

θ j = ∑ pj =1 wpθ pj = 1

22

and

∑

p j =1

wp = 1

(9)

It is evident that choosing weights (Wp) is very important for evaluating the performance of the network. So a suitable choice for Wp is the proportion of the input used at the p-th stage to the input used in the network (Cook et al., 2010).

∑ ∑u z ∑ ∑ ∑v

wp =

p

r

p

pq r

q

pq rj qp qp r rj

r

z

p = 1.....k

(10)

j = 1.....n

(11)

Thus, the overall efficiency can be rewired as in the form:

θj =

∑ ∑ ∑u ∑ ∑ ∑v p

q

p

r

q

r

pq r qp r

z rjpq z

qp rj

For computing the performance of DMU0 in the best condition we use the following model:

Max θ 0

s.t

θ j ≤ 1 ∀j

(12)

θ pj ≤ 1 ∀j∀p u

≥ 0, vrqp ≥ 0∀r∀p∀q

pq r

It is evident that θ j ≤ 1 is redundant constraint, therefore we rewrite model (12) as follows:

Max θ 0

s.t

(13)

θ pj ≤ 1 ∀j∀p urpq ≥ 0, vrqp ≥ 0∀r∀p∀q By substituting (8) and (10) in model (13), we have the following model:

Max

∑ ∑ ∑ p

q

pq pq r r ro

u z

S .t

∑ ∑ ∑v ∑ ∑ u z −∑ ∑ p

r

pq r

q

u

qp qp r r ro

q

pq rj

p

z

q

qp r

=1

v z

qp rj

(14)

≤ 0∀j ∀p

≥ 0, v ≥ 0∀r∀p∀q

pq r

qp r

Theorem1:if DMU0 is the unit under evaluation, then there is an optimal solution, say ( v* , u* ) . For index like L we have:

∑ ∑ ∑ p

q

* pq pq r r rl

u

z

− ∑ p ∑ q ∑ r vr*qp zrlqp = 0

Proof: we write model (14) as follows:

Max

∑ ∑ ∑ p

q

pq pq r r ro

u z

S .t

∑ ∑ ∑ v z =1 ∑ ∑ ∑ u z −∑ ∑ ∑ v p

p

q

r

pq r

q

pq rj

qp r

r

p

(15)

qp ro

q

r

qp r

z rjqp ≤ 0∀j

urpq ≥ 0, vrqp ≥ 0∀r∀p∀q It is evident that the feasible region in (15) is a subset of feasible in the region in (14), but the value of the objective function in both of them is equal. Therefore, the optimal solution of model (15) is the optimal solution of model (14). 23

Let ( v , u ) be the optimal solution for model (15), we know that if an optimal solution exists, then an optimal extreme point exists as well. Let ( v* , u* ) be the optimal extreme solution for model (15), so it lies on k = 2∑ q ∑ p s pq linearly independent hyperplans. At first, we prove that all variables are not binding in the *qp

optimal solution. Every v r

cannot be equal to zero, because in this case we would have:

∑∑∑v p

q

*qp r

z roqp = 0 ≠ 1

r

And every ur* pq cannot be equal to zero, because in this case the value of the objective function is equal to 0. So at most, the (k-2) of variables in optimality are binding. Moreover,

∑ ∑ ∑ p

q

*qp qp r r r0

v

z

= 1 is binding at every feasible solution, so this constraint is tight at

optimal solution. Consequently, a least one of the constraints in optimality should be binding. So

∑ ∑ ∑ p

q

* pq pq r r r1

u

z

∑ ∑ ∑ p

q

r

u r* pq z rjpq − ∑

p

∑ ∑ q

v r*qp z rjqp ≤ 0

r

− ∑ p ∑ q ∑ r vr*qp zrqp1 = 0 and since ( v* , u* ) is

the optimal solution for model (15) it is an optimal solution for (14) too. Therefore, the proof is complete. Theorem 2: if DMU1 ,...., DMU n be DMUs which are under evaluation, then there exists at least one relative efficient DMU. Proof: let ( v* , u* ) be the optimal solution for model (14) in evaluating DMU0 and by using theorem (1) we know that

