PRIORITIZING THE BANKING SERVICE QUALITY OF DIFFERENT BRANCHES USING FACTOR ANALYSIS, AHP AND TOPSIS METHODOLOGY-A CASE STUDY

International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 DOI: 10.5121/ijmvsc.2015.6406 65 PRIORITIZING THE BANKING SERVICE QUALITY OF DIFFERENT BRANCHES USING FACTOR ANALYSIS, AHP AND TOPSIS METHODOLOGY-A CASE STUDY G.V.S.K.Gautham 1 , Dr.G.Ram Babu 2 , D.Balaji Naik 3 , G.Gandhi Krishna 4 1 Department of Mechanical Engg, Andhra University, Visakhapatnam-530 007 2 Department of Mechanical Engg, Andhra University, Visakhapatnam-530 007 3 Department of Mechanical Engg, Andhra University, Visakhapatnam-530 007 4 Manager,MMGS-III,State Bank of India ABSTRACT In recent years, India’s service industry is developing rapidly. The objective of the study is to explore the dimensions of customer perceived service quality in the context of the Indian banking industry. In order to categorize the customer needs into quality dimensions, Factor analysis (FA) has been carried out on customer responses obtained through questionnaire survey. Analytic Hierarchy Process (AHP) is employed to determine the weights of the banking service quality dimensions. The priority structure of the quality dimensions provides an idea for the Banking management to allocate the resources in an effective manner to achieve more customer satisfaction. Technique for Order Preference Similarity to Ideal Solution (TOPSIS) is used to obtain final ranking of different branches. KEYWORDS Service Quality, Factor analysis, Analytic Hierarchy Process, Technique for Order Preference Similarity to Ideal Solution. 1. INTRODUCTION Banking sector in India is sound, adequately capitalized and well-regulated. It has always been one of the most preferred destinations for employment. In this decade, this sector has emerged as a sunrise sector in Indian economy. A large number of people are engaged with this sector from staff to executive level to operate the whole system .The major challenge to this sector at present is to ensure expected quality of service that the customer wishes. Factor analysis is one of the very useful techniques to summarize a large amount of data in a manageable way. Factor analysis attempts to identify underlying variables, or factors, that explain the pattern of correlations within a set of observed variables. Factor analysis is often used in data reduction to identify a small number of factors that explain most of the variance observed in a much larger number of manifest variables. It may used to define a relationship among sets of International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 66 many interrelated variables to be examined and represented in terms of a few underlying factors. This technique is applicable to identify the underlying dimensions or factors that explain the correlations among a set of variables. Factor analysis can be employed to determine the brand attributes that influence customer choice. In the current study this technique is used to determine the factors that influence the quality of banking service. The overall banking service is interdependent on the service attributes. The quality of those service attributes dominates the satisfaction of overall service of customer and this relationship can be depicted through a linear model stating overall satisfaction as dependent and others service attributes as independent variable. Analytic hierarchy process (AHP) is a structured technique for organizing and analyzing complex decisions .It is a multi-criteria decision making (MCDM) technique proposed by Saaty. It is a theory of measurement through pair wise comparisons and relies on the judgments of experts to derive priority scales. It is the scale that measure intangibles in relative terms. The comparisons are made using a scale of absolute judgments that represents, how much more one element dominates another with respect to a given attribute. The judgments may be inconsistent, how to measure inconsistency and improve the judgments, when possible to obtain better consistency is a concern of the AHP. The derived priority scales are synthesized by multiplying them by the priority of their parent nodes and adding for all such nodes. Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) was set forth by Hwang and Lin (1987). In this technique, “n” different alternatives are evaluated by “m” different attributes, the attributes being common to all the alternatives. The method belongs to MCDM (Multiple Criteria Decision Making) group of methods and identifies solutions from a finite set of alternatives based upon simultaneous minimization of distance from an ideal point and maximization of distance from a negative ideal point. Hence ranking of different alternatives can be done with the help of TOPSIS methodology. 2. METHODOLOGY 2.1 Factor analysis Factor analysis is carried out with a view to reduce the list of customer attributes. The data received from the questionnaire survey was carried out using statistical package for social sciences (SPSS) version 16.0. The factor analysis begins with the correlation matrix, in which the inter-correlations between the studied variables (customer attributes) are presented. The sample adequacy for the response data is examined through KMO and Bartlett’s tests. Communalities are determined and rotated component matrix is prepared. Scree plot is obtained for the data to identify the appropriate factors. Factors obtained through factor analysis are grouped and AHP methodology is employed to find out the weights. 2.2 Analytical hierarchy process Step 1: Establishment of pair-wise comparison matrix Setup the pair-wise comparison matrix of order n n × consists of n elements (requirements) in the rows and columns whose priorities are to be determined. International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 67 Step 2: Perform pair-wise comparisons of all the elements. Table 1.Saaty’s scale This comparison scale enables the decision-maker to incorporate experience and knowledge intuitively and indicate how many times an element dominates another with respect to the criterion. The decision-maker can express his preference between each pair of elements verbally as equally preferred, moderately preferred, strongly preferred, very strongly preferred and extremely preferred. These descriptive preferences would then be translated into numerical values 1, 3, 5, 7, 9 respectively, with 2, 4, 6 and 8 as intermediate values for comparisons between two successive judgments. Reciprocals of these values are used for the corresponding transposed judgments. For a matrix of order n , ( ) 1 / 2 n n − comparisons are required. After the pair-wise comparisons are completed, proceed for the next step to estimate the Eigen values of the matrix. Step 3: Estimation of the Eigen values of the matrix In this method, first sum the values in each column of the pair-wise comparison matrix and then divide each element in a column by the sum of its respective column. The resultant matrix is termed as the normalized pair-wise comparison matrix. Finally sum the elements in each row of the normalized pair-wise comparison matrix and divide the sum with the number of elements. The result of this computation is referred to as the priority matrix and is an estimation of the Eigen values of the matrix. Step 4: Checking the consistency of pair-wise judgments In order to verify the consistency of the pair-wise comparison matrix, Saaty proposed consistency index (CI) and consistency ratio (CR). The CI and CR are defined as follows. CI =     ; CR = . . Where   = maximum principal Eigen value of the comparison matrix n = number of elements (order of the pair-wise comparison matrix) Intensity of Importance Interpretation 1 Requirement i and j are of equal value. 