1 Introduction

Improving service quality is a top priority for firms that aim to differentiate their services in today’s highly competitive business environment. Researchers and practitioners attribute this emerging trend to two salient driving forces (Nakhai and Neves 2009). First, the service sector has become the dominant part of the economy, particularly in developed or industrialized nations. Second, products are being offered more and more bundled with services in response to customer needs (Geum et al. 2011). To cope with these changes, an increasing number of businesses are adopting the quality improvement programs originated in manufacturing, such as total quality management (TQM), Six Sigma, reengineering, benchmarking, quality function deployment, etc., to enhance service quality (Hoerl and Snee 2002). Intrinsically, the underlying principles of these quality improvement methodologies can be applied in a service-oriented business environment (Feigenbaum 1983; Ishikawa 1985; Deming 1986; Hensley and Dobie 2005). In particular, Six Sigma, which makes use of a series of well-defined steps (define, measure, analyze, improve, and control), has received a growing attention and interest from service firms owing to its customer-centric philosophy that produces satisfactory results for many world-class companies (Taghaboni-Dutta and Moreland 2004).

In fact, Six Sigma manifests itself in many ways and, as such, is a multifaceted conceptualization (Tjahjono et al. 2010). For example, Six Sigma may refer to a set of statistical tools for process improvement (Goh and Xie 2004), or an analysis methodology that utilizes scientific methods (Kumar et al. 2007). Some researchers further define Six Sigma as an operational philosophy of management that is beneficial to customers, shareholders, employees, etc. (Chakrabarty and Tan 2007), and a business culture, an organized structure that uses process improvement specialists aiming to achieve strategic objectives (Schroeder et al. 2008). This project-driven approach which utilizes the problem solver’s expertise and the support of collected data to form a structured way for improving the quality of targeted processes or products has been adopted by numerous leading companies in the world since its introduction in the late 1970s (Goh 2002). In the 1990s, it migrated to analysis of transactions within the manufacturing sector. Since then, it saw a shift to applying those concepts to the transactional activities in non-manufacturing industries.

The success of Six Sigma implementation, however, is contingent on several factors such as top management support, clear performance metrics, organizational understanding of work processes, etc. (Chakrabarty and Tan 2007). Among these critical success factors, selecting the right Six Sigma projects is essential to the early success and long-term acceptance within an organization (Adams et al. 2003). A body of research probes Six Sigma project selection and generates fruitful insights (Antony 2006; Su and Chou 2008; Yang and Hsieh 2009). Particularly, Antony (2006) and Su and Chou (2008) maintain that the project selection process should be systematic and respond to three important voices: the voice of the process, the voice of the customer, and the voice of the strategic business goals. Following this line of logic, Yang and Hsieh (2009) propose a project selection mechanism based on Taiwan national quality award criteria and Delphi fuzzy multiple criteria decision-making method. In contrast, many firms merely consider voices of the process and customer in project selection and regard Six Sigma as “operational”, ignoring its link to corporate strategies (Kwak and Anbari 2006). While Six Sigma research suggests a number of project selection approaches, these approaches generally provide only general guidelines for project selection (Antony 2006), or primarily for the manufacturing context (Su and Chou 2008; Yang and Hsieh 2009). Research on the service Six Sigma project selection using a systematic approach remains deficient (Heckl et al. 2010). In view of the rapid growth of service industries, this deficiency in service Six Sigma project selection research apparently warrants research attention.

This study thus aims to fill the research gap by developing a systematic approach to select Six Sigma projects in the service industries. This research work is contributive as it integrates various project selection scenarios and tools, and thereby establishes a framework to facilitate the project selection process. This study is organized as follows. First, the literature pertaining to current implementation of Six Sigma in service industries is reviewed. Second, the study proposes a framework to select service Six Sigma projects. Third, the study uses two case studies in banking and health care industries to demonstrate the implementation of the proposed framework. Finally, the study draws conclusions and suggests future research topics.

2 Six Sigma in services

Occasionally referred to as the “transactional Six Sigma,” service Six Sigma offers firms a disciplined approach to improve service efficiency (i.e., saving time and cost) and effectiveness (i.e., meeting the desirable attributes of a service) in the business processes (Antony et al. 2007). For example, the waiting time to be admitted into an emergency room is a measure of service efficiency, whereas cleanliness is a measure of service effectiveness in the hospital environment. Namely, for firms implementing service Six Sigma, the emphasis in process improvement lies on the above timeliness characteristics and service non-conformity characteristics (Antony 2004a). Service-oriented firms adopting Six Sigma gain benefits essentially from creating more consistent processes for service delivery that lead to satisfied customers and lower cost (Bisgaard and Freiesleben 2004). As is the case in manufacturing, Six Sigma proves to be a useful tool to promote quality in a broad range of services (Chakrabarty and Tan 2007; Nakhai and Neves 2009). Specifically, researchers acknowledge Six Sigma’s benefits in material and facility management (Holtz and Campbell 2004), education (Jenicke et al. 2008; Utecht and Jenicke 2009), innovation (Byrne et al. 2007; Cho et al. 2011), payroll and human resource (Hayen 2008), software development (Grant and Mergen 2009), etc.