∑ ∑ ∑ p

q

* pq pq r r r1

u

z

− ∑ p ∑ q ∑ r vr*qp zrqp1 = 0 . The efficiency of DMUl is gained by

solving the following model:

Max ∑∑∑ u rpq z rlpq p

q

r

S .t

∑∑∑ν z = 1 ∑∑∑u z − ∑∑∑ν qp r

p pq r

p

q

q r pq rj

r

qp r

p

(16)

qp rl

q

z rjqp ≤ 0∀ j

r

u rpq ≥ 0,ν rqp ≥ 0∀r If

∑ ∑ ∑ p

q

* qp qp r r r1

v

z

= 1 , the proof is evident. Otherwise

∑∑∑ν p

q

*qp r

z rlqp = α > 0

r

So we have:

v r*qp

∑∑∑ ( p

q

r

α

)z rlqp = 1 and ∑∑∑ ( p

q

r

u r* pq

α

v *qp )z rjpq − ∑∑∑ ( r )z rjqp ≤ 0∀ j p

q

r

α

* * These relations imply that  v , u  is a feasible solution for model (15) and since the optimal solution  

∝ ∝

for model (15) is the optimal solution in model (14); the objective value is equal to 1 for model (14) and (15). So DMUl is efficient. 24

4.4 Results In order to appraise the efficiency of food industries supply chains in the Iranian market, a network DEA model was applied as proposed in this paper. So in this section the overall performance was calculated directly by applying this model. This performance is obtained from combining the BSC approach and the network DEA model based on the above definition of efficiency. This paper evaluated the performance of 22 Iranian food supply chains via a network DEA model evaluation process which separates the four perspectives of BSC. Determining inputs and outputs to calculate efficiency is a significant step in efficiency evaluation. Since the aim of the present study was to evaluate the efficiency of supply chain, four sets of inputs and outputs which were shown in the previous section were necessary. On the other hand, the supply chain metrics were determined by expert ideas in section 4.1 and by applying the DEMATEL approach, the relationships among the four perspectives of BSC were determined. The choice of inputs and outputs was influenced by the literature on the DEA application in the supply chain efficiency measurement. The inputs and outputs used in earlier DEA application studies were listed in section 2. Based on the literature of DEA application in the supply chain, only the inputs and outputs in the supply chain efficiency prevalent in the literature of supply chain efficiency were listed. Then the most important indexes of efficiency in Iranian food industries were selected based on the expert view and the food industry’s staff ideals. In fact, four sets of indexes were determined. Based on the proposal model in this paper and section 4.3, the network of supply chain can be divided into four stages based on BSC approach. The first stage is learning and growth in which three categories (internal process, customer and financial perspectives) are considered as inputs and one category (customer perspectives) as the output. Hence, the inputs are production capacity utilization, average number of on time delivery per year, total inventory cost (Rial), customer response time, volume of new goods sold, volume of qualified delivery goods per year, return on investment (Rial), gross revenue (Rial), and profit before tax (Rial). The outputs of this stage are customer response time, volume of new goods sold and volume of qualified delivery goods per year. The next stage is internal process perspective in which one category (customer perspective) is considered as the input and one category (learning and growth perspective) as the output. The inputs are customer response time, volume of new goods sold, and volume of qualified delivery goods per year, while the outputs of the second stage are average of suggestions implemented per employee yearly, cost of information sharing (Rial), cost saving by supplier initiatives (Rial), and learning cost (Rial). The third stage is customer perspective. As shown in Fig.7, this stage has three categories (learning and growth, customer and financial perspectives) as inputs and four categories (learning and growth, internal process, customer and financial perspectives) as outputs. Hence, in this stage, the inputs are average of suggestions implemented per employee yearly, cost of information sharing (Rial), cost saving by supplier initiatives (Rial), learning cost (Rial), customer response time, volume of new goods sold, volume of qualified delivery goods per year, return on investment (Rial), gross revenue (Rial), and profit before tax (Rial). And finally, the outputs of this stage are average of suggestions implemented per employee yearly, 25