3 Requirement i has a slighter higher value than j 5 Requirement i has a strongly higher value than j 7 Requirement i has a very strongly higher value than j 9 Requirement i has an absolute higher value than j 2,4,6,8 These are intermediate scales between two adjacent judgments Reciprocals If requirement i has lower value than j International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 68 The value of   is obtained by first multiplying the pair-wise comparison matrix with the priority matrix. Then divide the first element of the resulting matrix by the first element of the priority matrix, the second element of the resulting matrix by the second element in the priority matrix, and so on. A single column matrix is obtained and the average of the elements of the matrix gives the value of   . The RI in the above equation represents the average consistency index for numerous random entries of same-order reciprocal matrices. The values of RI for matrices of order n are given in table 2 Table 2.Average value of RI for corresponding matrix order (Saaty, 1980) n RI n RI n RI n RI 1 0 5 1.12 9 1.45 13 1.56 2 0 6 1.24 10 1.49 14 1.57 3 0.58 7 1.32 11 1.51 15 1.59 4 0.90 8 1.41 12 1.48 If CR ≤ 0.1, then the estimate is accepted; otherwise, a new comparison matrix is solicited until CR ≤ 0.1 (Chang et al., 2007) In the present work, AHP is integrated with Factor analysis and TOPSIS to determine the priority structure of customers’ service quality attributes and ranking of different banks. 2.3 Technique for order preference similarity to ideal solution (TOPSIS) Step-1: Construct normalized decision matrix by using the formula, r ij =  (∑ ) / for i = 1,2,….,m ; j = 1,2,…,n Step-2: Construct the weighted normalized decision matrix. Multiply each column of the normalized decision matrix by its associated weight. An element of the new matrix is: v ij = w j * r ij Step-3: Now determine the positive ideal and negative ideal solutions using, Positive ideal solution : V j * = { max( v ij ) } Negative ideal solution : V j ′ = { min( v ij ) } Step-4: Calculate the separation measures for each alternative. The separation from the ideal alternative is: S i * =[∑( !"# v ij – v j * ) 2 ] 1/2 International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 69 Similarly, the separation from the negative ideal alternative is: S i ′ = [ ∑( !"# v ij – v j ′ ) 2 ] 1/2 Step-5: Calculate the relative closeness to the ideal solution C i * and the corresponding ranks of different Branches: C i * = $% $%&$∗ ; 0 < C i * < 1 3. CASE STUDY 3.1 Questionnaire survey In view of demonstrating methodology, a case study has been undertaken in 4 branches of State Bank of India, Visakhapatnam. After several discussions made with the experts in the quality service, a questionnaire was developed on the expectations of the customers from 5 dimensions of service quality namely Reliability, Responsiveness, Assurance, Empathy and Tangibles. The questionnaire was administrated to 150 customers in each branch (Hair et al.,1995). After receiving their comments, the questionnaire on customer attributes was revised and finalized. The respondents were asked to indicate the degree of importance of the customer attributes (variables from Q1 to Q30) in terms of a five - point Likert scale(1-Low,2-Average,3-Good,4-VeryGood,5Excellent). To make the study broader, respondents with age greater than 18 years familiar with the use of all modern technologies were chosen. Customers who don’t have time or not willing are omitted. People who come to bank on behalf of actual customers are omitted from the study. A total of 624 responses were received from the respondents and in which 46 responses are invalid as the respondents filled the questionnaires not properly. However, 578 responses were considered to carry out the factor analysis. The sample questionnaire is presented below: International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 70 Table 3.Sample questionnaire International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 71 3.2 Performing Factor Analysis Factor analysis is carried out with a view to reduce the list of customer attributes. Kaiser-Meyer Olkin measure of sampling adequacy (KMO) and Bartlett’s test of Sphericity were used to examine the appropriateness of the