Sunteți pe pagina 1din 14
Int. J. Production Economics 158 (2014) 77 – 90 Contents lists available at ScienceDirect Int. J.

Contents lists available at ScienceDirect

Int. J. Production Economics

journal homepage: www.elsevier.com/locate/ijpe

Economics journal homepage: www.elsevier.com/locate/ijpe The impact of barriers and bene fi ts of e-procurement on

The impact of barriers and bene ts of e-procurement on its adoption decision: An empirical analysis

Peral Tokta ş -Palut a , n , Ecem Baylav a , Seyhan Teoman b , Mustafa Altunbey c

Baylav a , Seyhan Teoman b , Mustafa Altunbey c a Department of Industrial Engineering, Do

a Department of Industrial Engineering, Do ğ u ş University, 34722 Istanbul, Turkey

b Logistics and Supply Chain Management Doctorate Program, Maltepe University, 34857 Istanbul, Turkey

c Logistics and Supply Chain Management Doctorate Program, Do ğ u ş University, 34722 Istanbul, Turkey

article info

Article history:

Received 25 April 2013 Accepted 25 June 2014 Available online 23 July 2014

Keywords:

E-procurement Barriers Bene ts Interpretive structural modeling Structural equation modeling

abstract

The aim of this study is to investigate the effects of the barriers and bene ts of e-procurement systems on the e-procurement adoption decision. An empirical analysis is performed for a retail store chain which operates in book and stationery sector in Turkey. Interpretive Structural Modeling (ISM) technique is applied to the selected barriers and bene ts to determine their contextual relationships. Inadequate IT infrastructure of suppliers/business partners is found to be the most important barrier in the adoption of e-procurement systems. On the other hand, the most signi cant bene t is determined as integrated information sharing. The barriers and bene ts, which have high driving power and the capability to in uence the other drivers, are then integrated into the structural equation model. The results denote that barriers (bene ts) of e-procurement systems have negative (positive) effect on the e-procurement adoption decision for the company. In addition, it is found that the effect of bene ts on the adoption decision is higher than that of the barriers. Thus, based on this analysis, it would be bene cial for the company to adopt the e-procurement system.

& 2014 Elsevier B.V. All rights reserved.

1. Introduction

Intensifying competition in today's business environment has highlighted the need to optimize the management of supply chains. Developments in the information and communication technology are important tools to manage supply chains effectively. In recent years, an increasing trend has been observed in the adoption of e-procurement systems, which help in the integration of the procurement process throughout the supply chain. The term e-procurement refers to the integration of procure- ment process, which includes operations such as negotiation, ordering, receipt, and post-purchase review ( Croom and Brandon-Jones, 2007 ). Morris et al. (2000) de ne e-procurement as a system that utilizes Internet technologies and services to automate and streamline an organization's processes from requisition to payment. E-procurement is not a recently discovered process as there have been many attempts of using electronic systems for devel- oping an automated procurement technology for organizations, such as electronic work ow systems and EDI ( Gunasekaran and Ngai, 2008 ). When the traditional procurement system is ana- lyzed, it can be seen that the process is implemented by using

n Corresponding author. Tel.: þ 90 216 544 5555; fax: þ 90 216 544 5535. E-mail address: ppalut@dogus.edu.tr (P. Tokta ş -Palut).

0925-5273/ & 2014 Elsevier B.V. All rights reserved.

phone, fax, and other communication channels, showing how the procurement system is developed over time ( Hawking et al.,

2004 ).

E-procurement, which has become one of the fundamental elements of a supply chain, is still in the process of development and several studies are performed on this subject. Gunasekaran and Ngai (2008) state that the supply chain of a company cannot be integrated successfully without the adoption of e-procurement systems. The authors conduct a questionnaire based survey in order to understand the adoption process of e-procurement in Hong Kong. Davila et al. (2003) carry out a research on the equilibrium point for the optimum e-procurement system that should be used by companies which have different attitudes towards technology. Panayiotou et al. (2004) work on a case study about the Greek purchasing process and indicate the problems which may occur. The authors also study e-procurement system design. Croom and Brandon-Jones (2007) assess the validity of the forecasts performed by earlier studies which made predictions about the changes and improvements that will be caused by e-procurement systems. Gunasekaran et al. (2009) analyze the current state of e-procurement in SMEs located in the Southcoast of Massachusetts, and they also examine the factors that affect the e-procurement adoption. Nguyen (2013) performs an empirical study for the transport and logistics companies in Australia, and determines the principal components affecting the e-business adoption decision.

78

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

In the light of the previous studies on e-procurement systems, this study focuses on the barriers and bene ts of e-procurement systems. Our aim is to investigate the effects of the barriers and bene ts on the e-procurement adoption decision. For this purpose, rst Interpretive Structural Modeling (ISM) is applied to the barriers and bene ts of e-procurement in order to understand the relations within these criteria. Talib et al. (2011) de ne ISM as an interpretive model, which de nes the relations between variables in complex systems by representing them in a hierarchical con guration. ISM is widely used to specify the interactions between the elements of a system, such as barriers and bene ts. Gorvett and Liu (2006) assess the risks in a rm by using ISM and analyze the risk pro le of the company. Singh and Kant (2008) use ISM in order to understand the relations between the barriers of knowledge management and specify the driving and dependent barriers. Chandramowli et al. (2011) perform a case study about the barriers of developing land ll sites in order to realize urban planning issues. Luthra et al. (2011) apply ISM to investigate the relations between the barriers of green supply chain management for India's automobile industry. The literature review shows that this is the rst study that applies ISM to the barriers and benets of e-procurement systems. In the second part of the study, the results of the ISM model are validated by using Structural Equation Modeling (SEM). Another point of interest is to analyze whether the effects of barriers or bene ts are higher on the e-procurement adoption decision. For this purpose, we perform an empirical analysis for a retail store chain which operates in book and stationery sector in Turkey. To the best of our knowledge, this is the rst study that integrates ISM and SEM techniques. SEM can be used for multivariate data as it is appropriate for illustrating the relations between exogenous and endogenous latent variables in one model ( Kline, 1998; Aibinu and Al-Lawati, 2010 ). SEM is a widely used tool especially for psychology, sociology, and econometrics ( Golob, 2003 ) and there are several studies applying SEM in the literature. Thus, we only mention the studies in a related eld. Madeja and Schoder (2003) examine the impacts of e-procurement adoption process on e-business success. Ho et al. (2004) investigate the direct and performance impacts of e-procurement on supply chain orientation from the buyerseller perspective. Lee and Quaddus (2006) conduct a study in Singapore about the impacts of buyersupplier relationship on e-purchasing adoption decision. Vaidyanathan and Devaraj (2008) study the impacts of the quality of the information ow process and the quality of logistics fulllment on the satisfaction of e-procurement performance. Konradt et al. (2012) analyze the effects of organizational and technical antecedence on usability and commercial transactions in B2B e-commerce. Finally, Devaraj et al. (2012) investigate the effects of mixing exibilities and purchase volumes on e-procurement performance. The rest of the paper is organized as follows. The research background and hypotheses are given in Section 2 . The methodol- ogy used in this study is explained in Section 3 . Section 4 presents the data analysis and results of the ISM model, measurement model, and structural model. Finally, concluding remarks are given in Section 5 .

