- Email: [email protected]
A study on process evaluation and selection model for business process management
Expert Systems with Applications 38 (2011) 6339–6350
Contents lists available at ScienceDirect
Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa
A study on process evaluation and selection model for business process management Chiwoon Cho ⇑, Seungsin Lee School of Industrial Engineering, University of Ulsan, Ulsan 680-749, Republic of Korea
a r t i c l e
i n f o
Keywords: Business process management Fuzzy AHP Balanced scorecard Process selection criteria
a b s t r a c t Currently, BPM is considered as the suitable framework for today’s process-centric trends and BPM may result in considerable rewards for companies adopting it. For successful BPM initiative, the selection of suitable processes for BPM is very important. However, it is difficult to evaluate systematically and reasonably business processes for enterprises that plan to introduce BPM. This paper describes a web-based business process evaluation model based on BSC and fuzzy AHP for BPM. A web-based business process evaluation system was implemented and it provides impartial and reasonable results to enterprises or concerned persons that have insufficient experience and knowledge about BPM. Thus, this paper demonstrates the applicability of fuzzy AHP and BSC concepts in business process evaluation and selection for BPM, and provides a systemic guidance in the decision-making process. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction Many surveys indicated that the gap between IT and business is growing and that might signal a change in how enterprise technology is run. There are also increasing reports of IT not meeting business needs. Therefore, there is a need for better communication and understanding between IT and Business. To overcome this gap, many companies emphasis on the importance of business processes and the role of IT. BPM (business process management) has emerged as a new breed of process-centric approaches for companies that consider processes to be fundamental business assets (Davenport, 1992; Hammer & Champy, 1993; Lee & Dale, 1998). However, to be fully effective, BPM must not be approached as simply another IT toolset but rather as an environment where a business-process-oriented view is the means of communicating business requirements throughout the organization. BPM solutions hold the promise of bringing a process-centric approach to IT solutions (Sinur, 2003; Smith & Finger, 2003). BPM uses a technology designed specifically to manage business processes. These BPM systems are an enabler of business innovation because of the dramatic potential for improving the performance and agility of companies (Verner, 2004). In BPM initiative, the selection of suitable processes for BPM is one of the most important issues. However, because of the wide variety of business processes existing, the selection of business process for BPM is extremely difficult and time-consuming task. The ⇑ Corresponding author. Tel.: +82 52 259 2287; fax: +82 52 259 2180. E-mail address: [email protected] (C. Cho). 0957-4174/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2010.11.105
other important factors contributing to the complexity of process selection are evaluation criteria imposed by the business characteristics of the individual organization and the existence of some level of uncertainty due to vagueness or fuzziness of the evaluation criteria. Owing to the unstructured nature of the problem, there are not many research works related to process selection problem (Weck, Klocke, Schell, & Ruenauver, 1997). Most are incomplete prototypes that consider only a limited number of evaluation criteria, and are not very successful to deal with the qualitative factors associated with the problem. In addition, very little attention has been paid for selecting business process for BPM from a set of alternatives under fuzzy environment. Furthermore, there are no reports of the deployment of such process evaluation tools for BPM or platforms on the web to make it widely accessible to potential users. This paper presents a web-based model called Biz_Tower for business process evaluation and selection for BPM initiative. The model employs BSC (balanced scorecard) and Fuzzy AHP (analytic hierarchy process), and decision algorithms to identify an appropriate solution. An example is given to demonstrate the application of Biz_Tower. 2. Related works 2.1. BPM process selection and BSC Many guides and considerations related to the process selection for BPM were presented from companies and research institutes (Kim, Cho, & Kim, 1994; Park, 1995; Weck et al., 1997). However, the standards or the standardized methods for BPM process selection do not exist and the same selection criteria also have a
6340
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
difference according to the characteristic of business and business environment of the target company. Table 1 summarizes the process selection criteria of the related works. However, there are limitations related to the BPM process selection in the set of work reported above. Most are incomplete prototypes that consider only a limited number of factors and the factors are so conceptual. Hence, more robust guides and standardized methodologies, which can evaluate both tangible and intangible factors and overcome the drawbacks, are required. BSC can be considered as an excellent tool for evaluating the BPM processes. BSC makes it possible to evaluate managerial processes and activities with unbiased viewpoints by providing both tangible financial aspect and intangible, non-financial aspects: customer, internal business process, and learning and growth. BSC, as illustrated in Fig. 1, is widely recognized and used (Kaplan & Norton, 1992, 1993, 1996). This framework views an organization’s performance from four key perspectives, with regard to which organizations should articulate their core vision, strategy, and goals before translating them into specific initiatives, targets, and measures (Butler, Letza, & Neale, 1997; Letza, 1996; Martinsons, Davison, & Dennis, 1999; Smith, 2007).
Table 1 Process selection criteria. References Davenport (1992)
Hammer and Champy (1993)
Lee and Dale (1998) Park (1995) Choi, Lee, and Choi (2005)
In this research, BSC is utilized for the BPM process selection to pursue overall optimization through a balanced view of various perspectives and add values by providing relevant and balanced information for decision makers. 2.2. Fuzzy AHP The fuzzy theory, which originated with Zadeh (1975), allows for the existence of some level of uncertainty in decision-making due to vagueness or fuzziness rather than due to randomness alone (Chen, 1996; Cheng, 1999; Cheng & Mon, 1994; Mon, Cheng, & Lin, 1994). The theory of fuzzy has evolved in various directions, for examples; treating fuzzy sets as precisely defined mathematical objects, and the linguistic approach. Based on those directions, fuzzy theory has been applied in a variety of fields in the last four decades (Bozdah, Kahraman, & Ruah, 2003; Weck et al., 1997). The AHP has become one of the most widely used multiplecriteria decision-making (MCDM) methods, and has been used to solve unstructured problems in a variety of areas (Chang, 1996; Lee & Kim, 2005). In this paper, the fuzzy theory is combined with the AHP to determine the criteria-values of each proposed process for BPM as well as the evaluation results of the proposed processes. 3. BSC-based process evaluation criteria for BPM
Process selection criteria Needs to Strategic importance improve Difficulty of Process scope improvement Many conflicts/high frequency/excessive non-structured communication Competition outperforming Continuous incremental improvements Most fundamental Most interactive Characteristic of process itself BPR performance ability BPM Business impact adequateness Implementation feasibility
BSC was applied to define the process evaluation criteria for BPM in this research because BSC has been the most suggestions for developing a framework for performance measurement and management. In addition, a good balanced scorecard contains several strategic or future-focused metrics, that tell the organization how it is doing, on its path towards its vision (Kaplan & Norton, 1992, 1993, 1996). We first base on the four perspectives of the BSC to prepare a list of evaluation criteria, and then have a several interviews with BPM experts. A questionnaire is designed using conventional AHP questionnaire format (Butler et al., 1997; Letza, 1996; Martinsons et al., 1999) and the four perspectives of BSC and the selected evaluation criteria are included as shown in Table 2. Each of the
Fig. 1. BSC approach (Kaplan & Norton, 1996).
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350 Table 2 Evaluation criteria hierarchy. Perspectives
Criteria
Financial
Process cost
Process trend Process budget
Customer
Agility
Customer management Satisfaction
Internal business
Learning and growth
Monitoring and management
Detailed criteria -
Visualization
-
Complexity
-
Productivity
Objective Potential
-
Direct cost Delay cost Input man-hour and position Frequency Number of occurrence Budget Appropriateness of budget and processing expenses Easiness of process change Necessity of rapid response to market Adequacy of processing time Level of customer segmentation Whether contact task with customer Reliability Assurance Responsiveness Service level Necessity of real-time monitoring Possibility of continuous improvement Level of process standardization Clearness of role and responsibility Complexity of task/ organization/IT system/interface Easiness of BPM implementation Adequacy of task assignment Task redundancy Level of manual Correlation with strategic goals Clearness of purpose Previous similar references Influence on other processes
four perspectives should be translated into corresponding metrics and measures that reflect strategic goals and objectives. For example, Table 3 shows the criteria matrix for financial perspective. The evaluation criteria can be changed according to the business characteristics of the individual organization and should be reviewed periodically and updated as necessary. A total of 29 questions are made in a questionnaire and the questionnaires are distributed to concerned persons. Quantitative and qualitative data collected through the questionnaire are used as the basic data for the fuzzy AHP described in next chapter. 4. Proposed model A business process evaluation model based on fuzzy AHP is explained in this chapter. 4.1. AHP In general, the procedures of AHP involve the following six essential steps (Chang, 1996; Cheng et al., 1994; Lee & Kim, 2005; Mon et al., 1994; Park, 1995; Weck et al., 1997).
