Analyzing Alternatives in Reverse Logistics for End of Life Computers With ANP and BSC

8/6/2019 Analyzing Alternatives in Reverse Logistics for End of Life Computers With ANP and BSC

1/30

Analyzing alternatives in reverse logistics for end-of-life

computers: ANP and balanced scorecard approach*

V. Ravia, Ravi Shankara,*, M.K. Tiwarib

aDepartment of Management Studies, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110 016, Indiab

Department of Manufacturing Engineering, National Institute of Foundry and Forge Technology, Jharkhand State,Ranchi 834003, India

Abstract

Activities in reverse logistics activities are extensively practiced by computer hardware industries. One of the

important problems faced by the top management in the computer hardware industries is the evaluation of various

alternatives for end-of-life (EOL) computers. Analytic network process (ANP) based decision model presented in

this paper structures the problem related to options in reverse logistics for EOL computers in a hierarchical form

and links the determinants, dimensions, and enablers of the reverse logistics with alternatives available to the

decision maker. In the proposed model, the dimensions of reverse logistics for the EOL computers have been taken

from four perspectives derived from balanced scorecard approach, viz. customer, internal business, innovation andlearning, and finance. The proposed approach, therefore, links the financial and non-financial, tangible and

intangible, internal and external factors, thus providing a holistic framework for the selection of an alternative for

the reverse logistics operations for EOL computers. Many criteria, sub-criteria, determinants, etc. for the selection

of reverse logistics options are interrelated. The ability of ANP to consider interdependencies among and between

levels of decision attributes makes it an attractive multi-criteria decision-making tool. Thus, a combination of

balanced scorecard and ANP-based approach proposed in this paper provides a more realistic and accurate

representation of the problem for conducting reverse logistics operations for EOL computers.

q 2005 Elsevier Ltd. All rights reserved.

Keywords: Reverse logistics; Balanced scorecard; Analytic network process; Multi-criteria decision making; Computer

hardware industry

0360-8352/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.cie.2005.01.017

Computers & Industrial Engineering 48 (2005) 327356

www.elsevier.com/locate/dsw

* This manuscript was processed by Area Editor Surendra Gupta.

* Corresponding author. Tel.: C91 11 26596421; fax: C91 11 26862620.

E-mail addresses: [email protected] (V. Ravi), [email protected] (R. Shankar), [email protected]

(M.K. Tiwari).
http://www.elsevier.com/locate/dswhttp://www.elsevier.com/locate/dsw


2/30

1. Introduction

It has been estimated that about 60 million computers enter the market every year in the USA and

over 12 million computers are disposed of every year. Out of these only about 10% areremanufactured or recycled (Platt & Hyde, 1997). The remaining may lead to enormous amount of

e-waste to be generated in a few years: 4 billion pounds of plastic, 1 billion pounds of lead, 1.9million pounds of cadmium, 1.2 million pounds of chromium, 400,000 lbs of mercury, etc. (Silicon

Valley Toxics Coalition, 2002). The National Safety Council in a report ranks computers as thenations fastest-growing category of solid waste by the Environmental Protection Agency (Hamilton,

2001). By 2004, there would be more than 315 million systems ready for disposal as opposed to 21million obsolete systems in 1998 (Bertagnoli, 2000). According to another estimate, about 500

million computers will be rendered obsolete by 2007 in the USA alone ( Hamilton, 2001). With theobsolescence rates on the rise (Blumberg, 1999) an important question that remains to be answered is

what can be done to these EOL computers both from economical and environmental point of view.Due to shortening of product life cycles, for products like consumer electronics, the recovery of

value from these consumer goods, after use, is becoming a necessity (Hillegersberg, Zuidwijk, vanNunen, & van Eijk, 2001). Several alternatives exist for disposing these EOL computers. Some of the

methods for handling the EOL products include temporary storage, recycling the product, disposingof the product via landfills, etc. (Jacoby, Berning, & Diettvorst, 1977). EPAs Municipal Solid Waste

FactBook reports that 29 states in USA have 10 years or more of landfill capacities remaining, 15states have between 5 and 10 years of landfill capacity remaining, and six states have less than 5

years of landfill capacity remaining (Rogers & Tibben-Lembke, 1998). But landfill usage may be ashort-term solution to the problem as for example, states like Massachusetts, Minnesota and

Wisconsin have either banned or are considering banning the dumping of the computer-related

equipment in their landfills (Stough & Benson, 2000). The German Packaging Ordinance of 1991mandate that industries organize the reclamation of reusable packaging waste, while local authoritiescontinue to handle the collection and disposal of the remaining waste. In Taiwan, proper disposition

of computers and electrical home appliances at their EOL phase has been strongly urged by thegeneral public because of the scarcity of landfill space and the hazardous materials contained in these

appliances (Shih, 2001). If offsetting of the increasing demand of landfills is to be done, enhancedefforts for recycling are needed, which directly requires the reverse logistics activities ( Barnes, 1982).

Reverse logistics provide many opportunities to reuse and create value out of this nearly omnipresentasset (Rogers & Tibben-Lembke, 1998).

Industries have started to realize that the reverse logistics can be used to gain competitive advantage

(Marien, 1998). An evaluation framework, which incorporates determinants and dimensions of reverselogistics, would be useful in configuring the post-activities associated with the EOL computers. One ofthe prime issues in this context is the evaluation of the various alternatives faced by computer

companies, which seek to undertake reverse logistics activities for the EOL computers. One such

approach, with an application of a systemic analysis technique is presented in this paper. This techniqueevaluates the various dimensions of reverse logistics through an analytic hierarchy network model.There are a number of variables affecting the reverse logistics, some of these are interdependent among

each other. Analytical Hierarchical Process (AHP) is one of the analytical tools, which can be used tohandle a multi-criteria decision-making problem (Saaty, 1980). However, a shortfall of AHP is that it

lacks in considering interdependencies, if any, among the selection criteria. Analytic Network Process

V. Ravi et al. / Computers & Industrial Engineering 48 (2005) 327356328


3/30

(ANP) is a similar technique, but can capture the interdependencies between the criteria under

consideration, hence allowing for a more systemic analysis. It can allow inclusion of criteria, bothtangible and intangible (difficult to quantify), which has some bearing on making the best decision

(Saaty, 1996). Further, many of these factors have some level of interdependency among them, thusmaking ANP modeling better fit for the problem under study.

The ANP model presented in this paper structures the problem related to selection of an alternative forthe reverse logistics option for EOL computers in a hierarchical form and links the determinants,

dimensions, and enablers of reverse logistics with different alternatives. One of the important issues forany strategic planning would be how the organization should prioritize the determinants and what policyelements or initiatives impact them (Wheelwright, 1978). The balanced scorecard is a performance

measurement system that allows managers to look at the business from four divergent importantperspectives: customer, internal business, innovation and learning, and finance (Kaplan & Norton,

1992). Brewer and Speh (2000) had used the concept of balanced scorecard to measure the supply chain

performance. In the proposed ANP model, the dimensions of the reverse logistics for the EOL computershave been taken from the four perspectives of the balanced scorecard, thus balancing as well as linkingthe financial and non-financial, tangible and intangible, internal and external factors. Therefore, theproposed framework provides a holistic approach to the selected multi-criteria decision-making problem

for EOL computers.This paper is further organized as follows. Section 2 provides a background of reverse logistics and its

application in various industries including the computer industry. Then a brief discussion of the

determinants, dimensions and their enablers of reverse logistics and alternatives to be evaluated in thismodel are provided. These characteristics are then used to structure the model. Later, the proposedmethodology for evaluating the decision model is presented and applied to a decision-making problem

faced by a small computer hardware company. This is followed by a discussion and managerial

implications of this research. Finally, we conclude the work with the limitations of this work anddirections for further research.

2. Reverse logistics

Reverse logistics is the movement of the goods from a consumer towards a producer in a channel ofdistribution (Murphy, 1986). Stock (1992) recognized the field of reverse logistics as being relevant for

business and society in general. Kopicki, Berg, Legg, Dasappa, and Maggioni (1993) paid attention tothe field and pointed out opportunities on reuse and recycling. Fleischmann, Bloemhof-Ruwaard,

Dekker, van der Laan, van Nunen, and Van Wassenhove (1997) had given a comprehensive review ofliterature of the quantitative models in reverse logistics. Reverse logistics programs in addition to thevarious environmental and the cost benefits can proactively minimize the threat of government

regulation and can improve the corporate image of the companies (Carter & Ellram, 1998). Reverselogistics is the process of planning, implementing, and controlling the efficient, cost effective flow of raw

materials, in-process inventory, finished goods and related information from the point of consumption tothe point of origin for the purpose of recapturing value or proper disposal (Rogers & Tibben-Lembke,1998). A reverse logistics defines a supply chain that is redesigned to efficiently manage the flow of

products or parts destined for remanufacturing, recycling, or disposal and to effectively utilize resources(Dowlatshahi, 2000).