cost of information sharing (Rial), cost saving by supplier initiatives (Rial), learning cost (Rial), production capacity utilization, average number of on time delivery per year, total inventory cost (Rial), customer response time, volume of new goods sold, volume of qualified delivery goods per year, return on investment (Rial), gross revenue (Rial), and profit before tax (Rial). Eventually, the final stage is the financial stage in which one category (customer perspective) is considered as the input and two categories (learning and growth and customer perspective) as the output. Therefore, the inputs of this stage are customer response time, volume of new goods sold, and volume of qualified delivery goods per year. Finally, the outputs of this stage are customer response time, volume of new goods sold, volume of qualified delivery goods per year, average of suggestions implemented per employee yearly, cost of information sharing (Rial), cost saving by supplier initiatives (Rial), and learning cost (Rial). In this study, the inputs and outputs data of April 2010 were used. In this phase, the proposal model for performance evaluation in the Iranian food supply chains was applied. As previously mentioned, this method calculates the overall efficiency directly. The results of this model are shown in Table 13. As shown in Table 13, DMU 1, 3, 5, 8, 12, 17, 21 and 22 are efficient, and the other DMUs are non efficient. For example, consider DMU 6 and DMU 7. These two DMUs are the least efficient. Although they are efficient in stage P, C and F, and have good activity in these stages, in stage L they are extremely weak. Results show that these units have been unaware of the learning and growth. The results of each stage also confirmed this point. As shown in Table 13, the efficiency score of stage L compared with other stages is lower. This shows that managers have less emphasis on learning and growth and that the managers' focus is on the customer. The DMUs are efficient on the customer perspective. Take in to DMU 2, 6, 7, 9, 11. Just as you observe, these DMUs are efficient in three stages (internal process, customer and financial) and stage L is non efficient. At first, this issue may not be important but this situation is not suitable for these chains. This issue presents relatively poor in performance of managers in this stage even this issue is an alarm for heads of these DMUs. Doubtless, learning and growth are important factors for permanence of the system and nonchalance of managers in this issue causes to establish difficulties for the system in the long term. Surly, poor performance in learning and growth perspective leads to decrease the efficiency of other stages to come. The manager of chain DMU 4 pays a special attention to the customer perspective and ignores other stages. Thus, stage C is efficient but other perspectives are inefficient, so poor performance of supply chain in terms of learning, internal process, and financial stages is a warning for the heads of DMUs. Unless they change the learning and economic policies, the performance of supply chain in terms of customer will be deteriorated in the near future. DMU15 is efficient in stage L, C and F but inefficient in stage P. To improve and develop efficiency of this DMU, managers must focus on the internal process perspective.

26

DMU(1)

Overall Efficiency 1

Stage L 1

Stage P 1

Stage C 1

Stage F 1

DMU(2)

0.929

0.73

1

1

1

DMUs

DMU(12)

Overall Efficiency 1

Stage L 1

Stage P 1

Stage C 1

Stage F 1

DMU(13)

0.926

1

0.96

1

0.86

DMUs

DMU(3)

1

1

1

1

0.99

DMU(14)

0.851

1

0.71

1

0.85

DMU(4)

0.915

0.71

0.92

1

0.93

DMU(15)

0.888

1

0.7

1

1

DMU(5)

1

1

1

1

1

DMU(16)

0.94

0.89

0.91

1

0.9

DMU(6)

0.584

0.11

1

1

1

DMU(17)

1

1

1

1

1

DMU(7)

0.671

0.14

1

1

1

DMU(18)

0.872

1

0.68

1

0.69

DMU(8)

1

1

1

1

1

DMU(19)

0.863

1

0.7

1

0.86

DMU(9)

0.951

0.95

1

1

1

DMU(20)

0.986

0.8

0.73

1

1

DMU(10)

0.984

1

0.98

1

1

DMU(21)

1

1

0.97

1

1

DMU(11)

0.948

0.31

1

1

1

DMU(22)