6341
– Define the unstructured problem and state clearly the objectives and outcomes. – Decompose the complex problem into a hierarchical structure with decision elements (criteria, detailed criteria, and alternatives). – Employ pair-wise comparisons among decision elements and form comparison matrices. – Use the eigenvalue method to estimate the relative weights of the decision elements. – Check the consistency property of matrices to ensure that the judgments of decision makers are consistent. – Aggregate the relative weights of decision elements to obtain an overall rating for the alternatives. In this research, AHP is used to generate the weighting of the four perspectives of the BSC and the weighting of the evaluation criteria. 4.1.1. Evaluation hierarchy As shown in Fig. 2, the hierarchical structure with perspectives and evaluation criteria (e.g., criteria and detailed criteria) is formulated. After constructing a hierarchy, each decision maker is asked to express relative importance of decision elements in the same level on a pair wise basis by a nine-point scale. 4.1.2. Weighting of the perspectives and the criteria The AHP is known as an eigenvector method. It means that the eigenvector corresponding to the largest eigenvalue of the pair wise comparisons matrix provides the relative priorities of the factors, and preserves ordinal preferences among the alternatives (Chang, 1996; Lee & Kim, 2005; Weck et al., 1997). In this study, the comparison of the importance of one perspective, criteria, sub-criteria over another was done by with help of BPM-TFT (task force team). As shown in Table 4, the responses collected with consensus of BPM-TFT can be used as input to the fuzzy AHP model. A square matrix is formed, as shown in Table 5, when every two perspectives are compared. The matrix has the property of the element Xij = 1/Xij. That means, if item i is 2 times as important as item j, then item j is 1/2 as important as item i. Secondly, the values are normalized by dividing the sum of each column into the elements of each column so that the sum of each column equals one as shown in Table 6. Lastly, the relative priorities for the perspectives are ascertained by dividing the number of elements (e.g., 4) in the row into the sum of the elements in each resulting row guaranteeing that the sum of relative priorities of factors which have the same parent node always equals one. These results indicate BPM-TFT consider customer perspective is most important with the highest priority at the 1st layer of decision hierarchy. The consistency property of the matrix is then checked to ensure the consistency of judgments in the pair-wise comparison. The consistency index (CI) and consistency ratio (CR) are defined as (Saaty, 1980):
Table 3 Criteria matrix for financial perspective. Perspectives
Criteria
Detailed criteria
Questions/measurement
Financial
Process cost
Direct cost Delay cost Input man-hour and position Frequency Number of occurrence Budget Appropriateness of budget and processing expenses
Ave. direct cost per transaction/As-Is analysis Ave. delay cost per unit time/As-Is analysis Ave. input man-hour per transaction and position/As-Is analysis Frequency of transaction/As-Is analysis Annual number of occurrence/As-Is analysis Budget assigned/As-Is analysis Questionnaire
Process trend Process budget
6342
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
Fig. 2. Decision hierarchy of BPM process selection.
Table 4 Example of pair-wise rating of perspectives.
Table 6 Relative importance of perspectives.
Layer
Perspective 1
Perspective 2
Inputs from BPM-TFT
1st layer
Financial Financial Financial Customer Customer Internal business
Customer Internal business Learning and growth Internal business Learning and growth Learning and growth
4:6 4:6 6:4 7:3 5:5 6:4
Table 5 Pair-wise comparison matrix. Customer
Internal business
Learning and growth
Financial Customer Internal business Learning and growth
1 3 3 0.3333
0.3333 1 0.2000 1
0.3333 5 1 0.3333
3 1 3 1
Sum
7.3333
2.5333
6.6667
8.0000
bmax n n1 CI CR ¼ RI
Customer
Internal business
Learning and growth
Weight
Financial Customer Internal business Learning and growth
0.1364 0.4091 0.4091 0.0455
0.1316 0.3947 0.0789 0.3947
0.0500 0.7500 0.1500 0.0500
0.3750 0.1250 0.3750 0.1250
0.1732 0.4197 0.2533 0.1538
Sum
1.00
1.00
1.00
1.00
1.00
4.2. Fuzzy
Financial
CI ¼
Financial
ð1Þ ð2Þ
where bmax is the largest eigenvalue of the matrix and n is the number of items being compared in the matrix. The CR is used to estimate directly the consistency of pair-wise comparisons. The CR is computed by dividing the CI by a value obtained from a table of random consistency index (RI). If the CR is less than 0.1, the comparisons are acceptable, otherwise not. In a similar fashion, the relative priorities for the criteria with respect to the perspectives and the detailed criteria with respect to the criteria are calculated. Based on this approach, the normalized priority weights of decision criteria can be determined and those are used to form the evaluation matrix (e.g., fuzzy judgment matrix) showing the final set of scores of the processes for BPM implementation.
Fuzzy AHP is designed to an alternative selection and justification problem by integrating the concept of fuzzy set theory and AHP. The use of fuzzy set theory allows the decision maker to incorporate both qualitative and quantitative data into the decision model. For this reason, decision makers usually feel more confident to give interval judgments rather than fixed value judgments. In this study, triangular fuzzy numbers are utilized to represent subjective judgments in order to capture the vagueness. 4.2.1. Triangular fuzzy number theory A fuzzy number is a special fuzzy set A = {(x, lA(x)), x 2 R}, where x takes its values on the real line, R: 1 < x < + 1 and lA(x) is a continuous mapping from R to the closed interval [0, 1]. A triangular fuzzy number denoted as A = (a, b, c), where a 6 b 6 c, has the following triangular type membership function;
8 0; > > > xa > > > < ba ; lA ðxÞ ¼ 1; > > cx > ; > > cb > : 0;
xa axb x¼b
ð3Þ
bxc cx
A triangular fuzzy number, A, is defined with the corresponding membership function as shown in Fig. 3. Alternatively, by defining the interval of confidence level a, the triangular fuzzy number can be characterized as
6343
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
4.2.2. Normalization of quantitative attributes Some elements in a fuzzy set may represent objective factors such as delay cost, direct cost, and input man-hour in Table 2. In case of the input man-hour, the membership function would represent the total man-hour required to perform the process. In essence, the membership function maps the total man-hour required associated with each process to the range between 0 and 1 as shown in Fig. 4. The less total man-hour required the process, the closer to 0 its degree of membership becomes and vice versa. It depends on how we define the fuzzy sets and the corresponding membership functions for quantitative attributes in advance. 4.3. The steps for BPM process evaluation with a case
Fig. 3. Triangular fuzzy number of fuzzy membership function.
8a 2 ½0; 1
ð4Þ
Aa ¼ ½aaL ; aaR ¼ ½ðb aÞa þ a; ðc bÞa þ c
a-Cut is known to incorporate the decision maker confidence over his/her preference of judgments. That is to say, because the judgment may be drawn by subjective and qualitative weights of decision makers, it may be considered that uncertainty can be contained. It will yield an interval set of values from a fuzzy number. For example, a = 0.5 will yield a set a0.5 = (aL, b, aR). In addition, degree of satisfaction for the judgment is estimated by the index of optimism k as L ¼ k½ðb aÞa þ a þ ð1 kÞ½ðc bÞa þ c
ð5Þ
The larger value of index k indicates the higher degree of optimism. While a is fixed, the fuzzy judgment matrix can be obtained after setting the index of optimism, k, in order to estimate the degree of satisfaction.
According to the above-mentioned process evaluation criteria and fuzzy AHP approach, a process evaluation model is developed as shown in Fig. 5. The goal here is to select the BPM target processes, satisfying all criteria in the best way. As a case study, the selection of the best BPM target process among a total of five sales processes of a construction equipment manufacturer was taken into consideration. This simplified example is chosen only for a better understanding of the main principles of the proposed model. In order to rate BPM target processes with respect to the criteria, triangular fuzzy numbers given in Table 7 is used in this case. In order to take the imprecision of human qualitative assessments into consideration, the triangular fuzzy numbers are defined with the corresponding membership function as shown in Fig. 6. Main steps for evaluating and selecting BPM target processes are as follows: Step 1: Construct the hierarchical structure with decision elements (e.g., BSC perspectives, criteria, and detailed criteria). A hierarchical structure is formed by using the overall goal as a root of the decision tree and making each major criterion a child as shown in Fig. 2. Step 2: Rate the relative importance (or priority) for each criterion among those which have the same parent node, and calculate the combined weight for the detailed criteria. Rating is
Table 7 Membership function of fuzzy number.
Fig. 4. Membership function for the input man-hour.
Fuzzy number
Membership function
1 x 9
(0, 0, 1) (x 1, x, x + 1) for x = 1, 2, 3, 4, 5, 6, 7, 8, 9 (8, 9, 9)
Fig. 5. BPM target process evaluation model.
6344
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350 Table 9 Fuzzy comparison matrix.