V. Ravi et al. / Computers & Industrial Engineering 48 (2005) 327356 329


4/30


5/30

3. Determinants of reverse logistics

Economic factors both directly and indirectly (de Brito & Dekker, 2003), legislation (de Brito &

Dekker, 2003), corporate citizenship (de Brito & Dekker, 2003; Rogers & Tibben-Lembke, 1998) andenvironmental and green issues (Rogers & Tibben-Lembke, 1998) are the four determinants of reverselogistics taken into account in this research. These are briefly described below.

3.1. Economic factors

Economics is seen as the driving force to reverse logistics relating to all the recovery options, where

the company receives both direct as well as indirect economic benefits. It is seen that companiescontinually strive for achieving cost savings in their production processes. If a firm does reverse logistics

well, it will make money (Stock, 1998). The recovery of the products for remanufacturing, repair,

reconfiguration, and recycling can lead to profitable business opportunities (Andel, 1997). Reverselogistics is now perceived by the organizations as an investment recovery as opposed to simplyminimizing the cost of waste management (Saccomano, 1997). A reverse logistics program can bringcost benefits to the companies by emphasizing on resource reduction, adding value from the recovery of

products or from reducing the disposal costs. Guide and Wassenhove (2003) give an example of the USfirm named ReCellular, which by refurbishing the cell phones, had gained economic advantage. Thus,the economic drivers of reverse logistics lead to direct gains in input materials, cost reduction, valueadded recovery and also in indirect gains by impeding legislation, market protection by companies,

green image for companies and for improvement in customer/supplier relations.

3.2. Legislation

Another important driver for the reverse logistics is legislation. Legislation refers to any jurisdiction

that makes it mandatory for the companies to recover its products or accept these back after the end-of-life of the product. These may include collection and reuse of products at the end of the product life

cycle, shift waste management costs to producers, reduce volume of waste generated, and the use ofincreased recycled materials. For example, the Waste Electrical & Electronics Equipment directiveencourages a set of criteria for collection, treatment and recovery of waste electrical and electronic

equipment and makes producers responsible for financing these activities (WEEE, 2003). There has alsobeen a restriction on the use of hazardous substances in the production processes, which facilitates thedismantling, and recycling of waste electronics. A reverse logistics decision for the EOL computers

should ensure that the end-of-life products are retired in a way that is compliant with existing legislation.

3.3. Corporate citizenship

Another driver for the reverse logistics is the corporate citizenship that concerns a set of values or

principles that impels a company or an organization to become responsibly engaged with reverse logisticsactivities. Reverse logistics activities can lead to increase of corporate image (Carter & Ellram, 1998). Agood example in this context would be of Paul Farrow, the founder of Walden Paddlers, Inc., whose

concern of the velocity at which consumer products travel through the market to the landfill, pushedhim to an innovative project of a 100-percent-recyclable kayak (Farrow, Johnson, & Larson, 2000).



6/30

In 1996, Hanna Andersson, a million direct retailer of infants and toddlers clothes developed a program

called Hannadowns in which they distributed the childrens gently worn returned clothes to schools,homeless shelters, and other charities (Spence, 1998). Nike, the shoe manufacturer encourages

consumers to bring their used shoes to the store where they had purchased them after their usage. Theyship these back to Nike plant where these are shredded and made into basketball courts and runningtracks. Nike also donates the material to the basketball courts and donates fund for building andmaintaining these courts, thus enhancing the value of brand ( Rogers & Tibben-Lembke, 1998). It is

seen from the last two examples that few firms are acting as good corporate citizens by contributing tothe good of the community and assisting the people who are probably less fortunate than their typicalcustomers.

3.4. Environment and green issues

Concern for the environment and green issues is also one of the drivers of reverse logistics. Thereverse logistics lead to benefits of environment (Byrne & Deeb, 1993; Carter & Ellram, 1998; Wu &Dunn, 1995). Hart (1997) proposes that the principle of the ecological footprint indicates the relevanceof greening initiatives for countries. Reverse logistics has led to competitive advantage to companies

which proactively incorporate environmental goals into their business practices and strategic plans(Newman & Hanna, 1996). Managers are giving increasing importance to the environmental issues

(McIntyre, Smith, Henham, & Pretlove, 1998). The environmental management has gained increasinginterest in the field of supply chain management. Handfield and Nichols (1999) mention greening as a

critical future avenue in this area. Murphy, Poist, and Braunschweig (1995) have found that 60% in agroup of 133 managers surveyed considered the issue of the environment to be a very important factorand 82% of them expected that the importance would increase in the years to come. A green image of

producing environmentally friendly products has become an important marketing element, which hasstimulated a number of companies to explore options for take-back and recovery of their products(Thierry, 1997). A reverse logistics operations for EOL computers should ensure that the environmental

and green issues are taken into account.

4. Dimensions of the reverse logistics

In this paper, we inherit the dimensions of the balanced scorecard, which allow the managers to

look at the business from four important perspectives, namely, the customer, internal business,

innovation and learning, and financial perspectives (Kaplan & Norton, 1992). Although the conceptof balanced scorecard has been primarily designed for the measurement of the system performance,in this model, we have used these dimensions to evolve a holistic framework towards the conduct ofreverse logistics operations for EOL computers. The four dimensions and their enablers are briefly

discussed below.

4.1. Customer perspective

This dimension depicts what a customer expects from the reverse logistics operations. Pochampallyand Gupta (2004) reported that the success of a prospective reverse supply chain depends heavily on



7/30

the participation of three important groups, viz. customers, local government officials, and supply chain

executives who have multiple, conflicting, and incommensurate goals and thus the potentials must beevaluated based on the maximized consensus among these three groups. For the customers, they opine

that the principal concern is convenience. A research study conducted on the logistics service providersin Singapore found that the voice of customer as the most important driver of the logistics management(Sum & Teo, 1999). The present day customers demand that manufacturers reduce the quantities ofwaste generated by their products. They demand clean and energy saving production processes from

their suppliers. They want that the potentially dangerous materials used in the production process bereplaced by those that minimize harm to users, more in line with the present-day eco-compatiblevalues. Today the customers are ready to pay more for a green product. In fact, customers drive the

corporation green (Vandermerwe & Oliff, 1990). There has been an increased acceptance from thecustomers for recycled goods and packaging due to concerns with the environment. Reverse logistics

also influences the customer service and satisfaction; as for example, the ability of companies to quickly

and efficiently handle the return of product for necessary repair is critical for its survival (Blumberg,1999). Thus, it is seen that the reverse logistics operations should offer services based on the customerperspective.

4.2. Internal business perspective

This dimension illustrates in what areas must the reverse logistics operations excel at so as to beable to achieve the target. Reverse logistics managers should focus on those critical internal

operations that would enable them to satisfy the customer needs. Information support is one of theways to develop linkages to achieve efficient reverse logistics operations (Daugherty, Myers, &

Richey, 2002). This is due to the availability of prompt and accurate information with which the

logistics managers are able to foresee the products that would be returned, thus aiding to the processof more efficient receipt and return of the products. The product recovery management is the ability

of a supply chain in recovering the economic (and ecological) value of a product as reasonably aspossible, thereby reducing the ultimate quantities of waste. The various product recovery options

could be repairing, refurbishing, remanufacturing, cannibalization, and recycling (Thierry et al.,1995). A sincere and committed effort from the top management is essential for successfuldeployment of reverse logistics programs (Carter & Ellram, 1998). New technologies are also

necessary. Many companies have product development programs encompassing design forenvironment for product recovery through disassembly. An example is the case of Xerox Europereported by Maslennikova and Foley (2000).

4.3. Innovation and learning perspective

This dimension of balanced scorecard focuses on whether the reverse logistics operations for EOLcomputers can continue to improve and create more value for customers by improving the efficiency.

Manufacturers and customers can reengineer their businesses to better serve the ultimate customers,rather than being regulated into positions that may not be of most advantageous to the channelmembers. This is very necessary to create competitive advantage as the customers become

more environment conscious (Marien, 1998). Reverse logistics is initiated as a strategic variablefor competitive reasons (Rogers & Tibben-Lembke, 1998). The environmental sustainability



8/30

and ecological performance of a company also depend on the suppliers ( Godfrey, 1998). So many

companies have started partnering and mentoring with their suppliers such as providing guidance toset up an environmental management system (EMS) to improve the operational efficiency (Hines &

Johns, 2001). Strategic alliances are made with various members of supply chain as thecompanies are realizing that the individual attempts at product reclamation make little sense botheconomically or environmentally (Cairncross, 1992). The Knowledge Management may also beused for logistics network. Smirnov, Pashkin, Chilov, and Levashova (2002) describe Knowledge

Source Network configuration approach (KSNet-approach) to knowledge logistics through knowledgefusion. Smirnov, Pashkin, Chilov, and Levashova (2001) present an ontology management in multi-agent system for knowledge logistics. Smirnov (2001) provides a profile-based configuring of

knowledge supply networks in the global business information environment. Thus, constantinnovation and learning processes are necessary for the successful conduct of reverse logistics

operations.