2.

Theoretical background and research hypotheses

2.1.

Theoretical background

Procurement is a vital issue in the retail sector and basic supply chain activities start with procurement. Hence, procurement directly affects the performance and ef ciency of the supply chain. In this study, we consider a retail store chain which operates in

book and stationery sector in Turkey. The company has an annual procurement volume of approximately 100 million USD, which covers its demand for more than 20,000 different stock keeping units provided by 185 suppliers. Currently, this company uses traditional procurement methods, but the senior management wants to decide whether to adopt an e-procurement system or not. As one would expect, there are some barriers to switching to an electronic procurement system for the company. On the other hand, the company will experience some bene ts from the e-procurement system when compared with the traditional pro- curement methods. The literature reviews on the barriers and bene ts of the e-procurement systems are given in Tables 1 and 2 , respectively. Based on the literature review, 20 barriers and 20 bene ts are determined. Then, these barriers and bene ts are discussed with eight top managers of the company and their numbers are decreased to 14 barriers and 15 bene ts. The barriers and bene ts used in this study and their explanations are given below.

2.1.1. Barriers of e-procurement systems

Although it is known that e-procurement systems have several bene ts for the company, the company would also face some barriers while switching to an e-procurement system. The barriers of e-procurement systems taken into consideration in this study are given below:

2.1.1.1. Bureaucratic disfunctionalities in practice. There are a number

of specic regulations and standards that have been developed for e-procurement applications which require that a bureaucratic procedure to be followed due to the nature of the legal institutions involved and embraces audit, accountability and compliance standards with national and international rules to ensure supply competition and transparency in the awarding of contracts.

2.1.1.2. Cost/benet concern. Emerges where the expenses outweigh

the benets of moving to electronic procurement.

2.1.1.3. External (supply chain/business partners) incompatibi-

lity. Suppliers' strategies and willingness are not sufcient to take

on new technologies such as e-procurement. There is a lack of readiness by external parties to engage in electronic interactions.

2.1.1.4. High investment cost of IT infrastructure/software. The company

cannot afford too high investment cost on a new IT infrastructure and

software which are necessary for e-procurement adoption.

2.1.1.5. Inadequate business processes to support e-procure-

ment. Specied aims, goals and context of the e-procurement application are not aligned with other organizational and managerial choices and not integrated with the organization's processes.

2.1.1.6. Inadequate IT infrastructure of suppliers/business partners. The

external parties of the supply chain do not have adequate IT infrastructure compatible with the e-procurement system.

2.1.1.7. Incompliance with company culture. Cultural issues such as

absence of a clear corporate policy, lack of a widely accepted solution, lack of leadership, lack of exible centralized control, resistance to change, and lack of information quality, produce a slowdown in adoption of e-procurement.

2.1.1.8. Inter-operability concerns with other systems used. It is a

technical issue related to lack of compatibility and interface with

other internal systems, mainly due to the fact that software companies have sought to make their product unique. In doing

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

Table 1 Literature review on barriers of e-procurement systems.

79

Barriers of e-procurement

References

Eadie et al.

Gunasekaran and Ngai

Eadie et al.

Farzin and Nezhad

Sitar

Eei et al.

This

(2007)

(2008)

 

(2010)

(2010)

(2011)

(2012)

study

Bureaucratic disfunctionalities in practice Cost/bene t concern External (supply chain/business partners ) incompatibility High investment cost of IT infrastructure/software Inadequate business processes to support e-procurement Inadequate IT infrastructure of suppliers/business partners Incompliance with company culture Inter-operability concerns with other systems used Lack of adequate technical/IT infrastructure Lack of e-procurement knowledge/skilled personnel Lack of exibility in process and documentation Lack of system integration with suppliers/business partners Lack of top management support Religious objections to internet Resistance to change of internal/external customers in supply chain Security, con dentiality, and authentication concerns Time needed for the implementation process Unsure as to the legal position of e-procurement

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

 

X

XX

X

X

X

X

X

X

X

X

 

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

Table 2 Literature review on bene ts of e-procurement systems.

 

Bene ts of e-procurement

References

Morris

Knudsen

Panayiotou

Wong

Eadie

Gunasekaran

Eadie et al.

Farzin and

Eei

This

et al.