Particularly, Six Sigma applications in health care and financial services have attracted much attention from both academics and practitioners (Antony 2004a; Heckl et al. 2010). The first health care organization to fully implement Six Sigma into its culture was Kentucky’s Commonwealth Health Corp., USA in partnership with General Electric. They jointly performed a Six Sigma project to improve radiology throughput and reduce the cost per radiology procedure, which generated a financial return of over $1.2 million (Thomerson 2001). Later, a similar Six Sigma project was accomplished in the film library of the radiology department of the University of Texas M.D. Anderson Cancer Center (Benedetto 2003). Overall, Six Sigma in health care has been adopted primarily to reduce medical errors and improve process efficiency (Johnstone et al. 2003; Drenckpohl et al. 2007; Gras and Philippe 2007; Printezis and Gopalakrishnan 2007; Gowen et al. 2008; Corn 2009).

By analogy, the focus of Six Sigma in financial services can be anything ranging from shortening transaction times at a particular branch to addressing a particular customer service issue at the call center. For example, Citibank lessened its service failure rate by more than 10% in its banking operations in 3 years, resulting in reduced waste, errors, and customer response time (Rucker 2000). Likewise, Fidelity Investments launched Six Sigma in 2002 as part of the initiative to move process analysis efforts to lean/Six Sigma (Nourse and Hays 2004); the goal was to improve customer satisfaction by “reducing variation caused by defects and waste or non-value added activities.” Continuing the above theme, Six Sigma research produces fruitful discussions on a range of issues in financial services (e.g., Jones 2004; Jiantong and Wenchi 2007; Uprety 2009).

2.1 Challenges for Six Sigma in services

Six Sigma research reveals a number of challenges in applying this methodology to service operations (Biolos 2002; Hensley and Dobie 2005; Antony 2006; Chakrabarty and Tan 2007; Nakhai and Neves 2009). In contrast to most manufacturing settings, these challenges in service Six Sigma mainly arise from the general lack of tangibility and measurability in a service process (Chakrabarty and Tan 2007), more human involvement (Benedetto 2003), difficulty in data collection (Nakhai and Neves 2009), and the lack of project selection paradigm (Heckl et al. 2010). For example, clearly defining the how and what of service failure in a service process can be arduous (Biolos 2002; Does et al. 2002; Smith 2003). Notably, a mere agreement on what constitutes a defect is potentially debatable in that it is problematic to reconcile both service provider’s and stakeholder’s perspectives and clarify who the customers are (e.g., clinicians versus patients in health care).

Other challenges for service Six Sigma include the ability to identify well-defined deliverables of the project, to pinpoint the beginning and ending of a service process, and to measure the performance of the process (Lanser 2000). Particularly, the measurement of process performance receives inadequate attention from many service-oriented businesses (Antony 2004b; Hensley and Dobie 2005). It is quite common to have some sort of measuring mechanism in place in a manufacturing context (e.g., number of defects per million parts produced), which provides an indicator of process performance and product quality. This system and practice, however, do not always translate into the service industries (Antony 2006). While the steps in the measurement phase are explicitly defined in manufacturing settings, whereas in services, measuring the process to satisfy customers’ needs is often a more general problem of data collection, quality, and integrity (Does et al. 2002; Hensley and Dobie 2005; Antony 2006; Heckl et al. 2010).

Service firms also find it challenging in establishing a systematic process to identify the sources of errors and drive them down. For example, firms usually develop process maps before initiating Six Sigma projects in manufacturing. Nonetheless, the use of flowcharts and process maps remains rare in many service processes (Antony et al. 2007). Additionally, service companies are often dependent on people processes. In this regard, service processes are more subject to noise or uncontrollable factors compared to manufacturing processes (Does et al. 2002), explaining partially why Six Sigma came slowly to health care and initially was met with some skepticism (Benedetto 2003). Intrinsically, this hesitancy results from disparities between processes driven by humans versus automated or engineered processes.