P1 P2 P3 P4 P5
Financial
Customer
Internal business
Learning and growth
(4.75, 5.56, 6.26) (2.41, 3.08, 4.08) (5.52, 6.48, 7.09) (4.19, 4.88, 5.73) (3.27, 4.19, 5.10)
(4.45, 5.45, 6.45) (4.53, 5.44, 6.44) (4.43, 5.43, 6.35) (5.13, 6.05, 6.83) (3.72, 4.50, 5.50)
(4.38, 5.38, 6.38) (4.87, 5.87, 6.84) (5.10, 6.10, 7.08) (4.75, 5.67, 6.67) (4.77, 5.77, 6.74)
(4.89, 5.89, 6.89) (4.50, 5.50, 6.50) (4.65, 5.65, 6.65) (5.15, 6.15, 7.15) (5.33, 6.33, 7.33)
Fig. 6. Fuzzy membership function for criteria.
done using the scaled pair-wise comparison method. As an example, Table 8 calculated the overall relative priorities and the combined weights associated with financial perspective. Step 3: Define the fuzzy sets and the corresponding membership functions for quantitative attributes. The membership functions map the values of quantitative attributes associated with each process to the range between 0 and 1 as shown in Fig. 4. Step 4: Construct the fuzzy comparison matrix. The fuzzy judgment matrix is constructed by using (3) and the fuzzy membership functions defined in Step 3. Table 9 shows the fuzzy judgment matrix of the five sales processes under consideration after multiplying the fuzzy number of each detailed criterion by the combined weight of the corresponding detailed criterion, and summing overall the detailed criteria. Step 5: Obtain the final raking. The evaluation values of the processes can be obtained by multiplying the fuzzy comparison matrix constructed in Step 4 by the local weights of each perspective, and summing over all the perspectives. Finally, after a-cut and k operations, the final ranking is obtained as shown in Table 10. In this example, the values, a = 0.5 and k = 0.5 are applied. According to Table 10, P3 (i.e., Process #3) is the best for BPM implementation and P5 is the worst.
5. Construction of web-based process evaluation system for BPM 5.1. System overview A web-based system called Biz_Tower for business process evaluation and selection for BPM is implemented based on the proposed model. Biz_Tower is written using JSP, MiPlatform(trial), Oracle 10g, and Apache Tomcat 6. It runs on any computer system with executable Internet Explorer. Biz_Tower consists of two areas largely: evaluation criteria management and process evaluation & selection. The main functions of the areas are briefly explained here:
Table 10 The final ranking of five sales processes.
Evaluation criteria management. 1. Add/delete/modify criteria: The basic information of criterion can be managed (i.e., newly entered/deleted/modified). 2. Providing process references: Some related process design examples are provided. Process evaluation and selection. 1. Data input: The responses of the questionnaire are stored including the relative importance between two decision elements. The results of As-Is analysis are also checked for quantitative factors. 2. Comparison priorities: Based on AHP concept, relative priorities are calculated. 3. Fuzzy comparison matrix: Based on the input of the questionnaire and relative priorities, fuzzy comparison matrices are formed. 4. Ranking: Obtain evaluation results of processes under consideration and final ranking of them. Biz_Tower provides final evaluation results and ranking of processes in terms of overall perspective and individual perspective as shown in Fig. 7. 5.2. Process evaluation for BPM on Biz_Tower A user can apply Biz_Tower after login on web as follows: Step 0: A new project can be established by selecting ‘‘Insert’’ and the name of the project can be entered as shown in Fig. 8.
Table 8 Relative importance of decision elements. Perspective
Perspective local weight
Criteria
Detailed criteria
Criteria local weight
Detailed criteria local weight
Combined weight
Financial
0.2052
Process cost
Direct cost Delay cost Input man-hour and position Frequency Number of occurrence Budget Appropriateness of budget and processing expenses
0.6864
0.2409 0.5485 0.2106 0.7500 0.2500 0.7500 0.2500
0.1654 0.3765 0.1446 0.1585 0.0528 0.0766 0.0255
Process trend Process budget
0.2114 0.1022
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
Fig. 7. Final evaluation values and ranking of processes.
Fig. 8. Establish a new project.
Fig. 9. Complete the evaluation hierarchy.
6345
6346
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
Step 1: Complete the evaluation hierarchy by selecting criteria to use as shown in Fig. 9. Step 2: The pair-wise comparison results of the user are input by selecting the number of the nine-point scale. Fig. 10 is an example of the pair-wise comparison results between the perspectives.
Step 3: The user is expected to type or provide a response to each question as indicated in Figs. 11 and 12. Step 4: After the all requested information is entered, the system can automatically generate fuzzy comparison matrix as shown in Fig. 13.
Fig. 10. Comparison results between perspectives.
Fig. 11. Data input for quantitative elements.
Fig. 12. Data input for qualitative elements.
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
6347
Fig. 13. Fuzzy comparison matrix.
Fig. 14. Final ranking of processes in terms of overall perspectives.
Fig. 15. Final ranking of processes in terms of financial perspective.
Step 5: The final evaluation values and ranking of processes in terms of overall perspectives are obtained as shown in Fig. 14, from which we get the best two processes: P3 and P1. The results in terms of individual perspective are also generated. As an example, Fig 15 shows the results of processes when only financial perspective is considered. In this case, the best two processes are P1 and P5. Users can apply the results to BPM implementation properly according to the business environment and conditions of their companies. In addition, the overall consistency ratio was also calculated by using (1) and (2) as 0.078 which is smaller than 0.1. It shows all of the judgments are consistent.
6. Result analysis In Biz_Tower, a database and user interface are designed and implemented. All data both entered via user interface and created during the evaluation are kept in a database so that they can be easily accessed for further analysis. Based on this data, the changes
of process evaluation results by adjusting the confidence level, a, and the index of optimism, k are explained in this chapter. The a-cut of a fuzzy number is the crisp set that contains all the elements of the universal set whose membership grades in the fuzzy number are greater than or equal to the specified value of a. The central value of a fuzzy number is the corresponding real crisp number and the spread of the number is the estimation from the real crisp number. The central value of a fuzzy number reacts to the real crisp number by adjusting the index of optimism, k. The changes of process evaluation values with a-cut and k applied are summarized in Table 11 and Fig. 16. According to Table 11 and Fig. 16, the less the degree of optimism k applied, the more the evaluation value of each process becomes and vice versa. When the confidence level a is set as 1.0 according to each degree of optimism k, there are no differences among the evaluation values. Fig. 17 shows the evaluation results of processes without a-cut applied according to each k. Fig. 18 shows the evaluation results of processes with a-cut applied as 0.7 according to each k. When a-cut is not applied, it is evaluated that P5 is more adequate than P1 when k is less than 0.3, but P1 is more adequate
6348
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
Table 11 Changes of process evaluation values with a-cut and k applied. Alpha
Lambda 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
P1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
6.24 6.16 6.07 5.98 5.89 5.80 5.71 5.62 5.53 5.44 5.35
6.06 5.99 5.92 5.85 5.78 5.71 5.53 5.56 5.49 5.42 5.35
5.87 5.82 5.77 5.72 5.67 5.61 5.56 5.51 5.46 5.40 5.35
5.59 5.56 5.62 5.59 5.55 5.52 5.49 5.45 5.42 5.39 5.35
5.50 5.49 5.47 5.46 5.44 5.43 5.41 5.40 5.38 5.37 5.35
5.32 5.32 5.33 5.33 5.33 5.34 5.34 5.34 5.35 5.35 5.35
5.13 5.16 5.18 5.20 5.22 5.26 5.26 5.29 5.31 5.33 5.35
4.95 4.99 5.03 5.07 5.11 5.15 5.19 5.23 5.27 5.31 5.35
4.76 4.82 4.88 4.94 5.00 5.06 5.12 5.18 5.23 5.29 5.35
4.58 4.66 4.73 4.81 4.89 4.96 5.04 5.12 5.20 5.27 5.35
4.39 4.49 4.58 4.68 4.78 4.87 4.97 5.06 5.16 5.26 5.35