(2003)

et al.

and

et al.

and

(2010)

Nezhad

et al.

study

(2000)

(2004)

Sloan

(2007)

Ngai (2008)

(2010)

(2012)

 

(2004)

Better management and control of suppliers Better utilization of staff Compliance with laws and regulations Cost savings in overall purchasing process Decentralization of power Decreased bureaucracy and redundancy Easier access to market data and enhanced intelligence Enhanced decision making Enhanced inventory management Improved communication and collaboration in supply chain Improved supply chain transparency Increased customer service levels Increased process quality and ef ciency Integrated information sharing Minimization of process errors On-line and real-time reporting

 

X

X

X

X

X

X

X

 

XX

X

X

 

X

X

X

X

X

X

X

X

X

X

X

 

X

 

X

X

X

X

 

X

X

X

X

X

XX

X

X

X

X

X

X

X

X

 

X

X

XX

X

X

X

 

X

X

X

X

X

X

 

X

X

X

X

X

X

 

X

X

X

 

X

X

X

X

X

X

X

X

Quicker response to problems through real-time X information

 

X

X

Reduced administration cost Reduced paperwork Reduction in processing time Simpli ed and streamlined purchasing process Standardization of process Time savings in overall purchasing process Wider range of suppliers

 

X

X

X

 

X

X

X

X

X

X

X

X

X

X

X

X

X

X

 

X

X

X

X

X

X

X

X

X

X

X

 

X

so, they have endeavored to stop migration of data between systems.

2.1.1.9. Lack of adequate

adequate IT infrastructure to carry out e-procurement processes and/or technology to operate IT.

technical/IT infrastructure. Need of

2.1.1.10. Lack of e-procurement knowledge/skilled personnel. Mainly

related to personnel issues such as older generations that have not kept up to the advances in IT related elds, but relying heavily on traditional forms and means of procurement.

2.1.1.11. Lack of exibility in process and documentation. Insuf cient

exibility of e-procurement systems can prevent system use in

80

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

certain circumstances and make exception handling necessary, and as a result jeopardize the expected bene ts.

2.1.1.12. Lack of system integration with suppliers/business part-

ners. Dif culties in integrating e-systems across rm boundaries in supply chains if suppliers lack capability.

2.1.1.13. Resistance to change of internal/external customers in supply

chain. Resistance to change is an inevitable aspect of human nature. Since the system users are generally unwilling to change their way of working to which they are accustomed, that might be a strong barrier in adopting e-procurement.

2.1.1.14. Security, con dentiality, and authentication concerns. The

e-procurement platform transacts con dential procurement data

and is exposed to several security threats. It would bring the dif culties with authentication and security into play.

2.1.2. Bene ts of e-procurement systems

E-procurement is a systematic way for increasing productivity and decreasing expenditures for businesses. With e-procurement, companies can gain some bene ts such that they improve their business ability and make their supply chain operations more observable. The bene ts of e-procurement systems considered in this study are given below:

2.1.2.1. Better management and control of suppliers. Provides buyers

to locate the suppliers with the best prices and quality, and

streamline the negotiation and contracting processes through enhance transparency and communication.

2.1.2.2. Cost savings in overall purchasing process. Increase in speed

and ef ciency in procurement process and communication, increased level of inventory-cycle, less need of operational personnel, savings on operating costs.

2.1.2.3. Decreased bureaucracy and redundancy. E-procurement

facilitates and speeds up internal control processes. It also eliminates inefcient approval procedures and needless duplication of services.

2.1.2.4. Easier access to market data and enhanced intelligence. Enables

to monitor and scan external sources of data and intelligence easily and to share information with others pro-actively.

2.1.2.5. Enhanced decision making. The decision-making process is

enhanced by e-procurement since relevant information is neatly organized and time-stamped.

2.1.2.6. Enhanced inventory management. Automated inventory

control and standardized procedures make inventory operations

easier to be managed, adjusted, and upgraded.

2.1.2.7. Improved communication and collaboration in supply chain.

E-procurement allows sections of electronic documentation to ow through the supply chain. As it improves the speed of returns and makes it easy to communicate requirements in a quicker and more accessible manner, it will result in a better understanding of requirements and due compliance.

2.1.2.8. Improved supply chain transparency. Transparency of

product speci cations, prices, contract details, such as the contractual conditions, time, terms of orders, etc., making these visible to relevant parties both internally and externally.

2.1.2.9. Increased process quality and ef ciency. Transparency in system and increased competition among suppliers through easier communication, intelligence and information access provide improvements in procurement process quality and efciency.

2.1.2.10. Integrated information sharing. Accelerates the ow of

important information between internal and external business partners; also provides real-time information sharing within a broader structure.

2.1.2.11. Minimization of process errors. Eliminating manual pro-

cesses and paperwork reduces the opportunity for human errors,

inaccuracies, and reworks.

2.1.2.12. On-line and real-time reporting. Real-time reporting

system that enables management to have a fast and reliable way to compare the spending with budget, allowing quick reaction to problems.

2.1.2.13. Quicker response to problems through real-time informa-

tion. Increased speed in transactions, tracking and reporting

through real-time information helps quicker problem solving and reactive decisions.

2.1.2.14. Simplied and streamlined purchasing process. E-procurement

solutions simplify the purchasing process by bringing all suppliers together, accessible from a single e-platform and eliminating the need for a paper form, while providing streamlined procedures to expedite order to payment processing.

2.1.2.15. Time savings in overall purchasing process. Simpli ed and

streamlined process; increase in information speed and transparency enables shorter time for evaluation and the decision-making process.

2.2. Research hypotheses

This study consists of two main parts. In the rst part, the aim is to analyze the interactions among the barriers and bene ts of e-procurement systems, and to determine the most signi cant ones. Then, the effects of these barriers and bene ts on the e-procurement adoption decision are investigated in the second part of the study. We have two main research hypotheses as follows:

Hypothesis 1 (H1). Barriers of e-procurement systems have nega- tive effect on the adoption decision.

Hypothesis 2 (H2). Bene ts of e-procurement systems have positive effect on the adoption decision.

In addition to the hypotheses given above, another point of interest is to understand whether the effects of the barriers or bene ts are higher on the e-procurement adoption decision. Accordingly, we may recommend the company to adopt the e-procurement system or not.