factor analysis. In this work, the factor analysis of the data received from the questionnaire survey was carried out using statistical package for social sciences (SPSS) version 16.0. The Bartlett’s test produces a χ 2 of 3625 with a significance level of 0.000, which shows that the sample taken from the total population under study is adequate. The KMO test produces a measure of 0.627, which further confirms that the adequacy of the sample. The test results are shown in the table 4. The results obtained from the Bartlett’s test and KMO test also indicate the suitability of the application of the factor analysis. Hence factor analysis is considered as an appropriate technique for further analysis of the data. Table 4.Result of KMO and Bartlett’s test KMO and Bartlett's Test Kaiser-Meyer-Olkin Measure of Sampling Adequacy. 0.627 Bartlett's Test of Sphericity Approx. Chi-Square 3.625E3 Df 435 Sig. .000 . In the language of the factor analysis the proportion of the variance of the particular variable that is due to common factors (shared with other variables) is called communality. Initial communalities are estimates of the variance in each variable accounted for by all components or factors. Extraction communalities are estimates of the variance in each variable accounted for by the factors (or components) in the factor solution. Small values indicate variables that do not fit well with the factor solution, and should possibly be dropped from the analysis. Table 5.Communalities Question Initial Extraction Q1 1.000 .730 Q2 1.000 .758 Q3 1.000 .604 Q4 1.000 .753 Q5 1.000 .534 Q6 1.000 .608 Q7 1.000 .649 Q8 1.000 .668 Q9 1.000 .639 Q10 1.000 .575 Q11 1.000 .663 Q12 1.000 .768 International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 72 Q13 1.000 .600 Q14 1.000 .596 Q15 1.000 .638 Q16 1.000 .671 Q17 1.000 .728 Q18 1.000 .705 Q19 1.000 .611 Q20 1.000 .582 Q21 1.000 .741 Q22 1.000 .630 Q23 1.000 .742 Q24 1.000 .708 Q25 1.000 .655 Q26 1.000 .604 Q27 1.000 .656 Q28 1.000 .654 Q29 1.000 .673 Q30 1.000 .610 A Scree Plot is a simple line segment plot that shows the fraction of total variance in the data as explained. A Scree plot is shown in figure 1 which indicates the Eigen values against the number of factors in order of extraction. Figure 1.Scree Plo International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 73 Table 6.Rotated Component matrix From the table 6, the factors obtained through factor analysis are grouped from 1 to 5 are labeled as Customer Service, Physical Features, Banking Facilities, System and Executive Innovation respectively and are summarized in the table 7 Component Question 1 2 3 4 5 Q14 .676 Q25 .626 Q22 .563 Q24 .559 Q5 .413 Q9 .564 Q20 .536 Q10 .519 Q23 .465 Q6 .584 Q18 .534 Q8 .434 Q21 .416 Q7 .612 Q11 .539 Q28 .538 Q29 .474 Q30 .418 Q16 .645 Q3 .569 International Journal of Managing Value and Supply Chains Table 15.Pairwise comparison matrix of different service quality dimensions and branches B I CS 0.3451 PF 0.2236 ST 0.2486 BF 0.1205 EI 0.0620 Figure 5.Weights of service quality dimension for By adding overall weights of the different dimensions to the table 11 we get, Table 16.Overall weights of different service quality dimensions and branches Weight 0.3675 Branch CS B I 0.4885 B II 0.1896 B III 0.0893 B IV 0.2329 3.5 TECHNIQUE FOR ORDER PREFERENCE SIMILARITY TO IDEAL SOLUTION (TOPSIS) Construct normalized decision matrix by using the formula, r ij =  ∑   0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 CS International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 wise comparison matrix of different service quality dimensions and branches B II B III B IV Overall Weights 0.3200 0.4383 0.3666 0.3675 0.2106 0.2260 0.1562 0.2041 0.3238 0.2024 0.3028 0.2694 0.1142 0.0830 0.1052 0.1057 0.0577 0.0512 0.0674 0.0476 Figure 5.Weights of service quality dimension for different branches By adding overall weights of the different dimensions to the table 11 we get, Table 16.Overall weights of different service quality dimensions and