P2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
5.97 5.87 5.77 5.67 5.57 5.47 5.37 5.27 5.17 5.07 4.97
5.78 5.70 5.52 5.54 5.46 5.38 5.30 5.22 5.14 5.06 4.97
5.60 5.53 5.47 5.41 5.35 5.28 5.22 5.16 5.10 5.04 4.97
5.41 5.35 5.32 5.28 5.23 5.19 5.15 5.10 5.06 5.02 4.97
5.22 5.19 5.17 5.14 5.12 5.10 5.07 5.05 5.02 5.00 4.97
5.03 5.02 5.02 5.01 5.01 5.00 5.00 4.99 4.98 4.98 4.97
4.84 4.85 4.87 4.88 4.89 4.91 4.92 4.93 4.95 4.96 4.97
4.65 4.58 4.71 4.75 4.78 4.81 4.84 4.88 4.91 4.94 4.97
4.46 4.51 4.56 4.61 4.67 4.72 4.77 4.82 4.87 4.92 4.97
4.27 4.34 4.41 4.48 4.55 4.62 4.59 4.76 4.83 4.90 4.97
4.08 4.17 4.26 4.35 4.44 4.53 4.62 4.71 4.80 4.88 4.97
P3
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
7.68 7.59 7.49 7.40 7.31 7.21 7.12 7.02 6.93 6.84 6.74
7.49 7.41 7.34 7.26 7.19 7.12 7.04 6.97 6.89 6.B2 6.74
7.30 7.24 7.19 7.13 7.07 7.02 6.96 6.91 6.85 6.80 6.74
7.11 7.07 7.03 7.00 6.96 6.92 6.89 6.85 6.81 6.78 6.74
6.91 6.90 6.88 6.86 6.84 6.83 6.81 6.79 6.78 6.76 6.74
6.72 6.72 6.73 6.73 6.73 6.73 6.73 6.74 6.74 6.74 6.74
6.53 6.55 6.57 6.59 6.61 6.64 6.66 6.68 6.70 6.72 6.74
6.34 6.38 6.42 6.46 6.50 6.54 6.58 6.52 6.66 6.70 6.74
6.15 6.21 6.26 6.32 6.38 6.44 6.50 6.56 6.62 6.68 6.74
5.95 6.03 6.11 6.19 6.27 6.35 6.43 6.51 6.58 6.66 6.74
5.76 5.86 5.96 6.06 6.15 6.25 6.35 6.45 6.55 6.64 6.74
P4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
5.96 5.87 5.77 5.67 5.58 5.48 5.38 5.29 5.19 5.09 5.00
5.77 5.59 5.52 5.54 5.46 5.38 5.31 5.23 5.15 5.07 5.00
5.58 5.52 5.46 5.41 5.35 5.29 5.23 5.17 5.11 5.05 5.00
5.39 5.35 5.31 5.27 5.23 5.19 5.15 5.11 5.07 5.03 5.00
5.20 5.18 5.16 5.14 5.12 5.10 5.08 5.06 5.04 5.02 5.00
5.01 5.01 5.01 5.00 5.00 5.00 5.00 5.00 5.00 5.00 5.00
4.82 4.84 4.85 4.87 4.89 4.91 4.92 4.94 4.96 4.98 5.00
4.63 4.66 4.70 4.74 4.77 4.81 4.85 4.88 4.92 4.96 5.00
4.44 4.49 4.55 4.60 4.66 4.72 4.77 4.83 4.88 4.94 5.00
4.25 4.32 4.40 4.47 4.55 4.62 4.70 4.77 4.85 4.92 5.00
4.06 4.15 4.24 4.34 4.43 4.53 4.62 4.71 4.81 4.90 5.00
P5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
6.30 6.20 6.10 6.00 5.90 5.80 5.70 5.60 5.50 5.40 5.30
5.10 6.02 5.94 5.85 5.78 5.70 5.52 5.54 5.45 5.38 5.30
5.90 5.84 5.78 5.72 5.66 5.60 5.54 5.48 5.42 5.36 5.30
5.70 5.56 5.52 5.58 5.54 5.50 5.46 5.42 5.38 5.34 5.30
5.50 5.48 5.46 5.44 5.42 5.40 5.38 5.36 5.34 5.32 5.30
5.30 5.30 5.30 5.30 5.30 5.30 5.30 5.30 5.30 5.30 5.30
5.10 5.12 5.14 5.16 5.18 5.20 5.22 5.24 5.26 5.28 5.30
4.90 4.94 4.98 5.02 5.06 5.10 5.14 5.18 5.22 5.26 5.30
4.70 4.76 4.82 4.88 4.94 5.00 5.06 5.12 5.18 5.24 5.30
4.50 4.58 4.56 4.74 4.82 4.90 4.98 5.06 5.14 5.22 5.30
4.30 4.40 4.50 4.60 4.70 4.80 4.90 5.00 5.10 5.20 5.30
than P5 when k is more than 0.5. When a-cut is applied as 0.7, we can see that P1 is always more adequate than P5. As a result, the final evaluation results and ranking of processes under consideration can be different according to user environment. The confidence level a and the index of optimism k have to be determined by expert group (i.e., BPM-TFT) of the target company. 7. Conclusion In this research, a model based on BSC and fuzzy AHP for evaluation and selection of business processes for BPM is proposed
and a web-based system called Biz_Tower is constructed to execute the model. The decision hierarchy is structured by the four major perspectives of BSC including financial, customer, internal business, and learning and growth, followed by criteria and detailed criteria. AHP is also used to generate the weighting of the four perspectives and the weighting of the evaluation criteria. Because human decision-making process usually contains vagueness, fuzzy theory is adopted to solve this problem. Some distinguished contributions of this research are as follows:
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
6349
Fig. 16. Overall evaluation results of processes.
Fig. 17. Process evaluation values without a-cut.
1. This research adopts the concept of BSC to construct a decision hierarchy for evaluation and selection of business processes for BPM. Based on interview with experts in this field and experience, a total of 29 detailed evaluation elements are finalized. 2. AHP and fuzzy set theory are adopted to develop a systematic evaluation model.
Fig. 18. Process evaluation values with a-cut applied as 0.7.
3. The a-cut and k operations are also implemented to consider a diversity of user environment and to compare the solutions obtained. 4. The comprehensive design nature of the system combined with the deployment on the web to make it accessible worldwide is what distinguishes Biz_Tower from all prior computer-aided
6350
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
BPM process selection systems that have been developed and reported in open literature. It provides impartial and reasonable results to enterprises or concerned persons that have insufficient experience and knowledge about BPM. References Bozdah, C. E., Kahraman, C., & Ruan, D. (2003). Fuzzy group decision making for selection among computer integrated manufacturing systems. Computers in Industry, 51, 13–29. Butler, A., Letza, S. R., & Neale, B. (1997). Linking the Balanced Scorecard to Strategy. Long Range Planning, 30(2), 242–253. Chang, D. Y. (1996). Applications of the extent analysis on fuzzy AHP. European Journal of Operational Research, 95, 649–655. Chen, S. M. (1996). Evaluating weapon systems using fuzzy arithmetic operations. Fuzzy Sets and Systems, 77, 265–276. Cheng, C. H., & Mon, D. L. (1994). Evaluating weapon system by analytical hierarchy process based on fuzzy scales. Fuzzy Sets and Systems, 63, 1–10. Cheng, C. H. (1999). Evaluating weapon systems using ranking fuzzy numbers. Fuzzy Sets and Systems, 107, 25–35. Choi, J. H., Lee, J. M., & Choi, H. J. (2005). A case study on the process selection and implementation of business process management in mobile telecommunication company. Information Systems Review, 8(1), 25–41. Davenport, T. H. (1992). Process innovation: Reengineering work through information technology. Harvard Business School. p. 337. Hammer, M., & Champy, J. (1993). Reengineering the corporation: A manifesto for business revolution. Harper Business. Kaplan, R. S., & Norton, D. P. (1992). The balanced scorecard measures that drive performance. Havard Business Review, 71–80. Kaplan, R. S., & Norton, D. P. (1993). Putting the balanced scorecard to work. Havard Business Review, 2–16.
Kaplan, R. S., & Norton, D. P. (1996). Using the balanced scorecard as a strategic management system. Havard Business Review, 75–85. Kim, S. H., Cho, Y. H., & Kim, J. K. (1994). Dynamic reengineering: The practical strategy of management innovation. The Korea Economic Daily, 91–111. Lee, S. C., & Kim, S. Y. (2005). Study on the solution selection model of information systems using fuzzy AHP. Entrue Journal of Information Technology, 4(1), 79–89. Lee, R. G., & Dale, B. G. (1998). Business process management: A review and evaluation. Business Process Management Journal, 4(3), 214–225. Letza, S. R. (1996). The design and implementation of the balanced business scorecard: An analysis of three companies in practice. Business Process Reengineering and Management Journal, 15(4), 245–266. Martinsons, M., Davison, R., & Dennis, T. (1999). The balanced scorecard: A foundation for the strategic management of information systems. Decision Support Systems, 25, 71–88. Mon, D. L., Cheng, C. H., & Lin, J. C. (1994). Evaluating weapon system using fuzzy analytic hierarchy process based on entropy weight. Fuzzy Sets and Systems, 62, 127–134. Park, D. J. (1995). A study on the determination of process selection factors in business reengineering. The Journal of Southern Seoul University, 1, 233–248. Saaty, T. L. (1980). The analytic hierarchy process. McGraw-Hill. Sinur, J. (2003). The business process management scenario. Gartner. Smith, H., & Finger, P. (2003). Business process management: The third wave. MeghanKiffer Press. Smith, R. F. (2007). Business process management and the balanced scorecard. John & Sons. Verner, L. (2004). BPM: The promise and the challenge. QUEUE. Available from http:// www.acmqueue.com. Weck, M., Klocke, F., Schell, H., & Ruenauver, E. (1997). Evaluating alternative production cycles using the extended fuzzy AHP method. European Journal of Operational Research, 100, 351–365. Zadeh, L. A. (1975). Fuzzy logic and approximate reasoning. Synthese, 30, 407–428.