3. Methodology

3.1. Sampling and data collection

As stated before, we perform an empirical analysis in this study and consider a retail store chain which operates in book and stationery sector in Turkey. The senior management wants to decide whether to switch from traditional procurement methods to an e-procurement system or not. For this purpose, a detailed literature review is performed to de ne the perceived barriers and

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

81

bene ts of e-procurement systems. 20 barriers and 20 bene ts are determined as a result of the literature review. In the second step, eight top managers from the company have rated these barriers and bene ts on a ve-point scale. By taking the geometric mean ( Aczél and Roberts, 1989 ) of their responses and after a brainstorming session, we get the nal set of barriers and bene ts that are rated above three. In the nal set, there are 14 barriers and 15 bene ts left, which are then entered into the ISM model. In the third step, to identify the contextual relationships within the barriers and bene ts of e-procurement systems, another brainstorming session is performed with eight top managers from the company and four experts from academia. Based on the outputs of this session, the ISM models for the barriers and bene ts of e-procurement systems are developed.

According to the results of the ISM models, 3 barriers (7 benets) of e-procurement systems are found to be the root causes of the other barriers (benets). These barriers and benets are then entered into the SEM model forming the fourth step of the study. In this step, another questionnaire is prepared to develop the SEM model. This questionnaire is also prepared on a ve-point scale. Here, the aim is to investigate the effects of these barriers and benets on the e-procurement adoption decision and the necessity of the e-procurement system for the company. The questionnaire is uploaded to the Intranet of the company. Among 916 people working in a related eld, 277 respondents answered the questionnaire, resulting in a response rate of 30.24%. Among these responses, 21 cases contain missing values, giving a percentage of 7.58%. On the other hand, regarding the variables, all variables except one contain missing values ranging between

Literature review List the barriers and benefits for e- procurement adoption process Define the research
Literature review
List the barriers and
benefits for e-
procurement adoption
process
Define the research
hypotheses
Expert opinion
Establish contextual
Test the measurement model using
confirmatory factor analysis
relationship between
variables
• Check model goodness-of-fit
• Check convergent validity
• Check discriminant validity
Develop structural
• Examine the standardized residual
covariances and modification indices
Develop reachability
self-interaction
No
matrix
No
matrix
Partition the
reachability matrix
into different levels
Is the measurement
model validated?
Yes
Develop the
reachability matrix
in its conical form
Test the structural model
• Check model goodness-of-fit
Draw a directed
graph
• Analyze the path coefficients
and factor loadings
Remove transitivity
from the digraph
Is the structural model
validated?
Replace variable
nodes with
Yes
relationship
statements
Analyze the effects of
barriers and benefits on e-
procurement adoption
decision
Yes
Is there any conceptual
inconsistency?
No
Obtain the base elements
of barriers and benefits
for the e-procurement
adoption process

INTERPRETIVE STRUCTURAL MODELING

STRUCTURAL EQUATION MODELING

Fig. 1. The methodology used in this study (ISM part is adapted from Ravi and Shankar (2005) ).

82

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

0.72% and 3.61%. Since sample size is especially important for SEM, Bayesian imputation ( Schafer and Graham, 2002 ) is performed to estimate the missing values. Readers are referred to Schafer and Graham (2002) and Chen and Âstebro (2003) for the advantages of Bayesian imputation over other techniques such as listwise dele- tion and regression imputation.

3.2. Interpretive structural modeling

This study combines two techniques: ISM and SEM. We will discuss ISM in this subsection, and SEM in the following one. ISM is found and handled by War eld (1973) and its roots come from graph theory. The ISM process transforms unclear, poorly articu- lated mental models of systems into visible, well-de ned models useful for many purposes ( Sage, 1977 ). Ravi and Shankar (2005) describe the signi cant characteristics of ISM as follows: (i) This methodology is interpretive as the judgment of the group decides whether and how different elements are related; (ii) it is structural as on the basis of the relationships, an overall structure is extracted from the complex set of elements; and (iii) it is a modeling technique as the speci c relationships and overall structure are portrayed in a digraph model. The steps involved in the ISM methodology are given below ( Ravi and Shankar, 2005 ; Govindan et al., 2012 ):

Step 1. Variables affecting the system under consideration are listed. Step 2. A contextual relationship is established among variables with respect to which pairs of variables would be examined. Step 3. A Structural Self-Interaction Matrix (SSIM) is developed for variables, which indicates pairwise relationships among variables of the system under consideration. Step 4. Reachability matrix is developed from the SSIM and the matrix is checked for transitivity. The transitivity of the con- textual relation is a basic assumption made in ISM. It states that if variable A is related to B , and B is related to C , then A is necessarily related to C . Step 5. The reachability matrix obtained in Step 4 is partitioned into different levels. Step 6. A digraph is drawn and the transitive links are removed based on the relationships given in the reachability matrix. Step 7. The resultant digraph is converted into an ISM model, by replacing variable nodes with statements. Step 8. The ISM model developed in Step 7 is reviewed to check for conceptual inconsistency and necessary modi cations are made.

In this study, after applying the ISM methodology to the barriers and bene ts of e-procurement systems, the barriers and bene ts that form the base of the hierarchies are then entered into the structural equation model to investigate their effects on the e-procurement adoption decision. The next subsection describes the SEM methodology.

3.3. Structural equation modeling

SEM is a multivariate technique that enables the researcher to simultaneously examine a series of interrelated dependence rela- tionships among the measured variables and latent constructs as well as between the latter ones ( Hair et al., 2010 ). This study uses Anderson and Gerbing's (1988) two-step approach, in which the measurement model is estimated prior to the structural model. The AMOS 16.0 software is used to test the measurement and structural models based on the maximum likelihood estimation method.

We test the measurement model using the con rmatory factor analysis (CFA) as described in Hair et al. (2010) . We investigate the model χ 2 and its signi cance rst. As Jöreskog and Sörbom (1993)

propose,

we also examine the χ 2 with the number of degrees of

freedom,

i.e., χ 2 = df . Additionally, we calculate other goodness-of-

t measures. Based on the recommendations of Hu and Bentler (1998) , we choose standardized root mean square (SRMR), goodness-of- t index (GFI), adjusted goodness-of- t index (AGFI), normed t index (NFI), comparative t index (CFI), and root mean square error of approximation (RMSEA).