In manufacturing, it is likely to eliminate most of human variability through automation, creating precise measurement of assignable causes of variation. Conversely, in services, especially in health care, the delivery of service such as patient care is largely a human process, and the causes of variability are often subtle and difficult to quantify. Patient variability too represents a key source of human variability; an acceptable medical care to one patient might be perceived as unsatisfactory by another patient. Additional sources of human behavioral characteristics engendering variability in services include friendliness, eagerness to help, honesty, etc., which are difficult to manage per se, and thus undermine the implementation of service Six Sigma (Antony 2004a). Furthermore, service-based companies may struggle with Six Sigma due to its intense data focus, difficulty in creating cultural changes for empowering Six Sigma leaders, and the low likelihood to capture the benefits of Six Sigma application immediately (Lanser 2000; Sehwail and DeYong 2003). Particularly, cost benefits generated from service Six Sigma projects may take time to realize, prompting managers to concede early (Sehwail and DeYong 2003). Above all, researchers consider project selection as a universal challenge for Six Sigma in services (Antony et al. 2007; Heckl et al. 2010).

2.2 Project selection for Six Sigma in services

Project selection has drawn notable attention in service Six Sigma due to its decisive role in the project success (Adams et al. 2003). Antony (2004a, 2006) indicates that project selection is one of the most critical success factors for the effective deployment of a Six Sigma program in service industries. Project selection refers to the process of choosing the best among alternative proposals on the basis of cost-benefit analysis so that the objectives of the organization will be achieved. In practice, the selection of Six Sigma projects in many service-oriented organizations is still based on pure subjective judgment (Raisinghani 2005; Antony 2006). Management has difficulty making project go/no go decisions and projects are generally initiated because management thinks they will make a contribution to quality (Antony 2006).

Research exposes several criteria for service Six Sigma project selection. Fundamentally, firms should choose projects in accordance with the firm’s goals and objectives (Gijo and Rao 2005; Antony et al. 2007) that tackle their business and customer problems (Does et al. 2002). As such, good Six Sigma projects possess characteristics that connect to business priorities, major importance to the organization, reasonable scope, etc. (Snee 2002; Antony 2006). Specifically, firms may rank potential projects based on five strategic imperatives: human resources, information technology, finance, quality and market growth, and expansion (Beaver 2004). Firms may also identify projects according to criteria such as financial return, customer satisfaction, resource required, risks, and alignment of strategic business goals and objectives (Antony 2004a; Chakrabarty and Tan 2007).

Chakrabarty and Tan (2008) suggest the following criteria for selecting service Six Sigma projects: measurable financial benefits, impact on business, linking to company’s business strategy, high probability of success, and far reaching impact. Most importantly, potential projects must be capable of reducing the defects or rework and have a high likelihood of being successfully completed on a tight time schedule (e.g., within 5 months) (Antony 2006). Furthermore, each project delivers a bottom-line financial result of $150,000 on an annual basis. In addition to these common criteria across service contexts, researchers propose several context-specific guidelines. For example, cost and frequency of problems represent the two major project selection criteria in banking services (Krupar 2003), whereas data availablility, clearly defined goals, milestones, timelines, and budgets serve as the key project selection guidelines in general financial services (Heckl et al. 2010). Likewise, service level, service cost, customer satisfaction, and clinical excellence constitute the four project selection criteria in health care (Sehwail and DeYong 2003). Table 1 summarizes the frequently endorsed service Six Sigma project selection criteria.

Table 1 Summary of service Six Sigma project selection criteria
Full size table

Quite a few tools are available for service Six Sigma project selection. For example, service-based companies may adopt the process map to identify those potential projects (Wyper and Harrison 2000; Krupar 2003; Antony 2004a; Raisinghani 2005; Antony et al. 2007; Nakhai and Neves 2009; Heckl et al. 2010). Likewise, the Failure Mode and Effects Analysis (FMEA) can be a useful problem-identification and ranking technique to develop the project list (Beaver 2004; Nakhai and Neves 2009). Furthermore, firms may follow the critical to quality (CTQ) breakdown approach to develop the candidate projects for service operations (Wyper and Harrison 2000; Sehwail and DeYong 2003; Heckl et al. 2010). Other commonly used tools and techniques include affinity diagrams, root cause analysis (RCA), and Pareto analysis (Antony 2004a; Antony et al. 2007). Service firms can also use these tools for other purposes in the subsequent phases (e.g., the Analyze phase) of a service Six Sigma project. On the other hand, service firms may employ empirical techniques (e.g., customer survey) via the designed questionnaire (Wyper and Harrison 2000; Raisinghani 2005), or the KANO’s model to facilitate the project selection process (Antony 2004a). Particularly, the KANO’s model helps understand the intended customer base through categorized customer expectations (i.e., basic, competitive and delight). Understanding each of these categories is essential to service Six Sigma project selection since it signifies that firms can better address customer needs, i.e., create specific value for different customers (Antony 2006). Table 2 summarizes the frequently adopted tools in service Six Sigma project selection.