Contents lists available at ScienceDirect
Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa
A study on process evaluation and selection model for business process management Chiwoon Cho ⇑, Seungsin Lee School of Industrial Engineering, University of Ulsan, Ulsan 680-749, Republic of Korea
a r t i c l e
i n f o
Keywords: Business process management Fuzzy AHP Balanced scorecard Process selection criteria
a b s t r a c t Currently, BPM is considered as the suitable framework for today’s process-centric trends and BPM may result in considerable rewards for companies adopting it. For successful BPM initiative, the selection of suitable processes for BPM is very important. However, it is difficult to evaluate systematically and reasonably business processes for enterprises that plan to introduce BPM. This paper describes a web-based business process evaluation model based on BSC and fuzzy AHP for BPM. A web-based business process evaluation system was implemented and it provides impartial and reasonable results to enterprises or concerned persons that have insufficient experience and knowledge about BPM. Thus, this paper demonstrates the applicability of fuzzy AHP and BSC concepts in business process evaluation and selection for BPM, and provides a systemic guidance in the decision-making process. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction Many surveys indicated that the gap between IT and business is growing and that might signal a change in how enterprise technology is run. There are also increasing reports of IT not meeting business needs. Therefore, there is a need for better communication and understanding between IT and Business. To overcome this gap, many companies emphasis on the importance of business processes and the role of IT. BPM (business process management) has emerged as a new breed of process-centric approaches for companies that consider processes to be fundamental business assets (Davenport, 1992; Hammer & Champy, 1993; Lee & Dale, 1998). However, to be fully effective, BPM must not be approached as simply another IT toolset but rather as an environment where a business-process-oriented view is the means of communicating business requirements throughout the organization. BPM solutions hold the promise of bringing a process-centric approach to IT solutions (Sinur, 2003; Smith & Finger, 2003). BPM uses a technology designed specifically to manage business processes. These BPM systems are an enabler of business innovation because of the dramatic potential for improving the performance and agility of companies (Verner, 2004). In BPM initiative, the selection of suitable processes for BPM is one of the most important issues. However, because of the wide variety of business processes existing, the selection of business process for BPM is extremely difficult and time-consuming task. The ⇑ Corresponding author. Tel.: +82 52 259 2287; fax: +82 52 259 2180. E-mail address: [email protected] (C. Cho). 0957-4174/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2010.11.105
other important factors contributing to the complexity of process selection are evaluation criteria imposed by the business characteristics of the individual organization and the existence of some level of uncertainty due to vagueness or fuzziness of the evaluation criteria. Owing to the unstructured nature of the problem, there are not many research works related to process selection problem (Weck, Klocke, Schell, & Ruenauver, 1997). Most are incomplete prototypes that consider only a limited number of evaluation criteria, and are not very successful to deal with the qualitative factors associated with the problem. In addition, very little attention has been paid for selecting business process for BPM from a set of alternatives under fuzzy environment. Furthermore, there are no reports of the deployment of such process evaluation tools for BPM or platforms on the web to make it widely accessible to potential users. This paper presents a web-based model called Biz_Tower for business process evaluation and selection for BPM initiative. The model employs BSC (balanced scorecard) and Fuzzy AHP (analytic hierarchy process), and decision algorithms to identify an appropriate solution. An example is given to demonstrate the application of Biz_Tower. 2. Related works 2.1. BPM process selection and BSC Many guides and considerations related to the process selection for BPM were presented from companies and research institutes (Kim, Cho, & Kim, 1994; Park, 1995; Weck et al., 1997). However, the standards or the standardized methods for BPM process selection do not exist and the same selection criteria also have a
6340
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
difference according to the characteristic of business and business environment of the target company. Table 1 summarizes the process selection criteria of the related works. However, there are limitations related to the BPM process selection in the set of work reported above. Most are incomplete prototypes that consider only a limited number of factors and the factors are so conceptual. Hence, more robust guides and standardized methodologies, which can evaluate both tangible and intangible factors and overcome the drawbacks, are required. BSC can be considered as an excellent tool for evaluating the BPM processes. BSC makes it possible to evaluate managerial processes and activities with unbiased viewpoints by providing both tangible financial aspect and intangible, non-financial aspects: customer, internal business process, and learning and growth. BSC, as illustrated in Fig. 1, is widely recognized and used (Kaplan & Norton, 1992, 1993, 1996). This framework views an organization’s performance from four key perspectives, with regard to which organizations should articulate their core vision, strategy, and goals before translating them into specific initiatives, targets, and measures (Butler, Letza, & Neale, 1997; Letza, 1996; Martinsons, Davison, & Dennis, 1999; Smith, 2007).
Table 1 Process selection criteria. References Davenport (1992)
Hammer and Champy (1993)
Lee and Dale (1998) Park (1995) Choi, Lee, and Choi (2005)
In this research, BSC is utilized for the BPM process selection to pursue overall optimization through a balanced view of various perspectives and add values by providing relevant and balanced information for decision makers. 2.2. Fuzzy AHP The fuzzy theory, which originated with Zadeh (1975), allows for the existence of some level of uncertainty in decision-making due to vagueness or fuzziness rather than due to randomness alone (Chen, 1996; Cheng, 1999; Cheng & Mon, 1994; Mon, Cheng, & Lin, 1994). The theory of fuzzy has evolved in various directions, for examples; treating fuzzy sets as precisely defined mathematical objects, and the linguistic approach. Based on those directions, fuzzy theory has been applied in a variety of fields in the last four decades (Bozdah, Kahraman, & Ruah, 2003; Weck et al., 1997). The AHP has become one of the most widely used multiplecriteria decision-making (MCDM) methods, and has been used to solve unstructured problems in a variety of areas (Chang, 1996; Lee & Kim, 2005). In this paper, the fuzzy theory is combined with the AHP to determine the criteria-values of each proposed process for BPM as well as the evaluation results of the proposed processes. 3. BSC-based process evaluation criteria for BPM
Process selection criteria Needs to Strategic importance improve Difficulty of Process scope improvement Many conflicts/high frequency/excessive non-structured communication Competition outperforming Continuous incremental improvements Most fundamental Most interactive Characteristic of process itself BPR performance ability BPM Business impact adequateness Implementation feasibility
BSC was applied to define the process evaluation criteria for BPM in this research because BSC has been the most suggestions for developing a framework for performance measurement and management. In addition, a good balanced scorecard contains several strategic or future-focused metrics, that tell the organization how it is doing, on its path towards its vision (Kaplan & Norton, 1992, 1993, 1996). We first base on the four perspectives of the BSC to prepare a list of evaluation criteria, and then have a several interviews with BPM experts. A questionnaire is designed using conventional AHP questionnaire format (Butler et al., 1997; Letza, 1996; Martinsons et al., 1999) and the four perspectives of BSC and the selected evaluation criteria are included as shown in Table 2. Each of the
Fig. 1. BSC approach (Kaplan & Norton, 1996).
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350 Table 2 Evaluation criteria hierarchy. Perspectives
Criteria
Financial
Process cost
Process trend Process budget
Customer
Agility
Customer management Satisfaction
Internal business
Learning and growth
Monitoring and management
Detailed criteria -
Visualization
-
Complexity
-
Productivity
Objective Potential
-
Direct cost Delay cost Input man-hour and position Frequency Number of occurrence Budget Appropriateness of budget and processing expenses Easiness of process change Necessity of rapid response to market Adequacy of processing time Level of customer segmentation Whether contact task with customer Reliability Assurance Responsiveness Service level Necessity of real-time monitoring Possibility of continuous improvement Level of process standardization Clearness of role and responsibility Complexity of task/ organization/IT system/interface Easiness of BPM implementation Adequacy of task assignment Task redundancy Level of manual Correlation with strategic goals Clearness of purpose Previous similar references Influence on other processes
four perspectives should be translated into corresponding metrics and measures that reflect strategic goals and objectives. For example, Table 3 shows the criteria matrix for financial perspective. The evaluation criteria can be changed according to the business characteristics of the individual organization and should be reviewed periodically and updated as necessary. A total of 29 questions are made in a questionnaire and the questionnaires are distributed to concerned persons. Quantitative and qualitative data collected through the questionnaire are used as the basic data for the fuzzy AHP described in next chapter. 4. Proposed model A business process evaluation model based on fuzzy AHP is explained in this chapter. 4.1. AHP In general, the procedures of AHP involve the following six essential steps (Chang, 1996; Cheng et al., 1994; Lee & Kim, 2005; Mon et al., 1994; Park, 1995; Weck et al., 1997).