After evaluating the t of the measurement model, we assess the convergent validity and discriminant validity of the constructs. We test the convergent validity by examining (i) the factor

loadings; (ii) the average variance extracted (AVE) for each latent

construct; and (iii) construct reliability (CR). To test the discrimi- nant validity, the AVE values for the constructs are compared with the square of the correlation estimate between these two con- structs ( Fornell and Larcker, 1981 ). After assessing the convergent and discriminant validities, we examine the standardized residual covariances and modi cation indices to identify the problems in the measurement model. Once the measurement model is validated, we test our struc-

tural model where our main focus is to test the hypothesized relationships. The goodness-of- t of the structural model is evaluated with the same measures that we use to test the measurement model. After evaluating the goodness-of- t mea- sures for the structural model, we nally analyze the path coef cients and loading estimates. The methodology used in this study is summarized in Fig. 1 .

4.

Data analysis and results

4.1.

Interpretive structural modeling

Recall that the rst part of this study aims at identifying the relationships within the barriers and bene ts of e-procurement systems. As explained in Section 3.1 , 14 barriers and 15 bene ts are determined and they are entered into the ISM model.

4.1.1. Structural self-interaction matrix (SSIM)

ISM methodology suggests using expert opinions based on various management techniques such as brainstorming, nominal technique, etc., in developing the contextual relationships among the variables ( Ravi and Shankar, 2005 ). Accordingly, to de ne the contextual relationships among the barriers and bene ts of e-procurement systems, a brainstorming session is performed with eight top managers from the company and four experts from academia. Similar to the previous studies (e.g., Ravi and Shankar, 2005; Govindan et al., 2012 ), four symbols are used to denote the

direction

of the relationship between barriers (bene ts) i and j :

V

Barrier (bene t) i will help to alleviate (achieve) barrier (Bene t) j ;

A

Barrier (bene t) j will be alleviated (achieved) by barrier (Bene t) i ;

X

Barriers (bene ts) i and j will help to achieve each other;

O

Barriers (bene ts) i and j are unrelated.

The structural self-interaction matrices for barriers and bene ts of e-procurement systems are given in Tables A1 and A2 , respectively. All tables for the ISM methodology can be found in Appendix A .

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

83

4.1.2. Initial reachability matrix

Once the SSIM is constructed, it is transformed into a binary matrix named as the initial reachability matrix. The rules for transforming SSIM into the initial reachability matrix are given in Table 3 . For instance, if the element ( i , j ) of the SSIM is V, then the element ( i , j ) of the initial reachability matrix becomes 1 and the

element ( j , i ) becomes 0. The initial reachability matrices for barriers and bene ts of e-procurement systems are given in Tables A3 and A4 , respectively.

4.1.3. Final reachability matrix

After developing the initial reachability matrix, this matrix is checked for transitivity resulting in the nal reachability matrix. Recall that transitivity is a basic assumption in ISM stating that if a variable A is related to B , and B is related to C , then A is necessarily related to C . The nal reachability matrices for barriers and bene ts of e-procurement systems can be found in Tables A5 and A6 , respectively.

4.1.4. Level partitions

The next step in ISM is to de ne the reachability and ante- cedent sets for each variable from the nal reachability matrix. The reachability set consists of the variable itself and the other

Table 3 Rules for transforming SSIM into the initial reachability matrix.

Structural self-interaction matrix

Initial reachability matrix

Element ( i , j )

Element ( i , j )

Element ( j , i )

V

10

A

01

X

11

O

00

variables which it may help achieve, whereas the antecedent set consists of the variable itself and the other variables which may help in achieving it. Thereafter, the intersection of these sets is derived for all the variables. The variables for which the reach- ability and intersection sets are the same constitute the top level in the ISM hierarchy. When a variable is assigned to a level, it is discarded from the variable set and the process is repeated for the remaining variables. The process continues until all variables are assigned to a level. The identi ed levels are then used to construct the digraph and the nal ISM model. The level partitions for barriers and bene ts of e-procurement systems are given in Tables A7 A12 and A13 A16 , respectively.

4.1.5. ISM model

After de ning the levels, the structural model is constructed from the nal reachability matrix, resulting in a digraph. Then, the transitive links are removed based on the relationships given in the nal reachability matrix, leading to the ISM model. The ISM

models for barriers and bene ts of e-procurement systems are given in Figs. 2 and 3 , respectively. Fig. 2 denotes that inadequate IT infrastructure of suppliers/ business partners is a very signi cant barrier for e-procurement systems since it lies at the bottom level of the hierarchy. This barrier then leads to the lack of e-procurement knowledge/skilled personnel and lack of adequate technical/IT infrastructure. These three barriers constitute the bottom two levels of the hierarchy, meaning that these are strong drivers and may be treated as the root causes of the remaining barriers. Therefore, top management should initiate a comprehensive strategic plan to improve the technical/IT infrastructure of the company and the whole supply chain. This strategy should also be supported by procurement of skilled personnel and by training programs to improve the technical expertise needed for e-procurement adoption. Overcoming these

11- Resistance to change of internal/ external customers in supply chain 2- Security, 1- Incompliance
11- Resistance to
change of internal/
external customers
in supply chain
2- Security,
1- Incompliance
confidentiality and
4- High investment
cost of IT
13- Cost/benefit
with company
authentication
culture
concerns
infrastructure/
software
7- Inadequate
business processes
to support e-
procurement
concern
12- Inter-operability
concerns with other
systems used
14- Lack of
flexibility in process
and documentation
3- Bureaucratic
disfunctionalities in
practice
6- External (Supply
chain/Business
partners)
incompatibility
10- Lack of system
integration with
suppliers/business
partners
5- Lack of e-
procurement
knowledge/skilled
personnel
8- Lack of adequate
technical/IT
infrastructure
9- Inadequate IT
infrastructure of
suppliers/business
partners

Fig. 2. ISM model for the barriers of e-procurement systems.