Table 2 Summary of service Six Sigma project selection tools
Full size table

A review of the related literature suggests the need for an organized approach incorporating the appropriate project selection criteria and tools, and extending them when necessary, for service Six Sigma project selection. Thus, the study proposes a step-by-step framework that guides the Six Sigma deployment team to select the right projects. The proposed framework is discussed in the following section.

3 Proposed framework

Overall, the proposed framework consists of four phases, namely, initial project identification, project value assessment, project complexity assessment, and project prioritization, which are described as follows:

3.1 Phase 1: initial project identification

A firm’s strategy is eventually carried out through projects. Hence, each service Six Sigma project should support the organization’s initiatives to realize the corporate vision and is expected to have a clear link to the organization’s business strategy (Antony et al. 2007). Specifically, initial identification of the projects must involve the following pivotal activities (Larson and Gray 2011): (1) review the organizational mission, (2) set long-range goals and objectives, (3) analyze and formulate strategies to reach objectives, and (4) establish portfolio of project choices. Nonetheless, the final project list is contingent on further assessment.

3.2 Phase 2: project value assessment

Firms adopting service Six Sigma typically aim to improve business, customer, and employee value via the project-oriented approach (Antony 2006). As such, firms enhance their service processes and provide consistent and reliable services to gain profits and build employee pride, satisfaction, etc. In this regard, firms should evaluate each candidate project’s value based on three criteria: (1) financial return, (2) cost, and (3) the impact on employee behavior. First, as aforementioned, service Six Sigma researchers largely agree on the importance of financial impact, especially in hard savings and soft savings, when selecting projects (Antony 2004a; Beaver 2004; Chakrabarty and Tan 2007). Second, researchers likewise stress the criticality of cost estimation in service Six Sigma project selection (Krupar 2003). While a number of approaches exist for estimating the financial return and cost associated with a particular project, the estimation is in itself a complex process (Larson and Gray 2011). The study thus suggests firms to estimate the financial return and cost based on the consensus method, a frequently used top-down approach for estimating project return and cost, for the following reasons (Larson and Gray 2011). It is ambitious to accurately predict the cost of any project even with new techniques such as activity-based costing. Furthermore, firms often use the top-down approach in the “need” phase of a project to get an initial financial estimate for the project considering that much of the information needed to derive accurate estimates is not available in the initial phase of the project.

Third, as noted above, there exists a high degree of customer contact in services. An employee’s attitude and behavior toward a customer may determine the customer’s wish to continue service with the firm. That is, employee behavior represents a pivotal consideration in Six Sigma deployment (Eckes 2003; Yang and Chen 2003; Gels 2005; Chakrabarty and Tan 2007), which leads to positive morale, satisfaction, etc., thereby creating a cultural change for Six Sigma (Operation Management Roundtable 2002; Sehwail and DeYong 2003). Hence, it is essential to consider a project’s potency to shape employee behavior when evaluating a service Six Sigma project’s value. Nonetheless, it remains problematic to measure a project’s impact on employee behavior in practice owing to its intangible nature. This study thus evaluates the impact on employee behavior via the Analytic Hierarchy Process (AHP), a tool for group decision-making based on pairwise comparisons (Saaty and Peniwati 2008) that is sometimes adopted in manufacturing-oriented Six Sigma (Su and Chou 2008). In AHP, every group member makes judgment on each project’s potential to enhance employee behavior, and compares these projects in pairs. Specifically, two questions are asked in each pairwise comparison: (1) Which project is more important with respect to the criterion (i.e., impact to promote employee behavior)? and (2) How strongly based on the 1–9 scale (i.e., 1: equally important; 9: extremely much more important)? AHP converts these evaluations to numerical values that can be processed and compared over the entire pool of the projects. A numerical weight or priority is derived for each project, allowing these projects to be compared to one another in a rational and consistent way. Finally, the study formulates the project value as follows:

$$ {\text{Value}} = \frac{\text{Return}}{\text{Cost}} \times {\text{Employee\_behavior}} $$

3.3 Phase 3: project complexity assessment

In addition to value, there is yet another critical dimension, i.e., complexity, in assessing the project. The study evaluates the complexity of a service Six Sigma project according to three criteria: (1) project scope, (2) data availability, and (3) risk. First, project scope refers to the mission of the project—a service for the organization’s client/customer, including project objective, deliverables, etc. (Heckl et al. 2010; Larson and Gray 2011). Defining the project scope sets the stage for developing an effective service Six Sigma project plan, whereas poorly defined scope typifies a frequently mentioned barrier to project success (Snee 2002; Larson and Gray 2011). Nonetheless, leaders of well-managed, large corporations often neglect the imperative need to define the scope for a service Six Sigma project (Antony 2004a).