1

1

1

1

1

Table 13. Results of evaluation of Iranian supply chains efficiency

Average of Overall Efficiency 92.30%

Average of Stage L 84.73%

Average of Stage P 92.09%

Average of Stage C 100%

Table 14. Average of supply chain efficiency

27

Average of Stage F 95.81%

As Table 14 represents, the result reveals that learning and growth perspective requires special attention with an average score of 84.73%, because the average score of this stage is lower than the other stages. In addition, the best average efficiency score is customer perspective with 100% efficiency. Of course, this average score is not unexpected. Due to the nature of food industry, the full attention to the customers is highly essential. In this study, all of the DMUs have paid full attention to the customer. The finding is supported by the real food industry context in Iran. As Table 14 shows, the BSC perspectives rankings are customer, financial, internal process, and learning and growth perspective, in order. As Table 13 represents, the DMUs 1, 5, 8, 12, 17 and 22 are efficient in each stage. Hence, the overall efficiency obtained from the proposal model for these DMUs are 100%. This point can illustrate the validity of the proposal model. The DMUs 2, 3, 6, 7, 9, 10, 11, 15 and 21 enjoy three efficient stages. As Table 13 shows other DMUs have several inefficient stages. Therefore, they are not efficient decision-making units. The proposal model also shows this reality. Although DMU 6 and 7 have three efficient stages, they have one stage with low efficiency score, 11% and 14% for learning and growth perspective respectively. The proposal model calculated a low overall efficiency score for these DMUs. This issue shows the validity of the proposal model. These DMUs are recommended to focus on learning and growth perspective indexes. Paying more attention to learning and growth input, these DMUs can increase efficiency score on this stage and consequently can increase overall efficiency. Based on the survey and the evaluation done by employing DEMATEL shown in Figure 5, it could be found that the learning and growth perspective has the lowest position in expert views. The result shows this point as well. Therefore, the managers must focus on this perceptive to improve this stage. Through improving this stage one can improve overall efficiency in decision-making units. Due the fall of learning and growth, the DMU 6 and 7 have very low score on this stage and they have the lowest overall efficiency scores of 58.4% and 67.1%, respectively. As Figure 5 and Table 11 present, it is observed that the greatest influential value (0.7405) is selected from one of the perspectives toward “customers” that has the highest value of (Di-Rj). Therefore, it is worth proposing that the most important perspective is the “customers”. Hence, the decision-making units have paid full attention to this stage. As Table 13 shows, all DMUs are efficient on this stage and average efficiency score on this stage is 100%. Following the customer perspective, the financial perspective is important, hence, the number of efficient DMUs in this stage is 15 units. This number takes the second position following the number of efficient DMUs on the customer perspective. This point is consistent with the results of DEMATEL method. If the ranking is done based on average efficiency score in each stage, learning and growth, internal process, financial, and customer perspective come in order. This is in li9ne with the DEMATEL results. This point shows the accuracy of the results of this paper.

28

DMUs DMU1 DMU2 DMU3

Benchmarks of stage1 (learning and growth)

Benchmarks of stage2 (Internal process )

Benchmarks of stage4 (Financial)

(2), (5), (6), (7), (9)

(2), (4), (5), (7), (8), (9), (13), (14), (21), (22) (2), (5), (6), (7), (9), (11)

(1),(8),(10),(12),(19)

DMU4 DMU5 DMU6 DMU7 DMU8 DMU9 DMU10 DMU11 DMU12 DMU13 DMU14 DMU15 DMU16 DMU17 DMU18 DMU19

(13), (14), (15), (19)

DMU20 DMU21

(10), (15), (17), (19)

(10), (15), (17), (19) (8), (13), (17), (19) (10), (12) (1), (6), (9), (17) (12), (13), (15)

(12), (13) ,(15)

(2), (7), (8), (21), (22) (2), (6), (7), (21) (5), (6), (12) (5), (6), (11), (12) (6), (12), (17) (2), (3), (4), (5), (8), (11), (12), (13), (14), (21) (5), (6), (11), (12), (16), (17) (2), (3), (7), (8), (11), (13), (17), (22)

(2), (7), (8), (22) (5), (7), (9), (22) (5), (6), (7), (9), (12) (12), (13), (17), (21) (2), (6), (7), (8)