6341
– Define the unstructured problem and state clearly the objectives and outcomes. – Decompose the complex problem into a hierarchical structure with decision elements (criteria, detailed criteria, and alternatives). – Employ pair-wise comparisons among decision elements and form comparison matrices. – Use the eigenvalue method to estimate the relative weights of the decision elements. – Check the consistency property of matrices to ensure that the judgments of decision makers are consistent. – Aggregate the relative weights of decision elements to obtain an overall rating for the alternatives. In this research, AHP is used to generate the weighting of the four perspectives of the BSC and the weighting of the evaluation criteria. 4.1.1. Evaluation hierarchy As shown in Fig. 2, the hierarchical structure with perspectives and evaluation criteria (e.g., criteria and detailed criteria) is formulated. After constructing a hierarchy, each decision maker is asked to express relative importance of decision elements in the same level on a pair wise basis by a nine-point scale. 4.1.2. Weighting of the perspectives and the criteria The AHP is known as an eigenvector method. It means that the eigenvector corresponding to the largest eigenvalue of the pair wise comparisons matrix provides the relative priorities of the factors, and preserves ordinal preferences among the alternatives (Chang, 1996; Lee & Kim, 2005; Weck et al., 1997). In this study, the comparison of the importance of one perspective, criteria, sub-criteria over another was done by with help of BPM-TFT (task force team). As shown in Table 4, the responses collected with consensus of BPM-TFT can be used as input to the fuzzy AHP model. A square matrix is formed, as shown in Table 5, when every two perspectives are compared. The matrix has the property of the element Xij = 1/Xij. That means, if item i is 2 times as important as item j, then item j is 1/2 as important as item i. Secondly, the values are normalized by dividing the sum of each column into the elements of each column so that the sum of each column equals one as shown in Table 6. Lastly, the relative priorities for the perspectives are ascertained by dividing the number of elements (e.g., 4) in the row into the sum of the elements in each resulting row guaranteeing that the sum of relative priorities of factors which have the same parent node always equals one. These results indicate BPM-TFT consider customer perspective is most important with the highest priority at the 1st layer of decision hierarchy. The consistency property of the matrix is then checked to ensure the consistency of judgments in the pair-wise comparison. The consistency index (CI) and consistency ratio (CR) are defined as (Saaty, 1980):
Table 3 Criteria matrix for financial perspective. Perspectives
Criteria
Detailed criteria
Questions/measurement
Financial
Process cost
Direct cost Delay cost Input man-hour and position Frequency Number of occurrence Budget Appropriateness of budget and processing expenses
Ave. direct cost per transaction/As-Is analysis Ave. delay cost per unit time/As-Is analysis Ave. input man-hour per transaction and position/As-Is analysis Frequency of transaction/As-Is analysis Annual number of occurrence/As-Is analysis Budget assigned/As-Is analysis Questionnaire
Process trend Process budget
6342
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
Fig. 2. Decision hierarchy of BPM process selection.
Table 4 Example of pair-wise rating of perspectives.
Table 6 Relative importance of perspectives.
Layer
Perspective 1
Perspective 2
Inputs from BPM-TFT
1st layer
Financial Financial Financial Customer Customer Internal business
Customer Internal business Learning and growth Internal business Learning and growth Learning and growth
4:6 4:6 6:4 7:3 5:5 6:4
Table 5 Pair-wise comparison matrix. Customer
Internal business
Learning and growth
Financial Customer Internal business Learning and growth
1 3 3 0.3333
0.3333 1 0.2000 1
0.3333 5 1 0.3333
3 1 3 1
Sum
7.3333
2.5333
6.6667
8.0000
bmax n n1 CI CR ¼ RI
Customer
Internal business
Learning and growth
Weight
Financial Customer Internal business Learning and growth
0.1364 0.4091 0.4091 0.0455
0.1316 0.3947 0.0789 0.3947
0.0500 0.7500 0.1500 0.0500
0.3750 0.1250 0.3750 0.1250
0.1732 0.4197 0.2533 0.1538
Sum
1.00
1.00
1.00
1.00
1.00
4.2. Fuzzy
Financial
CI ¼
Financial
ð1Þ ð2Þ
where bmax is the largest eigenvalue of the matrix and n is the number of items being compared in the matrix. The CR is used to estimate directly the consistency of pair-wise comparisons. The CR is computed by dividing the CI by a value obtained from a table of random consistency index (RI). If the CR is less than 0.1, the comparisons are acceptable, otherwise not. In a similar fashion, the relative priorities for the criteria with respect to the perspectives and the detailed criteria with respect to the criteria are calculated. Based on this approach, the normalized priority weights of decision criteria can be determined and those are used to form the evaluation matrix (e.g., fuzzy judgment matrix) showing the final set of scores of the processes for BPM implementation.
Fuzzy AHP is designed to an alternative selection and justification problem by integrating the concept of fuzzy set theory and AHP. The use of fuzzy set theory allows the decision maker to incorporate both qualitative and quantitative data into the decision model. For this reason, decision makers usually feel more confident to give interval judgments rather than fixed value judgments. In this study, triangular fuzzy numbers are utilized to represent subjective judgments in order to capture the vagueness. 4.2.1. Triangular fuzzy number theory A fuzzy number is a special fuzzy set A = {(x, lA(x)), x 2 R}, where x takes its values on the real line, R: 1 < x < + 1 and lA(x) is a continuous mapping from R to the closed interval [0, 1]. A triangular fuzzy number denoted as A = (a, b, c), where a 6 b 6 c, has the following triangular type membership function;
8 0; > > > xa > > > < ba ; lA ðxÞ ¼ 1; > > cx > ; > > cb > : 0;
xa axb x¼b
ð3Þ
bxc cx
A triangular fuzzy number, A, is defined with the corresponding membership function as shown in Fig. 3. Alternatively, by defining the interval of confidence level a, the triangular fuzzy number can be characterized as
6343
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
4.2.2. Normalization of quantitative attributes Some elements in a fuzzy set may represent objective factors such as delay cost, direct cost, and input man-hour in Table 2. In case of the input man-hour, the membership function would represent the total man-hour required to perform the process. In essence, the membership function maps the total man-hour required associated with each process to the range between 0 and 1 as shown in Fig. 4. The less total man-hour required the process, the closer to 0 its degree of membership becomes and vice versa. It depends on how we define the fuzzy sets and the corresponding membership functions for quantitative attributes in advance. 4.3. The steps for BPM process evaluation with a case
Fig. 3. Triangular fuzzy number of fuzzy membership function.
8a 2 ½0; 1
ð4Þ
Aa ¼ ½aaL ; aaR ¼ ½ðb aÞa þ a; ðc bÞa þ c
a-Cut is known to incorporate the decision maker confidence over his/her preference of judgments. That is to say, because the judgment may be drawn by subjective and qualitative weights of decision makers, it may be considered that uncertainty can be contained. It will yield an interval set of values from a fuzzy number. For example, a = 0.5 will yield a set a0.5 = (aL, b, aR). In addition, degree of satisfaction for the judgment is estimated by the index of optimism k as L ¼ k½ðb aÞa þ a þ ð1 kÞ½ðc bÞa þ c
ð5Þ
The larger value of index k indicates the higher degree of optimism. While a is fixed, the fuzzy judgment matrix can be obtained after setting the index of optimism, k, in order to estimate the degree of satisfaction.
According to the above-mentioned process evaluation criteria and fuzzy AHP approach, a process evaluation model is developed as shown in Fig. 5. The goal here is to select the BPM target processes, satisfying all criteria in the best way. As a case study, the selection of the best BPM target process among a total of five sales processes of a construction equipment manufacturer was taken into consideration. This simplified example is chosen only for a better understanding of the main principles of the proposed model. In order to rate BPM target processes with respect to the criteria, triangular fuzzy numbers given in Table 7 is used in this case. In order to take the imprecision of human qualitative assessments into consideration, the triangular fuzzy numbers are defined with the corresponding membership function as shown in Fig. 6. Main steps for evaluating and selecting BPM target processes are as follows: Step 1: Construct the hierarchical structure with decision elements (e.g., BSC perspectives, criteria, and detailed criteria). A hierarchical structure is formed by using the overall goal as a root of the decision tree and making each major criterion a child as shown in Fig. 2. Step 2: Rate the relative importance (or priority) for each criterion among those which have the same parent node, and calculate the combined weight for the detailed criteria. Rating is
Table 7 Membership function of fuzzy number.
Fig. 4. Membership function for the input man-hour.
Fuzzy number
Membership function
1 x 9
(0, 0, 1) (x 1, x, x + 1) for x = 1, 2, 3, 4, 5, 6, 7, 8, 9 (8, 9, 9)
Fig. 5. BPM target process evaluation model.
6344
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350 Table 9 Fuzzy comparison matrix.