84

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

6- Cost savings in overall purchasing process 11- Decreased 2- Minimization of process errors 5-
6- Cost savings in
overall purchasing
process
11- Decreased
2- Minimization of
process errors
5- Better
management and
control of suppliers
9- Enhanced
10- Enhanced
inventory
bureaucracy and
decision making
management
redundancy
14- Time savings in
overall purchasing
process
15- Increased
process quality and
efficiency
13- Improved
1- Easier access to
market data and
enhanced intelligence
3- Quicker response
to problems through
real-time
information
4- On-line and real-
time reporting
7- Improved supply
chain transparency
8- Simplified and
streamlined
purchasing process
communication and
collaboration in
supply chain
12- Integrated
information sharing

Fig. 3. ISM model for the bene ts of e-procurement systems.

Table 4 Indicators of the latent variables in the measurement and structural models.

Latent

Indicator Description

variable

Barriers

R

1

5. Lack of e-procurement knowledge/skilled personnel 8. Lack of adequate technical/IT infrastructure 9. Inadequate IT infrastructure of suppliers/business partners

R

2

R

3

Bene ts

N

1

1. Easier access to market data and enhanced intelligence 3. Quicker response to problems through real-time information 4. On-line and real-time reporting 7. Improved supply chain transparency 8. Simpli ed and streamlined purchasing process 12. Integrated information sharing 13. Improved communication and collaboration in supply chain

N

2

N 3

N 4

N

5

N

6

N

7

barriers will help to reduce the negative effects of all the other barriers up to internal/external resistance to change. The ISM model regarding the bene ts of e-procurement sys- tems indicates that the most signi cant bene t of e-procurement systems is integrated information sharing. Therefore, for the success of the whole project, enhancement and well-integration in information sharing within the whole supply chain is the key factor, which has a boosting effect on the other bene ts. As one can see from Fig. 3 , integrated information sharing leads to easier access to market data and enhanced intelligence; quicker response to problems through real-time information; on-line and real time reporting; improved supply chain transparency; simpli ed and streamlined purchasing process; improved communication and collaboration in supply chain; and nally cost savings in overall purchasing process that constitutes the top level of the hierarchy. According to the results of the ISM models (see Figs. 2 and 3 ), 3 barriers (7 bene ts) of e-procurement systems constitute the bottom two levels of the hierarchy and they are found to be the root causes of the other barriers (bene ts). These barriers and bene ts are then entered into the SEM model and they become the indicators of the latent variables as given in Table 4 .

4.2. Measurement model results

In this study, Anderson and Gerbing's (1988) two-step approach is used, i.e., the measurement model is estimated prior to the structural model. The AMOS 16.0 software is used to test the measurement and structural models based on the maximum likelihood estimation method. Although the data does not t to normal distribution very well (multivariate kurtosis ¼ 51.29; criti- cal ratio ¼ 25.24), there are several studies in the literature (e.g., Hu et al., 1992; Olsson et al., 2000 ) that shows the maximum like- lihood estimation even in the case of nonnormality outper- forms other estimation methods such as the generalized least squares or asymptotically distribution free methods ( Schermelleh- Engel et al., 2003 ). The measurement model consists of two latent variables, i.e., Barriers and Bene ts, with three and seven indicators, respec- tively. We test the measurement model using the con rmatory factor analysis (CFA) method as described in Hair et al. (2010) . Here, we present the results for the nal measurement model. The model χ 2 is calculated as 65.66 with 31 degrees of freedom and it is signi cant. However, because of the shortcomings of the χ 2 test statistic, too much emphasis should not be given to this test ( Schermelleh-Engel et al., 2003 ). Jöreskog and Sörbom (1993) propose to compare the magnitude of χ 2 with the number of degrees of freedom, i.e., χ 2 = df . Our measurement model yields χ 2 = df ¼ 2 : 12, which is below the threshold of 3.00. Additionally, we calculate other goodness-of- t measures. Based on the recom- mendations of Hu and Bentler (1998) , we choose standardized root mean square (SRMR), goodness-of- t index (GFI), adjusted goodness-of- t index (AGFI), normed t index (NFI), comparative t index (CFI), and root mean square error of approximation (RMSEA). The recommended values for these t indices and the results for the measurement model are given in Table 5 , which shows that all values are within the recommended ranges indicat- ing that the measurement model has a good t. After evaluating the t of the measurement model, one should also assess the construct validity that consists of two components:

convergent validity and discriminant validity. We test the convergent validity by examining the factor loadings rst. All factor loadings are signi cant at the 0.001 level and except two of them all are above 0.70. Only R 1 and R 3 have factor loadings of 0.65 and 0.67, respectively; but since these are

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

85

Table 5 Goodness-of- t measures for the measurement model.

Goodness-of- t measure

Recommended value ( Schermelleh-Engel et al. (2003) )

Result

χ 2 = df SRMR GFI AGFI NFI CFI RMSEA

 

r 3.00

2.118

r 0.10

0.052

Z 0.90

0.957

Z 0.85

0.923

Z 0.90

0.957

Z 0.95

0.977

r 0.08

0.064

Table 6 Results for the measurement model.

 

Latent variable

Indicator

 

Factor loading

AVE

CR

Barriers

0.490

0.742

 

R

1

0.654 n

R

2

0.774 n

R

3

0.665 n

Bene ts

0.624

0.924

 

N

1

0.743 n

N

2

0.790 n

N 3

 

0.748 n

N 4

0.809

n

N

5

0.865

n

N

6

0.756

n

N

7

0.810 n

n Signi cant at the 0.001 level (two-tailed).

above 0.50, they are also acceptable. Second, the average variance extracted (AVE) is calculated for each latent construct. AVE values are calculated as 0.49 and 0.62 for barriers and bene ts, respec- tively. Since an AVE of 0.50 or higher suggests adequate conver- gent validity, one can see that the AVE for barriers is just below this critical level. Third, we calculate the construct reliability (CR), which is also an indicator of convergent validity. CR values are calculated as 0.74 and 0.92 for barriers and bene ts, respectively. Since they both exceed the threshold of 0.70, this also validates convergent validity. The results for the measurement model can be found in Table 6 . To test the discriminant validity, the AVE values for the constructs are compared with the square of the correlation estimate between these two constructs ( Fornell and Larcker, 1981 ). The correlation estimate between Barriers and Bene ts is calculated as 0.30. Since AVE values for both constructs are greater than the squared correlation estimate, this result provides good evidence of discriminant validity. After assessing the convergent and discriminant validities, we examine the standardized residual covariances and modi cation indices to identify the problems in the measurement model. Standardized residual covariances should be less than | 2.50 | for a correct model. The results denote that only the standardized residual covariance between R 2 and N 4 is outside this range and it is calculated as 2.68. However, it is below another threshold of | 4.00 | , where existence of values above |4.00 | suggests a potentially unacceptable degree of error ( Hair et al., 2010 ). Therefore, the standardized residual covariances indicate no problem in the measurement model. On the other hand, we also examine the modi cation indices. High modi cation indices suggest that the t could be improved signi cantly by freeing the corresponding path to be estimated. Since the modi cation indices of the measure- ment model (available upon request) are reasonable, this result also provides good evidence of model validation.