DMU22 Table 15. Benchmarks for each stage

Table 13 presents the overall performance of supply chain and efficiency score of them in terms of four perspectives and divides DMUs into two groups: efficient and inefficient units. But this information is not enough for the manager of inefficient chains to improve performance of supply chain in different perspectives and finally improve overall efficiency score of the supply chain. Therefore, benchmarks for each inefficient stage have been given in Table 15. It should be noted that this table presents benchmarks for inefficient stages and based on Table 13, all DMUs are efficient in terms of customer perspective (stage 3) thus any discussion for benchmark is plausible for this stage. Table 15 shows that DMU 2 is inefficient in stage1 and based on the given information in column 2, DMU1, DMU8, DMU10, DMU12 and DMU19 are benchmarks for DMU2 (supply chain 2) in terms of learning (stage1), thus the manager of this supply chain can follow learning policy of these DMUs to improve performance supply chain in learning perspective; however, this DMU is efficient in other stages and it can be a benchmark for other DMUs. DMU2 is a benchmark for DMU4, DMU13, DMU14 and DMU 19 in stage 2 (internal process) and DMU4, DMU13 and DMU19 have the tendency to follow the economic policy of DMU 2 because DMU 2 is a benchmark for these DMUs in the financial stage. As Table 15 represents, in learning perspective, DMU3, DMU5, DMU18, DMU 28 and DMU 22 are not benchmarks for any supply chains but DMU 15 and DMU 17 are benchmarks for 5 supply chains, therefore, these supply chains are reliable references for learning and growth perceptive for other DMUs.

29

5. Conclusion One of the important points in evaluation of performance of organization is identifying of weak points in subunits and considering internal relations in the system. Management's target is establishing a balance between different divisions; hence at the first glance it seems that the BSC technique is suitable for establishment of a balance in a system but in this approach, relations between different perspectives are embedded to the model as a presupposition. On the other hand, BSC method can not specify efficient or inefficient DMUs. Several studies have been recorded on the combination of BSC and DEA model in the literature of DEA. Never the less, most of these models suffer deficiencies such as the inability to determine the relative efficiency of DMUs. In the present paper a general framework is proposed to evaluate the overall performance of the supply chain by means of the BSC and DEA model. In the first step, the efficiency measurements of the supply chain were determined via doing a thorough review of the related literature and relying on the scrutiny of the expert ideals. In the second step, these measurements were divided into the four perspectives of the BSC approach; The DEMATEL method was applied to determine the causal relationships and mutual influences among the perspectives; in the next step, the proposal model for performance evaluation was applied to Iranian food supply chains in April 2010. In this case study, a network structure obtained from the BSC, by applying the DEMATEL methods, was developed. By using this network structure, a network DEA was created based on all the relationships present among the four perspectives of BSC, and returnable relationships in particular. The results showed that DMU 1, 3, 5, 8, 12, 13, 17, 21 and 22 were efficient, and the other DMUs were inefficient. As it was shown in this paper, the efficiency score of stage L is lower than other stages and in stage C, all DMUs are efficient. This shows that managers emphasize learning and growth less and that the managers' focus is on customers. This issue presents relatively poor performance of the managers in this stage and is considered an alarm for heads for these DMUs. There is no doubt that learning and growth are important factors for permanence of the system and nonchalance of managers in this issue causes difficulties for the system in the long term. Surly, poor performance in learning and growth perspective leads to decries efficiency of other stages to come. This issue can be solved by passing off educational workshop for employees. Some topics for future research were recommended as well: The main focus of the present study was the evaluation of supply chains but based on the ideas presented in this paper the proposed model can be extended to rank efficient supply chains to identify supper efficient DMUs. Also, based on the market conditions or political and economic situations, managers inclined to recognize progress and regress of units between two times can employ the model presented. In addition, there exist some questions for managers in this field; for instance, managers who are inclined to recognize the relationship between the progress and regress of stages with progress and regress of the supply chain can employ the model presented here. In the economic issues, the aforementioned problems could be solved by Malmquist productivity index, so future studies can employ combination of ideas of this paper and frontier-shift effect, and Catch-up in order to solve the existing problems.