4.4. Finance perspective

This dimension of balanced scorecard indicates how the reverse logistics operations cater to theshareholders financial objectives. This indicates whether the companys strategy, implementation andexecution are contributing to bottom-line improvement. This can be enabled by following reverse

logistics activities of waste reduction, cost savings and recapturing value from the recovered products(Bacallan, 2000; Deere & Company, 1998; Hans & Byrne, 1993; Thierry et al., 1995). Effective

reverse logistics contributes to regaining value from reusing products or parts or from recyclingmaterials. There were more than 70,000 re-manufacturing firms in USA for jet and car engines, auto

parts and copiers, total sales amounting to 53 billion USD in 1998 (Lund, 1998). Reverse logisticscan improve the cost savings for the company. A research finding has reported that companies thatmake use of remanufacturing in the product recovery are estimated to save 4060% of the costs

compared to manufacturing a completely new product (Cohen, 1988; Heeb, 1989; Toensmeier, 1992;Wilder, 1988) while requiring only 20% of the effort (Lund, 1984; Sturgess, 1992). The reverselogistics process consists of recapturing value from the recovery of products (Rogers &

Tibben-Lembke, 1998). Implementing reverse logistics programs to reduce, reuse, and recyclewastes from distribution and other company processes produces tangible and intangible value(Kopicki et al., 1993). Kokkinaki, Dekker, de Koster, Pappis, and Verbeke (2001) give example of a

computer company that recaptured reasonable value from the computer product by supportingalternative uses for the products.

5. Alternatives for the conduct of reverse logistics operations

After review of literature and discussion with experts in the field of reverse logistics, both fromindustry and the academia, some of the important selection criteria for the conduct of reverse logisticsoperations are identified. For the purpose of illustration of our model, we analyze three distinct

alternatives. The three categories of alternatives are Third Party Demanufacturing (TPD), SymbioticLogistics Concept (SLC), and Virtual Reverse Logistics Network for PCs (VRL). These criteria have



9/30

been used in the proposed framework for the development of an ANP model. A brief description of these

three alternatives follows in Sections 5.15.3.

5.1. Third party demanufacturing (TPD)

Grenchus, Keene, and Nobs (1997) and White, Masanet, Rosen, and Beckman (2003) haveopined that most computer recovery businesses engage in demanufacturing operations to processobsolete and scrap end-of-line computer products. Lieb and Randall (1999) reported that third-partyexecutives viewed reverse logistics as an opportunity area and suggest that reverse logistics

activities performed by third-party providers may become more prevalent in future. This allows theoriginal equipment manufacturer (OEM) with the opportunity to focus on their core competencies,leaving the demanufacturing operations to private companies specializing in these functions. Spicer

and Johnson (2004) proposed the concept of Third Party Demanufacturing (TPD). It is defined as

an extended producer responsibility (EPR) approach in which private companies take up end-of-liferesponsibility for products on behalf of the OEM. In this arrangement, an OEM would pay a fee toa Product Responsibility Provider (PRP) that would ensure that the manufacturers product are

disposed in a way that is environmentally responsible while compliant with EPR legislation. Eachproduct of the OEM is marked with an identifier provided by the PRP. As the products are sold inthe market, the OEM pays money to its PRP for the end-of-life management of the products. Thus,

the liability of end-of-life management of the products is transferred to the PRP and themanufacturer has no further financial risk associated with the end-of-life product. After many years,when the product reaches to its end-of-life it enters a collection system. The product from the

collection system is delivered to a local recycler who works in partnership with the PRP. Therecycler demanufactures the product and receives payment from the PRP. They meetthe environmental targets set by the EPR legislation and also return the potential components

that the OEM desires for closed-loop recycling. The PRP on his part provides his recycling partnerwith technical support, including disassembly instructions and material and parts identification.Both the manufacturers and general public reap the benefits of this system. The benefits to the

manufacturers from this type of system are:

(i) they can focus on their core competencies and leave the demanufacturing process to the

specialized companies,(ii) this approach allows them to meet their product end-of-life responsibilities while

simultaneously eliminating the financial risk associated with the end-of-life uncertainties,

(iii) this provides them an opportunity to reap the benefits of better design through the competitivePRP bid process, and(iv) they can also tap the potential benefits derived out of the demanufacturing innovation and

efficiencies driven by competition in the recycling industry.

The benefits to the general public by this system are:

(i) third party demanufacturing provides a discerning and immediate economic feedback to the product

design process, driving improvements,(ii) the competition leads to the promotion of innovation in the demanufacturing industry, and



10/30

(iii) more importantly, it distributes the demanufacturing process to the local level, creating jobs andreduces the transportation inefficiencies.

Though the third-party take-back mode of EPR has the advantages described above, it suffers fromfew limitations. Some of the difficulties associated with this system are:

(i) identification of products at end-of-life,

(ii) problem of dismantling products and recycling materials, and(iii) difficulties faced by the local demanufacturers in finding suitable markets for the recyclable

materials and parts that they have removed from a wide variety of products.

The principal role of the PRP is to help solve these problems outlined above.

5.2. Symbiotic logistics concept (SLC)

The hyperdictionary defines symbiosis as the relation between two different species of organisms

that orgnanisms that are interdependent; each gains benefits from the other. Adler (1966) had used thisconcept and defined the symbiotic marketing as the alliance of resources or programs between two ormore independent organizations, designed to improve the marketing potential of each. A number of

factors like swift pace of technological change, the impact of shifting markets and consumer tastes, theinformation explosion, automation, the staggering financial burden of research and development, theinternationalization of business, and the growing ferocity of competition as everyone gets into everyone

elses business by means of mergers, acquisitions, an intensive new product development increases therelevance of symbiotic marketing (Adler, 1966). Turner, LeMay, and Mitchell (1994) have examined the

potential benefits of symbiotic relationships in reverse logistics. Symbiotic relationships assume a greatimportance as the competition becomes more global in the present day environment. Degher (2002) has

reported the take-back and recycling programs at Hewlett-Packard Ltd and concluded that electronicmanufacturers and government agencies should work together to better provide customers withenvironmentally responsible take-back and recycling programs. The new emphasis in business

community is on forming strategic alliances (Achrol, Reve, & Stern, 1983; Bowersox, 1990; Heide &John, 1990, 1992). One of the challenges faced by the management is to build flexibility in theirorganization, as vertical hierarchies are replaced with horizontal networks; traditional functions are

linked through interfunctional teams; and strategic alliances are formed with suppliers, customers oreven competitors (Hirschorn & Gilmore, 1992). The firms have realized the potential of the mutual

benefits arising from the working in concert rather than independently by pooling of the resources(Lambert & Stock, 1993).Symbiotic logistics is defined as the strategic alliance of two or more independent entities designed to

provide the desired level of customer service in accordance with the concept of integrated logisticsmanagement (Mitchell, LeMay, Arnold, & Turner, 1993). Reverse logistics force systems designed to

move goods and services forward to act in the reverse direction. Using the symbiotic concept could solvea number of problems of the reverse channels. It could provide the effective means for combating theproblems created by the need to maintain the reverse channel capability.

A number of factors like the political and legal, competition, technological, and economic conceptsincrease the relevance of symbiotic relationships in reverse logistics. Turner et al. (1994) provided two



11/30

examples of the problems associated with reverse logistics and the ways symbiotic relationships can help

them combating it. In the first example, they illustrated a symbiotic logistics relationship that solves thereverse logistics problem by circumventing true reverse channels. In the second, they illustrated

symbiotic relationships formed as a result of the environmental regulation that truly reflect a reverselogistics process. An example in this regard would be the German packaging ordinance of 1991, whichresulted in the companies working closely with competitors to put tough environmental policies intopractice (Cairncross, 1992; Cooke, 1992).

In the effective implementation of reverse logistics, companies have realized that individual attemptsat the product reclamation make little sense, both economically as well as environmentally. One reasonfor this could be that the volume of the returned products involved is probably too small to justify

individual effort. Thus, the logical solution to this problem would be to pool the resources with otherfirms in similar situations in order to gain economies. More importantly, the disruptive effective effects

of reverse channels of distribution could be minimized by the symbiotic relationships.