branches 0.2041 0.2694 0.1057 0.0476 PF ST BF 0.5556 0.2468 0.4189 0.1282 0.1690 0.0940 0.2237 0.4773 0.095 0.5015 0.1160 0.2565 0.1807 0.1575 0.2415 0.1382 TECHNIQUE FOR ORDER PREFERENCE SIMILARITY TO IDEAL SOLUTION (TOPSIS) decision matrix by using the formula,       /  for i = 1,2,….,m ; j = 1,2,…,n ST PF BF EI (IJMVSC) Vol. 6, No. 4, December 2015 80 wise comparison matrix of different service quality dimensions and branches Overall Weights 0.3675 0.2041 0.2694 0.1057 0.0476 different branches Table 16.Overall weights of different service quality dimensions and branches 0.0476 EI 0.1282 0.4773 0.2565 0.1382 TECHNIQUE FOR ORDER PREFERENCE SIMILARITY TO International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 81 Table 17.Normalized Decision Matrix Weights 0.3675 0.2041 0.2694 0.1057 0.0476 Branch CS PF ST BF EI B I 0.8420 0.9029 0.4195 0.7686 0.2235 B II 0.3268 0.2746 0.1598 0.4104 0.8322 B III 0.1539 0.1543 0.8526 0.2128 0.4472 B IV 0.4014 0.2936 0.2677 0.4431 0.2409 Construct the weighted normalized decision matrix. Multiply each column of the normalized decision matrix by its associated weight. An element of the new matrix is : v ij = w j * r ij Table 18.Weighted Normalized Decision Matrix CS PF ST BF EI B I 0.3094 0.1842 0.0496 0.0812 0.0106 B II 0.1200 0.0560 0.0414 0.0433 0.0396 B III 0.0565 0.0314 0.2296 0.0224 0.0212 B IV 0.1475 0.0599 0.0721 0.0468 0.0114 Now determine the positive ideal and negative ideal solutions using, Positive ideal solution : V j * = { max maxmax max( v ij ) } Negative ideal solution : V j ′ = { KLM( v ij ) } Hence, V j * = {0.3094, 0.1842, 0.2296, 0.0812, 0.0396} V j ′ = {0.0565, 0.0314, 0.04140, 0.0224, 0.0106} Now, calculate the separation measures for each alternative. The separation from the ideal alternative is: S i * =[∑ !"# v ij – v j * ) 2 ] 1/2 Table 20.Separation measure from Positive Ideal alternative CS PF ST BF EI S i * B I 0 0 0.0324 0 0.0008 0.1822 B II 0.0358 0.0164 0.0354 0.0014 0 0.298 B III 0.0639 0.0233 0 0.0034 0.0003 0.3014 B IV 0.0262 0.0154 0.0248 0.0011 0.0007 0.2611 Similarly, the separation from the negative ideal alternative is: International Journal of Managing Value and Supply Chains Table 21.Separation measure from Negative Ideal alternative CS B I 0.0639 0.0233 B II 0.0040 0.0006 B III 0 B IV 0.0082 0.0008 Calculate the relative closeness to the ideal solution C Branches: Table 22.Relative closeness and Ranks of branches BRANCHES BRANCH I BRANCH II BRANCH III BRANCH IV Figure 6.Overall ranking of Branches 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Branch I International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 S i ′ =  ∑ !"# v ij – v j ′ ) 2 ] 1/2 Table 21.Separation measure from Negative Ideal alternative PF ST BF EI 0.0233 0.0006 0.0034 0 0.0006 0 0.0004 0.0008 0 0.0354 0 0.0001 0.0008 0.0009 0.0005 0.0006 Calculate the relative closeness to the ideal solution C i * and the corresponding ranks of different C i * = $ % $ %&$ ∗ ; 0 < C i * < 1 Table 22.Relative closeness and Ranks of branches RESULT RANK 0.622 1 0.253 4 0.471 2 0.340 3 Figure 6.Overall ranking of Branches Branch II Branch III Branch IV (IJMVSC) Vol. 6, No. 4, December 2015 82 S i ′ 0.301 0.0761 0.1884 0.1022 * and the corresponding ranks of different International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 83 4. RESULTS AND DISCUSSION Quality service is essential to gain competitive advantage in the market place also it helps to sustain customer’s confidence. Profit strategy is always linked with excellent services as it results in more business with existing and new customers. The focus of this research is to identify the ways through which banks can improve the quality of their services and extend to which the quality affects the satisfaction level of customers. It was analyzed that the quality of service and customer satisfaction both are crucial factors for success in the business world. The results from the study show the Ranking of different Branches with respect to different attributes. The satisfaction