30

References Agarwal, A., Shankar, R., & Tiwari, M.K. (2006). Modeling the metrics of lean, agile and legible supply chain: An ANP-based approach. European Journal of Operational Research, 173, 211–225. Ahumada, O., & Villalobos, J.R. (2009). Application of planning models in the agri-food supply chain: a review. European Journal of Operation Research, 195, 1-20. Amado, C.A.F., Santos S.R.P., & Margues P.M. (2011). Integrating the Data Envelopment Analysis and the Balanced Scorecard approaches for enhanced performance assessment. Omega, The International Journal of Management Science. Article In Press. Angerhofer, B.J.,& Angelides, M.C. (2006). A model and a performance measurement system for collaborative supply chains. Decision Support Systems.42, 283- 301. Asoshen, A., Nalchigar, S., & Jamporazmey, M. (2010). Information technology project evaluation: An integrated data envelopment analysis and balanced scorecard. Expert Systems with Applications, 37, 5931-5938. Baker, N., & Freeland, J. (1975). Recent advances in R&D benefit measurement and project selection methods. The International Journal of Management Science, 21, 1164-1175. Ballou, R.H., Gilbert, S. M., & Mukherjee, A. (2000). New managerial challenges from supply chain opportunities. IEEE Engineering Management Review, (Third quarter): 7–16. Banker, R.D., Charnes, A., & Cooper, W.W. (1984). Some models for estimating technical and scale inefficiencies in data envelopment analysis. Management Science, 30, 1078-1092. Beamon, B. M. (1998). Supply chain design and analysis: models and methods. International Journal of Production Economics, 55, 281-294 Beamon, B. M. (1999). Measuring Supply Chain Performance. International Journal of Operations and Production Management, 19, 275-292. Beamon, B.M., Chen, V.C.P., (2001). Performance analysis of conjoined supply chains. International Journal of Production Research, 39(14): 3195-3218. Berrah, L., & Cliville´, V. (2007). Towards an aggregation performance measurement system model in a supply chain context. Production Planning & Control, 58, 709–719. Bhagwat, R., & Sharma, M.K. (2007). Performance measurement of supply chain management: A balanced scorecard approach. Computers & Industrial Engineering, 53, 43-62. Bigliardi, B., & Bottani, E. (2010). Performance measurement in the food supply chain: a balanced scorecard approach. Facilities, 28, 249-260. Chan, F.T.S. (2003). Performance measurement in a supply chain.International Journal of Advanced Manufacturing Technology, 21, 534-348. Chan, F.T.S., & Qi, H.J. (2003). An innovative performance measurement method for supply chain management. Supply Chain Management: An International Journal, 8(3-4): 209-223. Chang, S.L., Wang, R.C., & Wang, S.Y. (2007). Applying a direct multi-granularity linguistic and strategy-oriented aggregation approach on the assessment of supply performance. European Journal of Operational Research, 177, 1013–1025. Charnes, A., Cooper, W.W., & Rhodes, E. (1978). Measuring the efficiency of decision making units. European Journal of Operational Research, 2, 429-444. Chen, I.J., & Paulraj, A. (2004). Understanding supply chain management: critical research and a theoretical framework. International Journal of Production Research, 42, 131-163.