The main difference between TPD and SLC are the parties that are involved in the conduct of reverselogistics operations for end-of-life computers, such that these products are retried in an environmentallyfriendly manner. In the TPD, original equipment manufacturers outsource the demanufacturingoperations to third party private companies. The companies take up the responsibility of disposing of

these end-of-life products on behalf of OEMs. This approach allows the OEMs with the opportunity tofocus on its core competencies, leaving the demanufacturing operations to the specialized privatecompanies performing these functions. Thus, in this concept, only the OEM and the private companies

assigned by them to carry out reverse logistics operations are involved. On the other hand, in SLC theorganizations may forge strategic alliances with other firms, suppliers, customers, and may be even theircompetitors to achieve common business objectives. The relationships between the parties involved in

this process is symbiotic in that two or more dissimilar organizations essentially pool the resources in an

attempt to realize the benefits of reverse logistics not available to the parties individually. The symbioticrelationships between the parties involved are necessary as the volume of products is probably too smallto economically justify individual efforts at product reclamation. Thus, the parties involved can reapmutual benefits that arise from working in concert rather than working independently.

5.3. Virtual reverse logistics network for PCs

The Internet revolution has led to new forces of global competition, availability of increased

information, educated customers, rapid innovations, and changing relationships. E-commerce may takeplace among the businesses (B2B) or between business and consumers (B2C). Internet encompasses a

wider spectrum of potential commercial activities and offers information exchanges necessary for anelectronic marketplace intermediary as shown in Fig. 1. Kokkinaki, Dekker, Lee, and Pappis (2001) haveproposed a model of virtual reverse logistics network for the PCs. This network relies on e-commerce and

www technologies for remote monitoring and benchmarking, instead of physical transportation anddistribution. In this a configuration monitoring and benchmarking agent screens the computer that is about

to enter the end-of-use stream and registers the data in the system databases. This virtual reverse logisticsnetwork regards all incoming PCs as submitted offers. In this framework, the users or agents, consider thepossibility of explicitly registering requests or offers for PCs or modules that are matched automatically. A

decision support module provides the recommendations for reuse, remanufacturing or recycling of theend-of use PCs. An electronic marketplace matches the requests. Bakos (1998) defines an electronic



12/30

marketplace as medium, which facilitates the exchange of information, goods, services, and payments andin the process, thus creating economic value for buyers, sellers, market intermediaries, and for society atlarge. The unique feature of the electronic marketplace is that it brings multiple buyers and sellers together

(in a virtual sense) in one central market space. Various used products are for sale in these sites and thepotentialcustomers have a chanceof getting relevant information on them online, declare their interest andthe possibility of buying them. US-based electronic marketplaces (like www.ebay.com and www.onsale.

com) and EU electronic marketplaces (like www.partikulier.nl and www.qxl.com) trade a wide variety ofproducts entering reverse logistics chain, but some sectors like computers, electronics and hi-techequipment are particularly popular (Kokkinaki, Dekker, van Nunn, & Pappis, 1999).

The building blocks of the virtual reverse logistics networks for PCs are described below ( Kokkinaki,Dekker, Lee, et al., 2001):

(i) Repository. It holds all the application data. The persistent storage of the repository is guaranteed bythe making use of a relational database system.

(ii) Application server. It accepts and processes all HTTP requests.

(iii) Security. It is ensured by the controls system access in conjunction with the Java Cookietechnology.

(iv) Mailer. It sends a notification to the appropriate recipients of the system, every time a new request

or offer has been successfully submitted.(v) Presenter. It presents the appropriate JSP or html according to the user status (registered or not

registered) and according to the user request.

6. The decision environment

A graphical representation of the ANP model and decision environment is shown in Fig. 2. It can beseen that the overall objective is to carry out reverse logistics processes for EOL computers for the

critical business process defined as conduct operations. The determinants of reverse logistics(economic factors, legislation, corporate citizenship and environment and green issues) described in

Fig. 1. Internet applications (Coppel, 2000).

http://www.ebay.com/http://www.onsale.com/http://www.onsale.com/http://www.partikulier.nl/http://www.qxl.com/http://www.qxl.com/http://www.partikulier.nl/http://www.onsale.com/http://www.onsale.com/http://www.ebay.com/


13/30

Section 3 are modeled to have dominance over the dimensions of reverse logistics. The reverse logistics

attribute enablers are those that assist in achieving the controlling dimension of reverse logistics. Thus,

these are dependent on the dimension. Also, there are some interdependencies among the enablers,

hence the arrow arching back to the enablers decision level (Fig. 2). For example, CSF (customer

satisfaction) and GP (green products) are interdependent to some degree. In order to achieve customer

satisfaction, green products need to be produced by companies.

The reverse logistics implementation alternatives in this model are the specific projects or policies

that a decision maker wishes to evaluate, given the various attribute levels of the reverse logistics.

The various alternatives available to the decision maker in this example include third party

Fig. 2. ANP model for the reverse logistics operations for EOL computers.



14/30

demanufacturing, symbiotic logistics concept and the virtual reverse logistics network for PCs. In

Section 7, we briefly describe the benefits of the ANP process and apply it to a small company examplein PC manufacturing to explain the ANP methodology.

7. Methodology: the analytic network process

ANP (Saaty, 1996) is a comprehensive decision-making technique that captures the outcome of thedependence and feedback within and between the clusters of elements. Analytical Hierarchy Process

(AHP) serves as a starting point of ANP. The ANP is a coupling of two parts, where the first consists of acontrol hierarchy or network of criteria and sub-criteria that controls the interactions, while the secondpart is a network of influences among the elements and clusters. In fact, ANP uses a network without a

need to specify levels as in a hierarchy. The main reason for choosing the ANP as our methodology for

selecting the reverse logistics operations is due to its suitability in offering solutions in a complex multi-criteria decision environment. Some of the fundamental ideas in support of ANP are (Saaty, 1999):

ANP is built on the widely used AHP technique, ANP allows for interdependency, therefore ANP goes beyond AHP,

the ANP technique deals with dependence within a set of elements (inner dependence) and among

different sets of elements (outer dependence),

the looser network structure of the ANP makes possible the representation of any decision problemwithout concern for what criteria comes first and what comes next as in a hierarchy,

the ANP is a non-linear structure that deals with sources, cycles and sinks having a hierarchy of linearform with goals in the top level and the alternatives in the bottom level,

ANP portrays a real world representation of the problem under consideration by prioritizing not onlyjust the elements but also groups or clusters of elements as is often necessary, and

the ANP utilizes the idea of a control hierarchy or a control network in dealing with different criteria,

eventually leading to the analysis of benefits, opportunities, costs, and risks.

7.1. ANP as a qualitative tool

ANP is a multi-attribute, decision-making approach based on the reasoning, knowledge, and

experience of the experts in the field. ANP can act as a valuable aid for decision making involving bothtangible as well as intangible attributes that are associated with the model under study. ANP relies on theprocess of eliciting managerial inputs, thus allowing for a structured communication among decision

makers. Thus, it can act as a qualitative tool for strategic decision-making problems. Sarkis andSunderraj (2002) used the ANP model to conduct a comprehensive evaluation of qualitative andquantitative factors for hub location at Digital Equipment Corporation.

7.2. Advantages of ANP

ANP is a comprehensive technique that allows for the inclusion of all the relevant criteria; tangible aswell as intangible, which have some bearing on decision-making process (Saaty, 1996)



15/30

AHP models a decision-making framework that assumes uni-directional hierarchical relationshipamong decision levels, whereas ANP allows for more complex relationship among the decision levels

and attributes as it does not require a strict hierarchical structure.

In decision-making problems, it is very important to consider the interdependent relationship amongcriteria because of the characteristics of interdependence that exists in real life problems. The ANPmethodology allows for the consideration of interdependencies among and between levels of criteria

and thus is an attractive multi-criteria decision-making tool. This feature makes it superior from AHPwhich fails to capture interdependencies among different enablers, criteria, and sub-criteria (Agarwal& Shankar, 2003).

ANP methodology is beneficial in considering both qualitative as well as quantitative characteristicswhich need to be considered, as well as taking non-linear interdependent relationship among theattributes into consideration (Meade & Sarkis, 1999).

ANP is unique in the sense that it provides synthetic scores, which is an indicator of the relative

ranking of different alternatives available to the decision maker.

7.3. Disadvantages of ANP

Identifying the relevant attributes of the problem and determining their relative importance indecision-making process requires extensive discussion and brainstorming sessions. Also, data

acquisition is a very time intensive process for ANP methodology. ANP requires more calculations and formation of additional pair-wise comparison matrices as

compared to the AHP process. Thus, a careful track of matrices and pair-wise comparisons of

attributes is necessary.