P1 P2 P3 P4 P5
Financial
Customer
Internal business
Learning and growth
(4.75, 5.56, 6.26) (2.41, 3.08, 4.08) (5.52, 6.48, 7.09) (4.19, 4.88, 5.73) (3.27, 4.19, 5.10)
(4.45, 5.45, 6.45) (4.53, 5.44, 6.44) (4.43, 5.43, 6.35) (5.13, 6.05, 6.83) (3.72, 4.50, 5.50)
(4.38, 5.38, 6.38) (4.87, 5.87, 6.84) (5.10, 6.10, 7.08) (4.75, 5.67, 6.67) (4.77, 5.77, 6.74)
(4.89, 5.89, 6.89) (4.50, 5.50, 6.50) (4.65, 5.65, 6.65) (5.15, 6.15, 7.15) (5.33, 6.33, 7.33)
Fig. 6. Fuzzy membership function for criteria.
done using the scaled pair-wise comparison method. As an example, Table 8 calculated the overall relative priorities and the combined weights associated with financial perspective. Step 3: Define the fuzzy sets and the corresponding membership functions for quantitative attributes. The membership functions map the values of quantitative attributes associated with each process to the range between 0 and 1 as shown in Fig. 4. Step 4: Construct the fuzzy comparison matrix. The fuzzy judgment matrix is constructed by using (3) and the fuzzy membership functions defined in Step 3. Table 9 shows the fuzzy judgment matrix of the five sales processes under consideration after multiplying the fuzzy number of each detailed criterion by the combined weight of the corresponding detailed criterion, and summing overall the detailed criteria. Step 5: Obtain the final raking. The evaluation values of the processes can be obtained by multiplying the fuzzy comparison matrix constructed in Step 4 by the local weights of each perspective, and summing over all the perspectives. Finally, after a-cut and k operations, the final ranking is obtained as shown in Table 10. In this example, the values, a = 0.5 and k = 0.5 are applied. According to Table 10, P3 (i.e., Process #3) is the best for BPM implementation and P5 is the worst.
5. Construction of web-based process evaluation system for BPM 5.1. System overview A web-based system called Biz_Tower for business process evaluation and selection for BPM is implemented based on the proposed model. Biz_Tower is written using JSP, MiPlatform(trial), Oracle 10g, and Apache Tomcat 6. It runs on any computer system with executable Internet Explorer. Biz_Tower consists of two areas largely: evaluation criteria management and process evaluation & selection. The main functions of the areas are briefly explained here:
Table 10 The final ranking of five sales processes.
Evaluation criteria management. 1. Add/delete/modify criteria: The basic information of criterion can be managed (i.e., newly entered/deleted/modified). 2. Providing process references: Some related process design examples are provided. Process evaluation and selection. 1. Data input: The responses of the questionnaire are stored including the relative importance between two decision elements. The results of As-Is analysis are also checked for quantitative factors. 2. Comparison priorities: Based on AHP concept, relative priorities are calculated. 3. Fuzzy comparison matrix: Based on the input of the questionnaire and relative priorities, fuzzy comparison matrices are formed. 4. Ranking: Obtain evaluation results of processes under consideration and final ranking of them. Biz_Tower provides final evaluation results and ranking of processes in terms of overall perspective and individual perspective as shown in Fig. 7. 5.2. Process evaluation for BPM on Biz_Tower A user can apply Biz_Tower after login on web as follows: Step 0: A new project can be established by selecting ‘‘Insert’’ and the name of the project can be entered as shown in Fig. 8.
Table 8 Relative importance of decision elements. Perspective
Perspective local weight
Criteria
Detailed criteria
Criteria local weight
Detailed criteria local weight
Combined weight
Financial
0.2052
Process cost
Direct cost Delay cost Input man-hour and position Frequency Number of occurrence Budget Appropriateness of budget and processing expenses
0.6864
0.2409 0.5485 0.2106 0.7500 0.2500 0.7500 0.2500
0.1654 0.3765 0.1446 0.1585 0.0528 0.0766 0.0255
Process trend Process budget
0.2114 0.1022
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
Fig. 7. Final evaluation values and ranking of processes.
Fig. 8. Establish a new project.
Fig. 9. Complete the evaluation hierarchy.
6345
6346
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
Step 1: Complete the evaluation hierarchy by selecting criteria to use as shown in Fig. 9. Step 2: The pair-wise comparison results of the user are input by selecting the number of the nine-point scale. Fig. 10 is an example of the pair-wise comparison results between the perspectives.
Step 3: The user is expected to type or provide a response to each question as indicated in Figs. 11 and 12. Step 4: After the all requested information is entered, the system can automatically generate fuzzy comparison matrix as shown in Fig. 13.
Fig. 10. Comparison results between perspectives.
Fig. 11. Data input for quantitative elements.
Fig. 12. Data input for qualitative elements.
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
6347
Fig. 13. Fuzzy comparison matrix.
Fig. 14. Final ranking of processes in terms of overall perspectives.
Fig. 15. Final ranking of processes in terms of financial perspective.
Step 5: The final evaluation values and ranking of processes in terms of overall perspectives are obtained as shown in Fig. 14, from which we get the best two processes: P3 and P1. The results in terms of individual perspective are also generated. As an example, Fig 15 shows the results of processes when only financial perspective is considered. In this case, the best two processes are P1 and P5. Users can apply the results to BPM implementation properly according to the business environment and conditions of their companies. In addition, the overall consistency ratio was also calculated by using (1) and (2) as 0.078 which is smaller than 0.1. It shows all of the judgments are consistent.
6. Result analysis In Biz_Tower, a database and user interface are designed and implemented. All data both entered via user interface and created during the evaluation are kept in a database so that they can be easily accessed for further analysis. Based on this data, the changes
of process evaluation results by adjusting the confidence level, a, and the index of optimism, k are explained in this chapter. The a-cut of a fuzzy number is the crisp set that contains all the elements of the universal set whose membership grades in the fuzzy number are greater than or equal to the specified value of a. The central value of a fuzzy number is the corresponding real crisp number and the spread of the number is the estimation from the real crisp number. The central value of a fuzzy number reacts to the real crisp number by adjusting the index of optimism, k. The changes of process evaluation values with a-cut and k applied are summarized in Table 11 and Fig. 16. According to Table 11 and Fig. 16, the less the degree of optimism k applied, the more the evaluation value of each process becomes and vice versa. When the confidence level a is set as 1.0 according to each degree of optimism k, there are no differences among the evaluation values. Fig. 17 shows the evaluation results of processes without a-cut applied according to each k. Fig. 18 shows the evaluation results of processes with a-cut applied as 0.7 according to each k. When a-cut is not applied, it is evaluated that P5 is more adequate than P1 when k is less than 0.3, but P1 is more adequate
6348
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
Table 11 Changes of process evaluation values with a-cut and k applied. Alpha
Lambda 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
P1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
6.24 6.16 6.07 5.98 5.89 5.80 5.71 5.62 5.53 5.44 5.35
6.06 5.99 5.92 5.85 5.78 5.71 5.53 5.56 5.49 5.42 5.35
5.87 5.82 5.77 5.72 5.67 5.61 5.56 5.51 5.46 5.40 5.35
5.59 5.56 5.62 5.59 5.55 5.52 5.49 5.45 5.42 5.39 5.35
5.50 5.49 5.47 5.46 5.44 5.43 5.41 5.40 5.38 5.37 5.35
5.32 5.32 5.33 5.33 5.33 5.34 5.34 5.34 5.35 5.35 5.35
5.13 5.16 5.18 5.20 5.22 5.26 5.26 5.29 5.31 5.33 5.35
4.95 4.99 5.03 5.07 5.11 5.15 5.19 5.23 5.27 5.31 5.35
4.76 4.82 4.88 4.94 5.00 5.06 5.12 5.18 5.23 5.29 5.35
4.58 4.66 4.73 4.81 4.89 4.96 5.04 5.12 5.20 5.27 5.35
4.39 4.49 4.58 4.68 4.78 4.87 4.97 5.06 5.16 5.26 5.35