model will be helpful to determine the overall satisfaction of the banking sector service quality which provides the guidelines in further assessment, betterment and improvement process. Finally, this work introduces an approach that integrates AHP and TOPSIS algorithm to support Banking evaluation decisions. REFERENCES [1] Angur M.G, Natarajan R and Jahera J.S. (1999), “Service quality in banking industry: An assessment in a developing economy”, International journal of bank marketing, 17(3), 116-123 [2] Babakus E and Boller G.W (1992), “An Empirical assessment of SERVQUAL scale”, journal of business research 24(3), 253-68. [3] Baumann C, Burton S and Elliott G (2007), “Predicting consumer behavior in retail banking”, The journal of business and management, 13(1), 79-96 [4] Brown T.J, Churchill G.A and Peter J.P (1993), “Improving the Measurement of Service Quality”, Journal of retailing, 69(1), 127-39. [5] Hair et al (1995) stated that, “it is generally agreed that the minimum sample for appropriate use for statistical analysis is equal to or greater than 5 time of independent variables, but not less than 100”. [6] Hallowell R (1996), “The Relationships of Customer Satisfaction, Customer Loyalty and Profitability: An Empirical Study”, International Journal of Service Industry Management, Vol. 7, No. 4, pp. 27-42. [7] Hwang.C and Lin.M (1987), “Group Decision Making Under Multiple Criteria,’’ SpringerVerlag. [8] Lasser M. C, Monalis and Winson.D (2000), “Service Quality Perspectives and Satisfaction in Private Banking”, Journal of Services Marketing, (14) pp. (2-3). [9] Levesque.T and Mcdougall G.H.G (1996), “Determinants of customer satisfaction in retail banking” ,International Journal of Bank Marketing, 14 (7), 12-20 [10] Lin, H. C.(2010), ”Personnel Selection Using Analytic Network Process and Fuzzy Data Envelopment Analysis Approaches”, Computers & Industrial Engineering (59) 937-944. [11] Mahmoodzadeh.S, Shahrabi.J, Pariazar.M and Zaeri.M.S , “Project Selection by Using Fuzzy AHP and TOPSIS Technique”, World Academy of Science, Engineering and Technology 30,pp: 333338, 2007. [12] Mishra, A. (2009), “A study on customer satisfaction in Indian Retail Banking”, The IUP Journal of Management Research, Vol. VIII, No. 11, pp. 45-61 [13] Pal, M.N., and Choudhury (2009), “Exploring the dimensionality of service quality: An application of TOPSIS in the Indian Banking Industry”,Asia-Pacific Journal of Operational Research, 26 (1), 115-133 [14] Parasuraman.A, Zeithamal.V.A and Berry.L.L(1985), “SERVQUAL:A Multiple item scale for measuring customer perception of service quality”, Journal of Retailing, 64(1),12 - 40. [15] Rust.R.T and Oliver.R.C (1994), “Service quality: Insights and managerial implications from the frontier”. In Service Quality: New Directions in “Theory and Practice”. [16] Saaty.T.L, The Analytic Hierarchy Process, McGraw-Hill, New York, 1980. International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 4, December 2015 84 [17] Saaty.T.L (1990), ”How To Make A Decision: the Analytic Hierarchy Process. European Journal of Operation Research”,48(1), 9-26. [18] Shih, S.H., Shyur, H.J., and Lee, E.S., (2007) An extension of TOPSIS for group decision making mathematical and Computer Modeling, 45, 801–813. [19] Ta, H. P., and Har, K. Y. (2000) A study of bank selection decisions in singapore using the analytical hierarchy process International Journal of Bank Marketing, 18(4), pp170-180. [20] Venkata Subbaiah.K, Durga Prasad. K.G, Uma Bharathi.K , Soma Sekhara Rao.K (2011). “Integrating factor analysis and Analytic hierarchy process for library service quality” International Journal for Quality research UDK - 005.6:02. Original Scientific Paper (1.01) [21] Zahedi.F, “The analytic Hierarchy process-A survey of the method and its applications,” Interfaces, Vol. 16, pp.343-350, 1977. [22] Zavadskas.E.K and Turskis.Z (2011) Multiple Criteria Decision Making (MCDM) Methods In Economics: an Overview. Technological and Economic Development of Economy, 17(2), 397427.