The pair-wise comparison of attributes under consideration can only be subjectively performed, andhence their accuracy of the results depends on the users expertise knowledge in the area concerned.

8. A small PC manufacturing company example

The ANP model that is presented in this research has been evaluated in an actual computermanufacturing company, which was interested in the implementation of the reverse logistics practices.

Due to the limited budget constraints, the company wanted a systematic way to determine the best possibleoption for conducting the reverse logistics operations. The case experience helps us to understand in abetter way the advantages and disadvantages of the methodology from a practical point of view. Theanalysis and the implementation of the ANP model are presented in the following nine steps.

8.1. Step 1. Model development and problem formulation

In this step, the decision problem is structured into its important components. The relevant criteria andalternatives are chosen on the basis of the review of literature and discussion with few both from industryand academia. The relevant criteria and alternatives are structured in the form of a control hierarchy

where the criteria at the top level in the model have the highest strategic value. The top-level criteriain this model are economic factors (ECO), legislation (LEGI), corporate citizenship (CC), and



16/30

environmental and green issues (EGI). These four criteria are termed as the determinants. In the second

level of hierarchy, four sub-criteria termed as dimensions of the model is placed which supports all thefour determinants at the top level of hierarchy. These are customer perspective (CP), internal business

perspective (IBP), innovation and learning perspective (IBP), and financial perspective (FP). Forexample, good internal business processes helps in achieving the four determinants of ECO, LEGI, CC,and EGI. Similar relationships are valid for CP, IBP, and FP.

In this ANP model, each of the four dimensions has some enablers, which help achieve that particular

dimension. For example, the dimension IBP is supported by the enablers IT, PRO, CTM, and NTE.These enablers also have some interdependency on one another. For example, in the dimension IBP,enablers CTM and NTE are interdependent as a sincere commitment by the top management would be

necessary for procuring new technologies. The degree of interdependency may vary from case to caseand would be captured in later steps.

The strength of the ANP model is that the feedback and the network structure of the ANP makes

possible the representation of the decision problem without much concern for what comes first and whatcomes next in a hierarchy. The objective of this hierarchy is to select the best possible alternative thatwill best meet the goals of conducting effective reverse logistics in a computer industry. The ANP modelso developed is presented in Fig. 2. The alternatives that the decision maker wishes to evaluate are

shown at the bottom of the model.The opinion of the logistics manager of the company was sought in the comparisons of the relative

importance of the criteria and the formation of pair-wise comparison matrices to be used in the ANP

model. In this paper, mainly for the purpose of brevity, we present and illustrate the results only of thelegislation determinant. The results of all the four determinants would be used in the calculation ofreverse logistics overall weighted index (RLOWI), which indicates the score assigned to a reverse

logistics operation.

8.2. Step 2. Pair-wise comparison of four determinants

In this step, the decision maker is asked to respond to a series of pair-wise comparisons where twocomponents at a time are compared with respect to an upper level control criterion. These comparisons

are made so as to establish the relative importance of determinants in achieving the case companysobjectives. In such comparisons, a scale of 19 is used to compare two options (Saaty, 1980). In this ascore of 1 indicates that the two options under comparison have equal importance, while a score of 9

indicates the overwhelming dominance of the component under consideration (row component) over thecomparison component (column component) in a pair-wise comparison matrix. In case, a component has

weaker impact than its comparison component, the range of the scores will be from 1 to 1/9, where 1indicates indifference and 1/9 represents an overwhelming dominance by a column element over the rowelement. For the reverse comparison between the components already compared, a reciprocal value is

automatically assigned within the matrix, so that in a matrix aijajiZ1. The matrix showing pair-wisecomparison of determinants along with the e-vectors of these determinants is shown in Table 1.

The e-vectors (also referred to as local priority vector) are the weighted priorities of the determinantsand shown in the last column of the matrix. In this paper, a two-stage algorithm ( Saaty, 1980) is used forcomputing e-vector. For the computation of the e-vector, we first add the values in each column of thematrix. Then, dividing each entry in each column by the total of that column, the normalized matrix is

obtained which permits the meaningful comparison among elements. Finally, averaging over the rows is



17/30

performed to obtain the e-vectors. These e-vectors would be used in Table 9 for the calculation of reverselogistics overall weighted index (RLOWI) for alternatives.

8.3. Step 3. Pair-wise comparison of dimensions

In this step, a pair-wise comparison matrix is prepared for determining the relative importance of eachof the dimensions of reverse logistics (CP, IBP, ILP and FP) on the determinant of reverse logistics. In

the model, four such matrices would be formed one for each of the determinant. One such matrix for thelegislation determinant is shown in Table 2. From this table, the results of the comparison (e-vectors) of

the dimensions for the legislation determinant are carried as Pja in Table 8.

8.4. Step 4. Pair-wise comparison matrices between component/enablers levels

In this step, the decision maker is asked to respond to a series of pair-wise comparisons where twocomponents would be compared at a time with respect to an upper level control criterion. The pair-wise

comparisons of the elements at each level are conducted with respect to their relative influence towardstheir control criterion. In the case of interdependencies, components within the same level may be

viewed as controlling components for each other, or levels may be interdependent on each other.For a determinant, pair-wise comparison is done between the applicable enablers within a given

dimension cluster. The pair-wise comparison matrix for the dimension IBP under the LEGI determinantis shown in Table 3. For the pair-wise comparison, the question asked to the decision maker is, what isthe relative impact on internal business perspective by enabler a when compared to enabler b, in

improving the legislation?In Table 3, the relative importance of CTM when compared to IT with respect to IBP, in achieving the

legislation, is five. From Table 3 it is also observed that for the case company, the enabler CTM has

Table 2

Pair-wise comparisons of dimensions

Legislation CP IBP ILP FP e-vector

CP 1 1/6 3 2 0.2375

IBP 6 1 5 1/4 0.3931

ILP 1/3 1/5 1 6 0.1889

FP 1/2 4 1/6 1 0.1805

Table 1

Pair-wise comparison of determinants

Determinants ECO LEGI CC EGI e-vector

Economic factors 1 1/7 6 3 0.2378

Legislation 7 1 5 2 0.5433

Corporate

citizenship

1/6 1/5 1 1/3 0.0610

Environment and

green issues

1/3 1/2 3 1 0.1579



18/30

the maximum influence (0.4914) on IBP in improving the legislation. Similarly, IT has the minimuminfluence (0.0993) on IBP in improving the legislation. The number of such pair-wise comparison

matrices depends on the number of determinants and the dimensions in the ANP model. In this model, 15such pair-wise comparison matrices are formed. The e-vectors obtained from these matrices are

imported as AD

kja in Table 8.

8.5. Step 5. Pair-wise comparison matrices of interdependencies

Pair-wise comparisons are done to consider the interdependencies among the enablers. One suchcomparison is presented in Table 4.

It represents the result of LEGIIBP cluster with IT as the control attribute over other enablers. Thequestion asked to the decision maker for evaluating the interdependencies is when considering IT withregards to increasing legislation, what is the relative impact of enabler a when compared to enabler b? For

example, when considering IT, with regards to increasing legislation, what is the relative impact of PRO

when compared to CTM? From Table 4, it is observed that PRO (0.6249) has the maximum impact onIBPLEGI cluster with IT as the control enabler over others. It is also observed that the impact of NTE on

IT in LTRLEGI cluster is minimum (0.0916). Therefore, NTE is not a problem for the user company andit will have little impact on information technologies in IBPLEGI cluster. For each determinant, there

will be 15 such matrices at this level of relationship. The e-vectors from these matrices are used in theformation of super matrices. As there are four determinants, 60 such matrices will be formed. The

e-vectors from matrix in Table 4 have been used in sixth column of the super matrix in Table 6.

8.6. Step 6. Evaluation of alternatives

The final set of pair-wise comparisons is made for the relative impact of each of the alternatives (TPD,SLC and VRL) on the enablers in influencing the determinants. The number of such pair-wise

Table 3

Pair-wise comparison matrix for internal business perspective under the legislation determinant

Internal business

perspective

IT PRO CTM NTE e-vector

IT 1 1/2 1/5 1/2 0.0993

PRO 2 1 1/3 3 0.2497

CTM 5 3 1 2 0.4914

NTE 2 1/3 1/2 1 0.1596

Table 4

Pair-wise comparison matrix for enablers under legislation, internal business perspective and information technologies

Information

technologies

PRO CTM NTE e-vector

PRO 1 5 3 0.6249

CTM 1/5 1 7 0.2835

NTE 1/3 1/7 1 0.0916



19/30

comparison matrices is dependent on the number of enablers that are included in each of thedeterminants. In our present case, there are 15 enablers for each of the determinants, which lead to 60

such pair-wise matrices. One such pair-wise comparison matrix is shown in Table 5, where the impactsof three alternatives are evaluated on the enabler IT in influencing the determinant LEGI. The e-vectors

from this matrix are used in columns 68 of compatibility desirability indices matrix in Table 8. Thecolumns 68 in Table 8 correspond to TPD, SLC and VRL, respectively.