P2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
5.97 5.87 5.77 5.67 5.57 5.47 5.37 5.27 5.17 5.07 4.97
5.78 5.70 5.52 5.54 5.46 5.38 5.30 5.22 5.14 5.06 4.97
5.60 5.53 5.47 5.41 5.35 5.28 5.22 5.16 5.10 5.04 4.97
5.41 5.35 5.32 5.28 5.23 5.19 5.15 5.10 5.06 5.02 4.97
5.22 5.19 5.17 5.14 5.12 5.10 5.07 5.05 5.02 5.00 4.97
5.03 5.02 5.02 5.01 5.01 5.00 5.00 4.99 4.98 4.98 4.97
4.84 4.85 4.87 4.88 4.89 4.91 4.92 4.93 4.95 4.96 4.97
4.65 4.58 4.71 4.75 4.78 4.81 4.84 4.88 4.91 4.94 4.97
4.46 4.51 4.56 4.61 4.67 4.72 4.77 4.82 4.87 4.92 4.97
4.27 4.34 4.41 4.48 4.55 4.62 4.59 4.76 4.83 4.90 4.97
4.08 4.17 4.26 4.35 4.44 4.53 4.62 4.71 4.80 4.88 4.97
P3
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
7.68 7.59 7.49 7.40 7.31 7.21 7.12 7.02 6.93 6.84 6.74
7.49 7.41 7.34 7.26 7.19 7.12 7.04 6.97 6.89 6.B2 6.74
7.30 7.24 7.19 7.13 7.07 7.02 6.96 6.91 6.85 6.80 6.74
7.11 7.07 7.03 7.00 6.96 6.92 6.89 6.85 6.81 6.78 6.74
6.91 6.90 6.88 6.86 6.84 6.83 6.81 6.79 6.78 6.76 6.74
6.72 6.72 6.73 6.73 6.73 6.73 6.73 6.74 6.74 6.74 6.74
6.53 6.55 6.57 6.59 6.61 6.64 6.66 6.68 6.70 6.72 6.74
6.34 6.38 6.42 6.46 6.50 6.54 6.58 6.52 6.66 6.70 6.74
6.15 6.21 6.26 6.32 6.38 6.44 6.50 6.56 6.62 6.68 6.74
5.95 6.03 6.11 6.19 6.27 6.35 6.43 6.51 6.58 6.66 6.74
5.76 5.86 5.96 6.06 6.15 6.25 6.35 6.45 6.55 6.64 6.74
P4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
5.96 5.87 5.77 5.67 5.58 5.48 5.38 5.29 5.19 5.09 5.00
5.77 5.59 5.52 5.54 5.46 5.38 5.31 5.23 5.15 5.07 5.00
5.58 5.52 5.46 5.41 5.35 5.29 5.23 5.17 5.11 5.05 5.00
5.39 5.35 5.31 5.27 5.23 5.19 5.15 5.11 5.07 5.03 5.00
5.20 5.18 5.16 5.14 5.12 5.10 5.08 5.06 5.04 5.02 5.00
5.01 5.01 5.01 5.00 5.00 5.00 5.00 5.00 5.00 5.00 5.00
4.82 4.84 4.85 4.87 4.89 4.91 4.92 4.94 4.96 4.98 5.00
4.63 4.66 4.70 4.74 4.77 4.81 4.85 4.88 4.92 4.96 5.00
4.44 4.49 4.55 4.60 4.66 4.72 4.77 4.83 4.88 4.94 5.00
4.25 4.32 4.40 4.47 4.55 4.62 4.70 4.77 4.85 4.92 5.00
4.06 4.15 4.24 4.34 4.43 4.53 4.62 4.71 4.81 4.90 5.00
P5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
6.30 6.20 6.10 6.00 5.90 5.80 5.70 5.60 5.50 5.40 5.30
5.10 6.02 5.94 5.85 5.78 5.70 5.52 5.54 5.45 5.38 5.30
5.90 5.84 5.78 5.72 5.66 5.60 5.54 5.48 5.42 5.36 5.30
5.70 5.56 5.52 5.58 5.54 5.50 5.46 5.42 5.38 5.34 5.30
5.50 5.48 5.46 5.44 5.42 5.40 5.38 5.36 5.34 5.32 5.30
5.30 5.30 5.30 5.30 5.30 5.30 5.30 5.30 5.30 5.30 5.30
5.10 5.12 5.14 5.16 5.18 5.20 5.22 5.24 5.26 5.28 5.30
4.90 4.94 4.98 5.02 5.06 5.10 5.14 5.18 5.22 5.26 5.30
4.70 4.76 4.82 4.88 4.94 5.00 5.06 5.12 5.18 5.24 5.30
4.50 4.58 4.56 4.74 4.82 4.90 4.98 5.06 5.14 5.22 5.30
4.30 4.40 4.50 4.60 4.70 4.80 4.90 5.00 5.10 5.20 5.30
than P5 when k is more than 0.5. When a-cut is applied as 0.7, we can see that P1 is always more adequate than P5. As a result, the final evaluation results and ranking of processes under consideration can be different according to user environment. The confidence level a and the index of optimism k have to be determined by expert group (i.e., BPM-TFT) of the target company. 7. Conclusion In this research, a model based on BSC and fuzzy AHP for evaluation and selection of business processes for BPM is proposed
and a web-based system called Biz_Tower is constructed to execute the model. The decision hierarchy is structured by the four major perspectives of BSC including financial, customer, internal business, and learning and growth, followed by criteria and detailed criteria. AHP is also used to generate the weighting of the four perspectives and the weighting of the evaluation criteria. Because human decision-making process usually contains vagueness, fuzzy theory is adopted to solve this problem. Some distinguished contributions of this research are as follows:
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
6349
Fig. 16. Overall evaluation results of processes.
Fig. 17. Process evaluation values without a-cut.
1. This research adopts the concept of BSC to construct a decision hierarchy for evaluation and selection of business processes for BPM. Based on interview with experts in this field and experience, a total of 29 detailed evaluation elements are finalized. 2. AHP and fuzzy set theory are adopted to develop a systematic evaluation model.
Fig. 18. Process evaluation values with a-cut applied as 0.7.
3. The a-cut and k operations are also implemented to consider a diversity of user environment and to compare the solutions obtained. 4. The comprehensive design nature of the system combined with the deployment on the web to make it accessible worldwide is what distinguishes Biz_Tower from all prior computer-aided
6350
C. Cho, S. Lee / Expert Systems with Applications 38 (2011) 6339–6350
BPM process selection systems that have been developed and reported in open literature. It provides impartial and reasonable results to enterprises or concerned persons that have insufficient experience and knowledge about BPM. References Bozdah, C. E., Kahraman, C., & Ruan, D. (2003). Fuzzy group decision making for selection among computer integrated manufacturing systems. Computers in Industry, 51, 13–29. Butler, A., Letza, S. R., & Neale, B. (1997). Linking the Balanced Scorecard to Strategy. Long Range Planning, 30(2), 242–253. Chang, D. Y. (1996). Applications of the extent analysis on fuzzy AHP. European Journal of Operational Research, 95, 649–655. Chen, S. M. (1996). Evaluating weapon systems using fuzzy arithmetic operations. Fuzzy Sets and Systems, 77, 265–276. Cheng, C. H., & Mon, D. L. (1994). Evaluating weapon system by analytical hierarchy process based on fuzzy scales. Fuzzy Sets and Systems, 63, 1–10. Cheng, C. H. (1999). Evaluating weapon systems using ranking fuzzy numbers. Fuzzy Sets and Systems, 107, 25–35. Choi, J. H., Lee, J. M., & Choi, H. J. (2005). A case study on the process selection and implementation of business process management in mobile telecommunication company. Information Systems Review, 8(1), 25–41. Davenport, T. H. (1992). Process innovation: Reengineering work through information technology. Harvard Business School. p. 337. Hammer, M., & Champy, J. (1993). Reengineering the corporation: A manifesto for business revolution. Harper Business. Kaplan, R. S., & Norton, D. P. (1992). The balanced scorecard measures that drive performance. Havard Business Review, 71–80. Kaplan, R. S., & Norton, D. P. (1993). Putting the balanced scorecard to work. Havard Business Review, 2–16.
Kaplan, R. S., & Norton, D. P. (1996). Using the balanced scorecard as a strategic management system. Havard Business Review, 75–85. Kim, S. H., Cho, Y. H., & Kim, J. K. (1994). Dynamic reengineering: The practical strategy of management innovation. The Korea Economic Daily, 91–111. Lee, S. C., & Kim, S. Y. (2005). Study on the solution selection model of information systems using fuzzy AHP. Entrue Journal of Information Technology, 4(1), 79–89. Lee, R. G., & Dale, B. G. (1998). Business process management: A review and evaluation. Business Process Management Journal, 4(3), 214–225. Letza, S. R. (1996). The design and implementation of the balanced business scorecard: An analysis of three companies in practice. Business Process Reengineering and Management Journal, 15(4), 245–266. Martinsons, M., Davison, R., & Dennis, T. (1999). The balanced scorecard: A foundation for the strategic management of information systems. Decision Support Systems, 25, 71–88. Mon, D. L., Cheng, C. H., & Lin, J. C. (1994). Evaluating weapon system using fuzzy analytic hierarchy process based on entropy weight. Fuzzy Sets and Systems, 62, 127–134. Park, D. J. (1995). A study on the determination of process selection factors in business reengineering. The Journal of Southern Seoul University, 1, 233–248. Saaty, T. L. (1980). The analytic hierarchy process. McGraw-Hill. Sinur, J. (2003). The business process management scenario. Gartner. Smith, H., & Finger, P. (2003). Business process management: The third wave. MeghanKiffer Press. Smith, R. F. (2007). Business process management and the balanced scorecard. John & Sons. Verner, L. (2004). BPM: The promise and the challenge. QUEUE. Available from http:// www.acmqueue.com. Weck, M., Klocke, F., Schell, H., & Ruenauver, E. (1997). Evaluating alternative production cycles using the extended fuzzy AHP method. European Journal of Operational Research, 100, 351–365. Zadeh, L. A. (1975). Fuzzy logic and approximate reasoning. Synthese, 30, 407–428.