4.3. Structural model results and hypothesis testing

After testing the measurement model, we test our structural model and our main focus is to test the hypothesized relation- ships. The structural model consists of two exogenous variables (Barriers and Bene ts) and one endogenous variable (Adoption). Recall that the exogenous variables Barriers and Bene ts have three and seven indicators, respectively. On the other hand, the endogenous variable Adoption has a single indicator, which is measured through the question that investigates the necessity of the e-procurement system for the company. The goodness-of- t of the structural model is evaluated with the same measures used to test the measurement model. The model χ 2 is calculated as 71.62 with 39 degrees of freedom and its p value is 0.001. The normed χ 2 is χ 2 = df ¼ 1 : 84. The goodness-of- t measures for the structural model are given in Table 7 . The results denote that all the measures are within desirable ranges and the structural model ts the data well. After evaluating the goodness-of- t measures for the structural model, we turn our attention to the path coef cients and loading estimates, which are given in Fig. 4 . When we compare the loading estimates of the measurement and structural models, we can see that they are nearly the same and the maximum difference is 0.007. This is an evidence of stability among the measured indicator variables and supports the validity of the measurement

Table 7 Goodness-of- t measures for the structural model.

Goodness-of- t measure

Recommended value ( Schermelleh-Engel et al. (2003) )

Result

χ 2 = df SRMR GFI AGFI NFI CFI RMSEA

r 3.00

1.836

r 0.10

0.050

Z 0.90

0.956

Z 0.85

0.926

Z 0.90

0.955

Z 0.95

0.979

r 0.08

0.055

R1 0.655 0.775 R2 Barriers 0.663 R3 H1: −0.154 * N1 0.303 Adoption N2 0.747
R1
0.655
0.775
R2
Barriers
0.663
R3
H1: −0.154 *
N1
0.303
Adoption
N2
0.747
0.785
N3
H2: 0.484 **
0.741
0.807
N4
Benefits
0.869
N5
0.755
N6
0.811
N7

* p -value = 0.023 ** p-value < 0.001

Fig. 4. Results of the path analysis.

86

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

model ( Hair et al., 2010 ). From Fig. 4 , it can be seen that the path coef cient estimate between barriers and adoption is 0.15 and it is signi cant at the 0.05 level. Thus, our rst hypothesis is supported, i.e., barriers of e-procurement systems have negative effect on the adoption decision. On the other hand, the path coef cient estimate between bene ts and adoption is 0.48 and it is signi cant at the 0.001 level. Thus, our second hypothesis is also supported, i.e., bene ts of e-procurement systems have positive effect on the adoption decision. As stated before, besides these hypotheses, another point of interest is to examine whether the barriers or bene ts of e-procurement systems have higher impact on the adoption decision. Since the path coef cient estimate of bene ts is higher than that of the barriers, we conclude that bene ts of an e-procurement system overweigh its barriers. Thus, it would be bene cial for the company to adopt the e-procurement system. However, it is also noteworthy that the squared multiple correlations (i.e., R 2 ) of Adoption is calculated as 0.21. In other words, bene ts and barriers explain 21% of the variance of Adoption. Therefore, adding other latent constructs that may affect the e-procurement adoption decision would increase the variance explained.

5. Conclusion

This study investigates the effects of the barriers and bene ts of e-procurement systems on the e-procurement adoption decision. An empirical analysis has been performed for a retail store chain which operates in book and stationery sector in Turkey. The rst aim of the study is to determine the best route for the company to follow at their transition period from traditional methods of purchasing to e-procurement adoption. This is achieved by measuring the driving power of each potential barrier and bene t as well as the inter-relations within them. The second and the ultimate aim of the study is to put forth whether barriers or bene ts have higher impact on the company's e-procurement adoption decision by using an approach based on cost/bene t analysis. The results of the rst part of the study denote that among the barriers of e-procurement systems, inadequate IT infrastructure of suppliers/business partners is the most important barrier in the adoption of e-procurement systems. This barrier then leads to two other important barriers, which are lack of adequate technical/IT infrastructure and lack of e-procurement knowledge/skilled personnel. Thus, in tackling the barriers of e-procurement systems, the management of the company should focus more on IT/technical infrastructure of the supply chain, and skills,

knowledge and capability of human resources for adopting e- procurement. Overcoming these barriers will help to reduce the negative effects of all other barriers up to internal/external resistance to change. When we turn back to the bene ts of e-procurement systems, integrated information sharing forms the base of the hierarchy and it is found as the most signi cant bene t. Therefore, enhancement and well-integration in information sharing within the supply chain is the key factor for the success of e-procurement adoption. An improvement in integrated information sharing helps to achieve all other bene ts, which are easier access to market data and enhanced intelligence; quicker response to problems through real-time information; on-line and real time reporting; improved supply chain transparency; simpli ed and streamlined purchasing process; improved communication and collaboration in supply chain; and nally cost savings in overall purchasing process. The results of the second part of the study denote that barriers (bene ts) of e-procurement systems have negative (positive) effect on the e-procurement adoption decision and the effect of bene ts on the adoption decision is higher than that of the barriers, which might be considered as indirect costs of adoption. In other words, the potential bene ts of e-procurement systems have stronger impacts than the costs that the company may put up with due to the potential barriers. Thus, based on the results of our cost/ bene t analysis, it would be bene cial for the company to adopt the e-procurement system. It is noteworthy that the results of this study might differ for other companies and sectors. Thus, as a further study, the models could be enhanced with other factors and a sector-base analysis could be performed from a larger perspective. Besides, the method proposed in this paper could be considered as an instrument and a guide for change management for the companies. Accordingly, companies could apply this method to help their transition decisions from traditional systems to e-systems.