8.7. Step 7. Super matrix formation

The super matrix allows for a resolution of the interdependencies that exist among the elements

of a system. It is a partitioned matrix where each sub-matrix is composed of a set of relationshipsbetween and within the levels as represented by the decision makers model. In this model, there

are four super matrixes for each of the four determinants of reverse logistics hierarchy network,which need to be evaluated. One such super matrix M, shown in Table 6, presents the results of therelative importance measures for each of the enablers for the legislation determinant of the reverselogistics.

The values of the elements of the super matrix M have been imported from the pair-wise comparisonmatrices of interdependencies (for example, Table 4). As there are 15 such pair-wise comparisonmatrices, one for each of the interdependent enablers in the legislation, there will be 15 non-zero

columns in this super matrix. Each of the non-zero values in the column is the relative importance weightassociated with the interdependent pair-wise comparison matrices.

In the next stage, the super matrix M is made to converge to obtain a long-term stable set of weights.For convergence to occur, super matrix needs to be column stochastic, i.e. the sum total of each of the

columns of the super matrix needs to be one. Raising the super matrix M to the power 2kC1, where kis anarbitrarily large number, allows for the convergence of the interdependent relationships (Meade &Sarkis, 1999). In this example, convergence is reached at M55. The converged super matrix is shown in

Table 7.

8.8. Step 8. Selection of the best alternative for a determinant

The selection of the best alternative depends on the outcome of the desirability index. The

desirability index, Dia, for the alternative i and the determinant a is defined as (Meade & Sarkis, 1999)

DiaZXJ

jZ1

XKja

kZ1

PjaADkjaA

IkjaSikja; (1)

Table 5

Matrix for alternatives impact on enabler in influencing determinant

IT TPD SLC VRL e-vector

TPD 1 1/5 1/7 0.0692

SLC 5 1 1/4 0.2437

VRL 7 4 1 0.6871



20/30

Table 6

Super matrix M for legislation before convergence

CON CS GP CSF IT PRO CTM NTE COM MOS FSA KM

CON 0 0.0745 0.1947 0.2199

CS 0.0880 0 0.0881 0.0873

GP 0.7173 0.7471 0 0.6928

CSF 0.1947 0.1784 0.7172 0

IT 0 0.4837 0.5106 0.0848

PRO 0.6249 0 0.3728 0.7144CTM 0.2835 0.3487 0 0.2008

NTE 0.0916 0.1676 0.1166 0

COM 0 0.1383 0.6093 0.5936

MOS 0.2179 0 0.3112 0.3124

FSA 0.7247 0.6957 0 0.0940

KM 0.0574 0.1660 0.0795 0

WR

CSA

REV


21/30

Table 7

Super matrix M for legislation after convergence

CON CS GP CSF IT PRO CTM NTE COM MOS FSA KM

CON 0.1621 0.1621 0.1621 0.1621

CS 0.0807 0.0807 0.0807 0.0807

GP 0.4142 0.4142 0.4142 0.4142

CSF 0.3430 0.3430 0.3430 0.3430

IT 0.2996 0.2996 0.2996 0.2996

PRO 0.3549 0.3549 0.3549 0.3549CTM 0.2316 0.2316 0.2316 0.2316

NTE 0.1139 0.1139 0.1139 0.1139

COM 0.3149 0.3149 0.3149 0.3149

MOS 0.2150 0.2150 0.2150 0.2150

FSA 0.3857 0.3857 0.3857 0.3857

KM 0.0844 0.0844 0.0844 0.0844

WR

CSA

REV


22/30

where

Pja is the relative importance weight of dimension of reverse logistics j on the determinant of reverse

logistics a,ADkja is the relative importance weight for reverse logistics attribute enabler kof dimension of reverselogistics j in the determinant of reverse logistics control hierarchy networka for the dependency (D)

relationships between component levels,

AIkja is the stabilized relative importance weight (determined by the super matrix) for reverse logisticsattribute enabler k of dimension of reverse logistics j in the determinant of reverse logistics control

hierarchy network a for interdependency (I) relationships within the reverse logistics attributeenablers component level,

Sikja is the relative impact of reverse logistics implementation alternative i on reverse logistics

attribute enabler kof dimension of reverse logistics j of reverse logistics control hierarchy networka,

Kja is the index set of reverse logistics attribute enablers for dimension of reverse logistics j in forreverse logistics determinant control hierarchy a, and

J is the index set for the dimensions of reverse logistics (same for all control hierarchies).

Table 8 shows the desirability indices for the compatibility determinant (Di legislation). It is based on

the legislation hierarchy using the relative weights obtained from the pair-wise comparison of

alternatives, dimensions and weights of enablers from the converged super matrix.

These weights are used to calculate a score for the determinants of reverse logistics overall weighted

index (RLOWI) for each of the alternative being considered. In Table 8, the values of second column are

imported from Table 2, which are obtained by comparing the relative impact of the dimensions on the

legislation determinant. For example, in improving the legislation, the role of internal business

perspective is found to be most important (0.3931), which is followed by CP (0.2375), ILP (0.1889), and

Table 8

Legislation desirability indices

Dimensions Pja Enablers ADkja AIkja

S1kja S2kja S3kja TPD SLC VRL

CP 0.2375 CON 0.0952 0.1621 0.1865 0.1265 0.6870 0.0007 0.0005 0.0025

CP 0.2375 CS 0.0328 0.0807 0.1004 0.2256 0.6740 0.0001 0.0001 0.0004

CP 0.2375 GP 0.6593 0.4142 0.0692 0.2437 0.6871 0.0045 0.0158 0.0446

CP 0.2375 CSF 0.2127 0.3430 0.2569 0.2942 0.4489 0.0045 0.0051 0.0078

IBP 0.3931 IT 0.0993 0.2996 0.0692 0.2437 0.6871 0.0008 0.0029 0.0080

IBP 0.3931 PRO 0.2497 0.3549 0.6248 0.2834 0.0918 0.0218 0.0099 0.0032IBP 0.3931 CTM 0.4914 0.2316 0.0839 0.2110 0.7051 0.0038 0.0094 0.0315

IBP 0.3931 NTE 0.1596 0.1139 0.1007 0.2255 0.6738 0.0007 0.0016 0.0048

ILP 0.1889 COM 0.1399 0.3149 0.6307 0.2717 0.0976 0.0052 0.0023 0.0008

ILP 0.1889 MOS 0.1590 0.2150 0.3089 0.3584 0.3327 0.0020 0.0023 0.0021

ILP 0.1889 FSA 0.5421 0.3857 0.0670 0.6612 0.2718 0.0026 0.0261 0.0107

ILP 0.1889 KM 0.1590 0.0844 0.0549 0.2897 0.6554 0.0001 0.0007 0.0017

FP 0.1805 WR 0.6406 0.4576 0.2795 0.3570 0.3635 0.0148 0.0189 0.0192

FP 0.1805 CSA 0.0669 0.1335 0.2377 0.1551 0.6072 0.0004 0.0003 0.0010

FP 0.1805 REV 0.2925 0.4089 0.1130 0.6519 0.2351 0.0024 0.0141 0.0051

Total desirability indices 0.0644 0.1099 0.1435



23/30

FP (0.1805). The values in the fifth column of Table 8 are the stable independent weights of enablers

obtained through converged super matrix (Table 7). The next three columns are from the pair-wisecomparison matrices giving the relative impact of each of the alternatives on the enablers. The final three

columns represent the weighted values of the alternatives (Pja!ADkja!AIkja!Sikja) for each ofthe enablers. For the purpose of illustration, the value corresponding to TPD for CON is 0.0007(0.2375!0.0952!0.1621!0.1865Z0.0007). The summations of these results, for the legislation ofeach of these alternatives, are presented in the final row of Table 8. These results indicate that the VRL

with a value of 0.1435 has maximum influence on the legislation. It is followed by SLC (0.1099) andTPD (0.0644). Till this step, the analysis has been conducted only for the legislation determinant.Similar analysis is carried out for other three determinants. In the next step, an index would be calculated

to capture the achievement of overall goal of selecting an alternative.