31

Chen, T.Y., & Chen, L.H. (2007). DEA performance evaluation based on BSC indicators incorporated. International Journal of Productivity and Performance Management, 56, 335-357. Chen, T.Y., & Yu, O.S. (1997). Performance evaluation of selected US utility commercial lighting demand side management programs. Energy Engineering, 94, 50-66. Chen, T.Y., Chen, C.B., &Peng, S.Y. (2008).Firm operation performance analysis using data envelopment analysis and balanced scorecard. International Journal of Productivity and Performance Management, 57, 523-539. Chi Sun, C. (2010). A performance evaluation model by integrating fuzzy AHP and fuzzy TOPSIS methods. Expert Systems with Applications, 37, 7745-7754. Chia, A., Goh, M., & Hum, S.H. (2009). Performance measurement in supply chain entities: balanced scorecard perspective. Benchmarking: An International Journal, 16, 605-620. Cook, W.D., Zhu, J., Bi, G., &Ynage, F. (2010). Network DEA: Additive efficiency decomposition. European Journal of Operational Research, 207, 1122-1129. De Toni, A., & Tonchia, S. (2001). Performance measurement systems: models, characteristics and measures. International Journal of Operations & Production Management, 21, 46-70. Detoro, I. (1995). The 10 pitfalls of benchmarking. Quality Progress, 28, 3-61. Dixon, J.R., Nanni, A.J., & Vollmann, T.E. (1990). The new performance challenge: measuring operations for world class competition. Dow Jones, Irwin, Homewood, IL. Eilat, H., Golany, B., & Shtub, A. (2007). R&D project evaluation: An integrated DEA and balanced scorecard approach. The International Journal of Management Science, 36, 895-912. Eilat, H., Golany, B., & Shtub, A. (2005). Constructing and evaluating balanced portfolios of R&D projects with interactions: A DEA based methodology. European Journal of Operation Research, 172, 1018-1039. Ellram, L.M., Zsidisin, G.A., Siferd, S.P., & Stanly, M.J. (2002). The Impact of Purchasing and Supply management Activities on Corporate Success. Journal of Supply Chain Management, 38, 4-7. Farrell, M. J. (1957). The Measurement of Productive Efficiency. Journal of Royal Statistical Society, 120, 253-281. Farris, M.T., & Hutchison, P.D. (2002). Cash to cash: the new supply chain management metric. International Journal of Physical Distribution & Logistics Management, 32, 288-296. Felix, T.S., & Qi, H.J. (2003).An innovative performance measurement method for supply chain management, Supply Chain Management, 8, 209-223. Ferguson, B.R. (2000). Implementing supply chain management. Production and Inventory Management Journal, march, 64-67 Ghalayini, A.M., & Noble, J.S. (1996).The changing basis of performance measurement. International Journal of Operations & Production Management, 16, 63-88. Globerson, S. (1985). Issues in developing a performance criteria system for an organization. International Journal of Production Research, 23, 639-646. Golany, B. (1988). An interactive MOLP procedure for the extension of DEA to effectiveness analysis. The Journal of The Operational Research Society, 39, 725-734. Gunasekaran, A., Patel, C., & Tirtiroglu, E. (2001). Performance measures and metrics in a supply chain environment, International Journal of Operations & Production Management, 21, 71-87. Gunasekaran, A., Patel, C., Ronald, E., & McGaughey, R. (2004). A framework for supply chain performance measurement, International Journal of Production Economics, 87, 333–348. Hausman, W. H. (2002). Supply chain performance metrics. Corey Billington, Terry H., Hau L., and Hohn N. (Ed.).The practice of supply chain management.Kluwer Academic Publishiers. 32