Acknowledgments

The authors are grateful to the Editor and the referee for their valuable comments and suggestions.

Appendix A. Tables for the ISM methodology

See Tables A1-A16 .

Table A1 Structural self-interaction matrix for barriers of e-procurement systems.

Barriers

14

13

12

11

10

9

8

7

6

5

4

3

2

1

1. Incompliance with company culture 2. Security, con dentiality, and authentication concerns 3. Bureaucratic disfunctionalities in practice 4. High investment cost of IT infrastructure/software 5. Lack of e-procurement knowledge/skilled personnel 6. External (Supply chain/Business partners) incompatibility 7. Inadequate business processes to support e-procurement 8. Lack of adequate technical/IT infrastructure

O

O

O

V

O

O

O

A

O

O

O

V

O

V

O

A

V

A

O

O

A

A

O

O

O

O

O

O

V

O

O

O

V

O

O

O

O

X

O

V

O

A

A

O

A

A

V

O

O

V

O

O

O

V

V

V

O

V

V

X

A

O

O

X

O

X

V

O

O

O

V

O

V

V

V

A

9. Inadequate IT infrastructure of suppliers/business partners

V

O

V

V

V

10. Lack of system integration with suppliers/business partners

V

O

A

V

11. Resistance to change of internal/external customers in supply chain

A

A

A

12. Inter-operability concerns with other systems used

V

O

13. Cost/bene t concern

O

14. Lack of exibility in process and documentation

P. Tokta ş -Palut et al. / Int. J. Production Economics 158 (2014) 77 90

Table A2 Structural self-interaction matrix for bene ts of e-procurement systems.

87

Bene ts

15

14

13

12

11

10

9

 

8

7

6

5

4

 

3

2

1

1. Easier access to market data and enhanced intelligence 2. Minimization of process errors 3. Quicker response to problems through real-time information 4. On-line and real-time reporting 5. Better management and control of suppliers 6. Cost savings in overall purchasing process 7. Improved supply chain transparency 8. Simpli ed and streamlined purchasing process 9. Enhanced inventory management

 

V

V

X

A

A

V

V

 

A

A

V

X

A

 

X

O

X

V

A

A

O

O

V

A

A

V

O

O

X

V

V

X

A

X

V

V

O

A

V

V

X

V

V

X

X

X

V

V

V

X

V

V

 

X

V

X

A

A

X

V

A

X

V

A

O

A

A

A

A

A

A

A

V

V

X

X

V

V

V

V

V

V

A

O

X

V

O

X

V

A

A

A

X

 

10. Enhanced decision making

 

X

V

A

A

X

11. Decreased bureaucracy and redundancy

 

X

V

A

A

12. Integrated information sharing

 

V

V

X

13. Improved communication and collaboration in supply chain

 

V

V

14. Time savings in overall purchasing process

 

X

15. Increased process quality and ef ciency

 

Table A3 Initial reachability matrix for barriers of e-procurement systems.

 

Table A5 Final reachability matrix for barriers of e-procurement systems.

 

Barriers

1

2

3

4

5

6

7

8

9

10

11

12

13

14

Barriers

1

2

3

4

5

6

7

8

9

10

 

11

12

 

13

14

1

1010000000

1

0

0

0

1

1

0

1

0

00

 

1 a 000

 

1

0

0

0

2

0100000000

1

0

0

1

2

0

1

0

0

00

1 a 000

1

0

0

1

3

0010001000

1

0

0

0

3

1

a

1

a

1

0

00

1

000

1

1

a

0

1 a

4

0001000000

1

0

1

0

4

0

0

0

1

00

0

000

1

0

1

0

5

0001111000

1

0

0

1

5

1

a

1

a

0

1

11

1

001

a

1

1

a

1

a

1

6

0101010001

1

1

0

1

6

0

1

0

1

01

1 a 001

1

1

1

a

1

7

1100001000

1

1

0

1

7

111 a 0

 

00

1

001

a

1101

8

0001000101

1

1

0

1

8

0

1

a

0

1

01

a

1

a

101

1

1

1

a

1

9

0001010111

1

1

0

1

9

0

1

a

0

1

01

1

a

111

1

1

1

a

1

10

0100010001

1

0

0

1

10

0101 a 01

1 a 001

1

1

a

0

1

11

0000000000

1

0

0

0

11

0

0

0

0

00

0

000

1

0

0

0

12

0100001001

1

1

0

1

12

1

a

1

0

0

01

a

1

001

1

1

0

1

13

0001000000

1

0

1

0

13

0

0

0

1

00

0

000

1

0

1

0

14

0000001000

1

0

0

1

14

1

a

1 a

0

0

00

1

000

1

1

a

0

1

 

a

Values obtained by incorporating transitivity.

 

Table A4 Initial reachability matrix for bene ts of e-procurement systems.

 

Table A6 Final reachability matrix for bene ts of e-procurement systems.

 

Bene ts

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Bene ts

1

2

3

4

5

6

7

8

9

10

11

12

 

13

14

15

1

1010110011

0

0

1

1

1

1

1

1 a

1

1

a

111 a 1 a

 

11

 

1

a

1

a

111

2

0110010010

0

0

0

1

1

2

1 a 111 a 1 a 10011 a

1

a

0

1 a

1

1

3

1111110011

1

0

1

1

1

3

1111111 a 1 a 11

11 a 111

4

1011111111

1

1

1

1

1

4

1

1 a 11111111

11

 

1

1

1

5

1000111011

0

0

1

1

1

5

1

1

a

1 a

1 a 1111 a

 

11

1 a

1

a

111

6

0000010000

0

0

0

0

0