8.9. Step 9. Calculation of reverse logistics overall weighted index (RLOWI)

The RLOWI for an alternative i (RLOWIi) is the summation of the products of the desirability indices(Dia) and the relative importance weights of the determinants (Ca) of the reverse logistics overall

weighted index. It is represented as:

RLOWIiZX

DiaCa

For example, the RLOWI for VRL is calculated as:

RLOWIVRLZ 0:2378!0:1687C 0:5433!0:1435C0:061!0:1697C 0:1579!0:1707

Z0:1554:

The final results are shown in Table 9.It is observed from Table 9 that VRL is the most-suited alternative for the reverse logistics operations

for the case company. SLC and TPD follow this alternative. It is observed from Table 9 that legislationplays a major role in the conduct of reverse logistics operations. It is also observed from the secondcolumn of this table that VRL (0.1687) is found to be more economic as compared to SLC (0.0821) andTPD (0.0423). The difference among these can probably be attributed to increased information

availability and other advanced IT capabilities which virtual reverse logistics network offers. These canbe used to diminish uncertainty on configuration, condition and place of origin, thus enabling better

planning and control of the reverse logistics networks. These results should be seen in the light of thecharacteristics of the case company and the inputs provided by its logistics manager in the pair-wise

comparison.Table 9

Reverse logistics overall weighted index (RLOWI) for alternatives

Alternatives ECO LEGI CC EGI RLOWI Normalized

values for

RLOWI

Weights 0.2378 0.5433 0.061 0.1579

TPD 0.0423 0.0644 0.0717 0.0717 0.0607 0.1894

SLC 0.0821 0.1099 0.1069 0.1189 0.1045 0.3260

VRL 0.1687 0.1435 0.1697 0.1707 0.1554 0.4846



24/30

9. Discussion and managerial implications

In this section, we first discuss the results of the model. Later, we present few suggestions to the

prospective users of this model. Finally, we discuss the managerial implications of this model and somegeneralization of results.

The major contribution of this research lies in the development of a comprehensive model, whichincorporates diversified issues for conducting reverse logistics operations for EOL computers. It

considers a balanced view on four perspectives namely customers, internal business, innovation andlearning, and financial for the conduct of reverse logistics. This is similar to the balanced scorecard

proposed by Kaplan and Norton (1992) who had emphasized the measurement of the performancesystems based on these four perspectives. The proposed ANP model in this paper, not only guides the

decision makers for the efficient conduct of reverse logistics operations but also enable them to visualizethe impact of various criteria in the arrival of the final solution. Further, the interdependencies among the

various criteria can be effectively captured using the ANP technique, which has rarely been applied inthe context of the conduct of reverse logistics on EOL computers.

For the case undertaken in this study, the results indicate that VRL is the first choice of the casecompany, which is followed by SLC and TPD. The choice of the case company towards VRL may be

attributed to the advanced IT, supply chain solutions and change management capabilities. In thecomputer industry, IT-enablement and distribution planning can lead to enhanced competitiveness. It is

relevant to discuss here the priority values of the determinants of reverse logistics, which influence thisdecision. From Table 1, it is seen that legislation (CaZ0.5433) is the most important determinant for the

conduct of reverse logistics operations. Economic factors (0.2378), environment and green issues(0.1579), and corporate citizenship (0.0610) follow it. In fact legislation and the economic factors put

together had virtually determined the conduct of reverse logistics operations for EOL computers. These

implications are straightforward as the legislation and enactment of laws like the extended producerresponsibility has forced the companies to reuse the products and also seek secondary markets for theirproducts after the end of the product life cycle. Virtual reverse logistics networks enable higher visibilityon products data, even before products enter the return flows. Also, information flows in virtual reverse

logistics networks enable coordination among multi-echelons of reverse logistics networks and take

advantage of economies of scale for transportation. The results also favor to environmental and greenissues. By giving importance to the environmental and green issues in their production practices, thecompany is assuming the role of a corporate citizen for a social cause.

Table 9 shows the RLOWI for the alternatives. It is observed from the table, that VRL excels over the

other two alternatives for all the four determinants. It is also observed that VRL is found to be more

economical than SLC and TPD. Though, in the illustrated example, the model has been described forthree distinct alternatives, it can accommodate more than three at the cost of complexity. Further, thesame model presented here can be used for comparison when these alternatives belong to only one

category or from two different categories. In the light of the results obtained for the case company, it is tobe noted that the results obtained are valid for the case company in its own decision environment.

Therefore, these results do not establish the supremacy of one alternative over another. Moreimportantly, it is the decision environment of the user, which makes one alternative superior to other.

Though the proposed ANP model is based on a sound algorithm for systemic decision-making, care mustbe taken in its application, the reason being that the user has to compare the reverse logistics operations

on a number of pair-wise comparison matrices.



25/30

Despite illustrative example deriving data set from a case company some generalization of results are

possible. The corporate environment is fast moving towards an environment-conscious supply chain andreverse logistics is an obvious candidate for being managed better. Information technology and

information sharing are of prime importance in the effectiveness of reverse logistics. Virtual logisticsnetworks provide superiority due to this reason. Products like PCs, which are reconfigurable need to besuitably managed in the reverse logistics. The criteria and dimensions identified in the proposed modelare quite generic and with marginal adjustments can be used for different product also.

10. Conclusion

The reverse logistics practices may cost in millions of dollars for company. The implementation ofthese may be a risky endeavor for the top management as it involves financial and operational aspects,

which can determine the performance of the company in the long run. However, with the legislativemeasures tightening up, there are not many options. The question now is not whether to go for it or notbut which framework to pick up. This research is relevant in this sense. The ANP model presented in this

paper structured the problem of conduct of reverse logistics for EOL computers in a hierarchical formand linked the determinants, dimensions, and enablers of the reverse logistics and the alternativesavailable to the decision maker for a computer industry. It can aid the top management in the evaluation

of the various alternatives available with them as it measures the relative strengths of impacts betweenelements in the hierarchical model.

Previously, the firms had overemphasized short-term financial performance. But no single measure

can provide a clear performance focus on the critical areas of business ( Kaplan & Norton, 1992). Thesame can be applicable to the conduct of the reverse logistics operations, as sole emphasis on the

finance (economic) aspect would not provide a complete picture of the real situation. The modelpresented in this paper gives a holistic view of the various criteria affecting reverse logistics

operations for EOL computers. It is holistic in the sense that it inherits the principle of balancedscorecard, which measures the performance of a firm in the four dimensions of customer, internalbusiness, innovation and learning, and finance (Kaplan & Norton, 1992). Although the concept of

balanced scorecard has been primarily designed for the measurement of the performance, we in thismodel have used these dimensions, in the dimension hierarchy of the ANP model in order to obtain aholistic framework towards the conduct of reverse logistics operations. It thus provides the decision

makers with a balanced framework for reverse logistics for the conduct of reverse logisticsoperations, thus enhancing the value and clarity of the decision-making process needed by the top

management.Thus, a combination of balanced scorecard and ANP approach proposed in this paper can provide tothe decision maker a more realistic and accurate representation of the problem for conducting reverse

logistics operations for EOL computers. A major contribution of this research study lies in its linkage ofvarious issues of the reverse logistics in a single systemic framework. It is an attempt in this regard to aid

the decision makers in the complex task of prioritizing their options. This decision model integrates andrelies upon the various characteristics of reverse logistics of determinants, dimensions, enablers and theirrelationships. The utility of the ANP methodology in integrating both quantitative as well as the

qualitative characteristics, which need the attention of the decision maker in arriving at the best possiblesolution, assumes tremendous value.



26/30

The model developed in this paper has a few limitations as well. The formation of the pair-wise

comparison matrices and data acquisition is a tedious and time-consuming task. In this case, 141pair-wise matrices are formed. Also, more importantly, the results reported in this research are based

on the opinion of the logistics manager of the case company. Thus, the pair-wise comparison of thecriteria always depends on the users knowledge and familiarity with the firm, its operations, and itsindustry. Therefore, the biasing of the manager to some criteria might have influenced the results.Although, we have tried to minimize this by checking the consistency of comparison using method

of consistency-ratio check as suggested by Saaty (1980). Hence, the identification of the relevantattributes to the problem under consideration, the determination of their relative importance incomparison to others require extensive brainstorming sessions, and the accumulation of expertise and

knowledge within the organization. Since many of the issues in the pair-wise comparisons are cross-functional in nature, a team of managers from various functional departments should be assigned the

responsibility of comparison. Delphi method may also be a promising technique that may be

explored in this regard.The logistics manager in this case computer company considered the approach to lead to an objectiveanalysis of the situation and is currently implementing the virtual reverse logistics networks for PCs asan approach to the conduct of reverse logistics operations for EOL computers in their organization. The

balanced view of all the four dimensions of the customer, internal business, innovation and learning, andfinancial with their enablers in the reverse logistics has enhanced the clarity of the decision making bythe logistics manager.