Hoek, R.I. (1998). Measuring the unmeasurable – measuring and improving performance in the supply chain. Supply Chain Management, 3, 187-191. Huang, S.H., Sheoran, S.K., & Wang, G. (2004). A review and analysis of supply chain operations reference (SCOR) model. Supply Chain Management: An International Journal, 9, 23-29. Kao, C., & Hwang, S.N. (2008). Efficiency decomposition in two-stage data envelopment analysis: An application to non-life insurance companies in Taiwan. European Journal of Operational Research, 183, 418-429. Kaplan, R., & Noton, D., (1996). The balanced scorecard. HBS Press, Boston. Kaplan, R.S., & Norton, D.P. (1992). The balanced scorecard: measures that drive performance. Harvard Business Review, 70, 9-71. Keagan, D.P., Eiler, R.G., & Jones, C.R. (1989). Are your performance measures obsolete? Management Accounting, June, 47-134. Lambert, D.M. (1998). Supply chain management: Implementation issues and research opportunities. International Journal of LogisticManagement, 9(2), 1-19. Lambert, D.M., & Cooper, M.C. (2000). Issues in supply chain management. Industrial Marketing Management, 29, 6584. Li, S., SubbaRao, S., Ragu-Nathan, T.S., & Ragu-Nathan, B. (2005). Development and validation of a measurement instrument for studying supply chain practices. Journal of Operations Management, 23, 618-641. Lockamy, A., & McCormack, K. (2004). Linking SCOR planning practices to supply chain performance: an exploratory study. International Journal of Operations & Production Management, 24, 1192-1218. Lohman, C., Fortuin, L.,& Wouters, M. (2004). Designing a performance measurement system: A case study. European Journal of Operational Research, 156, 267-286. Min, H., Min, H., & Joo S. (2008). A data envelopment analysis-based balanced scorecard for measuring the comparative efficiency of Korean luxury hotels. International Journal of Quality & Reliability Management, 25, 349-365. Morphy, E. (1999). Measuring up. Export Today, 15, 52-57. Neely, A., & Bourne, M. (2000). Why measurement initiatives fall. Measuring Business Excellence, 4, 3-7. Neely, A., Gregory, M., & Platts, K. (1995). Performance measurement system design: a literature review and research agenda. International Journal of Operations and Production Management, 15, 80-116. Park, J.H., Lee, J.K., & Yoo, J.S. (2005). A framework for designing the balanced supply chain scorecard. European Journal of Information Systems, 14, 335-346. Parker, C. (2000). Performance measurement. Work Study, 49, 63-66. Richard, R.C. (2003). Setting benchmarks and evaluating balanced scorecards with data envelopment analysis. Benchmarking: An International Journal, 10, 226-245. Russell, K. (2001). Supply chain management. Computer World, 35(1): 32. Samuel H., Sunil S., & Ge, W. (2004). A research and analysis of supply chain operations reference (SCOR) model. Supply Chain Management: An International Journal, 9, 23-29. Schnetzler, M.J., Sennheiser, A., & Schonsleben, P. (2007). A decomposition-based approach for the development of a supply chain strategy. Production Economics, 105, 21-42. Sharma, M.K., & Bhagwat, R. (2007). An integrated BSC-AHP approach for supply chain management evaluation.Measuring Business Excellence. 11, 57-68. Sheridan, J.H. (1993). Where bench markers go wrong.Iindustry Week, 242, 28-34.

33

Soni, G., & Kodali, R. (2010). Internal benchmarking for assessment of supply chain performance. Benchmarking: An International Journal, 17,44-76. Spronk, J. & Post, T. (1999). Performance benchmarking using interactive data envelopment analysis. European Journal of Operational Research, 115, 472-487. Stephens, S. (2001). Supply chain operations reference model version 5.0: a new tool to improve supply chain efficiency and achieve best practice. Information Systems Frontiers, 3, 471-476. Stewart, G. (1995). Supply chain performance benchmarking study reveals keys to supply chain excellence. Logistics Information Management, 8,38-44. Supply Chain Council, (2006). Supply Chain Operations Reference Model: Overview of SCOR Version 8.0, available at: www.Supply Chain.org Tamura, H., & Akazawa, K. (2005). Stochastic DEMATEL for structural modeling of a complex problematique for realizing safe, secure and reliable society. Journal of Telecommunications and Information Technology, 4, 139– 146. Tone, K., & Tsutsui, M. (2009). Network DEA: A slack-based measure approach. European Journal of Operational Research, 197, 243-252. Tseng, M.L. (2010). Implementation and performance evaluation using the fuzzy network balanced scorecard. Computer and Education, 55, 188-201. Valderrama, T.G., Mendigorri, & E.M., Bordoy, D.R. (2009). Relating the perspectives of the balanced scorecard for R&D by means of DEA. European Journal of Operation Research, 196, 1177-1189. Varma, S., Wadhwa, S., & Deshmukh, S.G. (2008). Evaluating petroleum supply chain performance. Asia Pacific Journal of Marketing and Logistics, 20, 343-356. Vorst, J.V.d.,& Beulens, A. (2001). Identifying sources of uncertainty to generate supply chain redesign strategies. International Journal of Physical Distribution & Logistics Management decision, 32(6): 409-430. Wang, J.C. & Wu, H. (2006). Corporate Performance Efficiency Investigated by Data Envelopment Analysis and Balanced Scorecard. Journal of American Academy of Business, 9, 312-318. Wu, H.Y., Lin, Y.K., & Chang, C.H. (2011). Performance evaluation of extension education centers in universities based on the balanced scorecard. Evaluation and Program Planning, 34, 37-50.

34