A possible extension of this research study might be to study the preferences of the user companiescorresponding to different sizes and sectors, where these criteria may be modeled as per the choice ofcompanies. The model may also be subjected to sensitivity analysis. User-friendly software may also be

developed on the basis of the model.

References

Achrol, R. S., Reve, T., & Stern, L. (1983). The environment of marketing channel dyads: A framework for comparative

analysis. Journal of Marketing, 47, 5567.

Adler, L. (1966). Symbiotic marketing. Harvard Business Review, 44, 5971.

Agarwal, A., & Shankar, R. (2003). On-line trust building in e-enabled supply chain. Supply Chain Management: An

International Journal, 8(4), 324334.

Andel, T. (1997). Reverse logistics: A second chance to profit. Transportation and Distribution, 38(7), 6164.

Bacallan, J. J. (2000). Greening the supply chain. Business and Environment, 6(5), 1112.

Bakos, Y. (1998). The emerging role of electronic marketplaces on the Internet. Communications of the ACM, 41, 3542.Barnes, J. (1982). A problem in reverse logistics. Journal of Macromarketing, 2(2), 3137.

Bertagnoli, L. (2000). Computers get second chance. Crains Chicago Business, October 2, 23.

Blumberg, D. (1999). Strategic examination of reverse logistics and repair service requirements, needs, market size, and

opportunities. Journal of Business Logistics, 20(2), 141159.

Boon, J. E., Isaacs, J. A., & Gupta, S. M. (2002). Economic sensitivity for end of life planning and processing of personal

computers. Journal of Electronics Manufacturing, 11(1), 8193.

Bowersox, D. (1990). The strategic benefits of logistics alliances. Harvard Business Review, 68(4), 3645.

Brewer, P. C., & Speh, T. W. (2000). Using the balanced scorecard to measure supply chain performance. Journal of Business

Logistics, 21(1), 7593.

Byrne, P. M., & Deeb, A. (1993). Logistics must meet the green challenge. Transportation and Distribution, 34(2), 3337.



27/30

Cairncross, F. (1992). How Europes companies position to recycle. Harvard Business Review, 70, 3337.

Carter, C., & Ellram, L. (1998). Reverse logistics: A review of literature and framework for future investigation. Journal of

Business Logistics, 19(1), 85102.

Cohen, M. (1988). Replace, rebuild or remanufacture. Equipment Management, 16(1), 2226.

Cooke, J. A. (1992). Its not easy being green. Traffic Management, 37(12), 4247.

Coppel, J. (2000). E-commerce: Impacts and Policy Challengers. OECD Economics Department, Working Paper No.252,

Organization for Economic Co-operation and Development OECD, June, http://www.oecd.org/dataoecd/42/48/2087433.

pdf

Daugherty, P. J., Myers, M. B., & Richey, R. G. (2002). Information support for reverse logistics: The influence of relationship

commitment. Journal of Business Logistics, 23(1), 85106.

de Brito, M. P., & Dekker, R. (2003). A framework for reverse logistics. ERIM report series research in Management, https://ep.

eur.nl/retrieve/439/ERS-2003-045-LIS.pdf

Deere & Company (1998), Deere & Company and Springfield Manufacturing Corporation announce joint venture.

PRNewswire, Deere & Company.

Degher, A. (2002). HPs worldwide take back and recycling programs: Lessons on improving program implementation. IEEE

international symposium on electronics and the environment on 69 May 2002 (pp. 224227).

Dowlatshahi, S. (2000). Developing a theory of reverse logistics. Interfaces, 30(3), 143154.

Farrow, P. H., Johnson, R. R., & Larson, A. L. (2000). Entrepreneurship, innovation, and sustainability strategies at Walden

Paddlers, Inc. Interfaces, 30(3), 215225.

Ferguson, N., & Browne, J. (2001). Issues in EOL product recovery and reverse logistics. Production Planning and Control,

12(5), 534547.

Ferguson, R. B. (2000). IBM Dell move to reverse logistics. e-Week, 17(40), 20.

Fleischmann, M., Bloemhof-Ruwaard, J. M., Dekker, R., van der Laan, E., van Nunen, J. A. E. E., & Van Wassenhove,

L. N. (1997). Quantitative models for reverse logistics: A review. European Journal of Operational Research, 103,

117.

Godfrey, R. (1998). Ethical purchasing: Developing the supply chain beyond environment. In T. Russel (Ed.), Greener

purchasing: Opportunities and innovations (pp. 244251). Sheffield: Greenleaf Publishing, 244251.

Grenchus, E., Johnson, S., & McDonell, D. (2001). Improving environmental performance through reverse logistics at IBM.

Proceedings of the 2001 IEEE international symposium on 79 May 2001, Denver, CO, USA (pp. 236240).Grenchus, E., Keene, R., & Nobs, C. (1997). Demanufacturing of information technology equipment. Proceedings of IEEE

symposium on electronics and the environment on 57 May 1997, Sanfrancisco, USA (pp. 157160).

Guide, V. D. R., & Srivastava, R. (1998). Inventory buffers in recoverable manufacturing. Journal of Operations Management,

16, 551568.

Guide, V. D. R., & Wassenhove, L. N. (2003). Full cycle supply chains. USA: Carnegie Mellon Press.

Gungor, A., & Gupta, S. M. (1999). Issues in environmentally conscious manufacturing and product recovery: A survey.

Computers and Industrial Engineering, 36(4), 811853.

Hamilton, A. (2001). How do you junk your computer. Time, 157(6), 7071.

Handfield, R. B., & Nichols, E. L. (1999). Introduction to supply chain management. Englewood Cliffs, NJ: Prentice-Hall.

Hans, H., & Byrne, P. M. (1993). Reducing waste through reverse logistics. Foundry Management and Technology, 121(8),

2829.

Hart, S. L. (1997). Beyond greening: Strategies for a sustainable world. Harvard Business Review, 75(1), 6776.Heeb, H. B. (1989). Can old workhorses still pull their weight? Manufacturing Engineering, 102(4), 7476.

Heide, J. B., & John, G. (1990). Alliances in industrial purchasing: The determinants of joint action in buyersupplier

relationships. Journal of Marketing Research, 27, 2436.

Heide, J. B., & John, G. (1992). Do norms matter in marketing relationships? Journal of Marketing, 56, 3244.

Hillegersberg, J., Zuidwijk, R., van Nunen, J., & van Eijk, D. (2001). Supporting return flows in the supply chain.

Communications of the ACM, 44(6), 7479.

Hines, F., & Johns, R. (2001). Environmental supply chain management: Evaluating the use of environmental mentoring

through supply chain Greening of Industry Network Conference on January 2125, Bangkok.

Hirschorn, L., & Gilmore, T. (1992). The new boundaries of the boundaryless company. Harvard Business Review, 70(3),

104115.

http://www.oecd.org/dataoecd/42/48/2087433.pdfhttp://www.oecd.org/dataoecd/42/48/2087433.pdfhttp://https//ep.eur.nl/retrieve/439/ERS-2003-045-LIS.pdfhttp://https//ep.eur.nl/retrieve/439/ERS-2003-045-LIS.pdfhttp://https//ep.eur.nl/retrieve/439/ERS-2003-045-LIS.pdfhttp://https//ep.eur.nl/retrieve/439/ERS-2003-045-LIS.pdfhttp://www.oecd.org/dataoecd/42/48/2087433.pdfhttp://www.oecd.org/dataoecd/42/48/2087433.pdf


28/30

Jacoby, J., Berning, C., & Diettvorst, T. (1977). What about disposition? Journal of Marketing, 41(4), 2328.

Jayaraman, V., Guide, V. D. R., & Srivastava, R. (1999). A closed-loop logistics model for use within a recoverable

manufacturing environment. Journal of Operational Research Society, 50(5), 497509.

Kaplan, R. S., & Norton, D. P. (1992). The balanced scorecardMeasures that drive performance. Harvard Business Review,70(1), 7179.

Kelle, P., & Silver, E. A. (1989). Forecasting the returns of reusable containers. Journal of Operations Management, 8,

1735.

Knemeyer, A. M., Ponzurick, T. G., & Logar, C. M. (2002). A qualitative examination of factors affecting reverse logistics

systems for end-of-life computers. International Journal of Physical Distribution and Logistics Management, 32(6),

455479.

Kokkinaki, A. I., Dekker, R., de Koster, M. B. M., Pappis, C., & Verbeke, W. (2001). From e-trash to e-treasure:

How value can be created by the new e-business models for reverse logistics, https://ep.eur.nl/retrieve/1010/

feweco20010219160545.pdf

Kokkinaki, A. I., Dekker, R., Lee, R., &

Documents

Analyzing Alternatives in Reverse Logistics for End of Life Computers With ANP and BSC