Second, data availability exemplifies another critical consideration when selecting service Six Sigma projects (Yang and Chen 2003; Hensley and Dobie 2005; Heckl et al. 2010). Particularly, data on the financial impact, cost of poor quality (COPQ), service level and quality, and service outcomes are crucial to the success of a project. However, data are mostly subjective and cannot be measured directly in services. Furthermore, there exists a general problem of data quality in non-manufacturing Six Sigma projects (Does et al. 2002). Above all, representing data in the form the project team desires poses a potential challenge for service firms even if the data are available. Similar to how the impact on employee behavior is measured above, the study uses AHP to evaluate the relative level of project scope and data availability for each project. In particular, the project scope is measured based on the overall assessment of deliverables, whereas data availability is evaluated according to the required effort to collect the necessary data in project implementation.

Finally, the study evaluates the potential risk involved in a project. Researchers consistently address this selection criterion for a typical or non-Six Sigma project (Stewart and Mohamed 2002; Enea and Piazza 2004; Daniels and Noordhuis 2005; Kulak et al. 2005; Lefley 2006; Baird et al. 2008). The consideration of risk, however, receives insufficient attention in the discussion of Six Sigma projects (Antony 2004b; Su and Chou 2008). In light of the existence of prevailing uncontrollable factors in a service setting (Does et al. 2002; Hensley and Dobie 2005; Antony 2006), evaluating the risk prior to project implementation appears to be crucial. Thus, this study incorporates the risk factor into project complexity assessment. Specifically, the risk is assessed via FMEA in terms of risk score according to the following three dimensions, namely, impact, occurrence, and detection (Adachi and Lodolce 2005; Su and Chou 2008). Each dimension is rated based on a five-point Likert-type scale anchored by 1 (lowest level of impact, least likelihood to occur, or highest possibility to detect prior to the occurrence) and 5 (highest level of impact, highest likelihood to occur, or lowest possibility to detect in advance). The weighting of the risk is then based on the overall risk score. For example, a risk yielding a minimum impact with a very low likelihood of occurrence, and an easy detection might score a 1 (1 × 1 × 1 = 1). Conversely, a high-impact risk with a high probability and impossible to detect in advance would score 125 (5 × 5 × 5 = 125). This broad range of numerical scores allows for easy stratification of risk based on overall significance for these projects. Accordingly, the study defines and formulates the project complexity as follows:

$$ {\text{Complexity}} = \frac{\text{Scope}}{{{\text{Data\_availability}}}} \times {\text{Risk}} $$

The “Appendix” presents the aforementioned measurement items in detail.

3.4 Phase 4: project prioritization

In the final phase, project prioritization is determined. The priority system can be managed by the Six Sigma project office or the quality initiatives management group. The study modifies the matrix originally developed by Matheson and Matheson (1998) to prioritize the projects (see Fig. 1). The vertical axis represents a project’s complexity, whereas the horizontal axis corresponds to a project’s potential commercial and organizational value. The grid has four quadrants, each with different service Six Sigma project dimensions.

Fig. 1
figure 1

Project prioritization matrix

Full size image

Specifically, any project falling into quadrant I is considered non-viable or No–Go (NG) as it renders low value while involving high level of complexity. Hence, it appears inefficient to perform this type of projects. The projects in quadrant II, though bringing forth relatively low value as well, can be regarded as just-do-it or low-hanging-fruit projects since they are perceived as less complex. Exemplary service Six Sigma projects should fall under quadrants III and IV. Projects in quadrant III are prospective candidates for green belt (GB) projects in that they yield high value and require less effort. In other words, they are eligible for a large-scale implementation in the service organization. In contrast, the projects assigned to quadrant IV can be problematic in making go/no go decisions. On the one hand, they are regarded as highly valuable and should be reckoned as black belt (BB) projects. On the other hand, the extremely complicated nature may impede the implementation of these projects. Thus, the project team may create a threshold in project complexity to determine whether these projects should be categorized as BB or NG projects. Notably, NG projects in quadrant IV, though perceived as disqualified to be the service Six Sigma projects, bear deliberation on the possibility of being other types of projects. Indeed, these arduous projects may be potentially contributive to the organization.

4 Case studies

The study presents two case studies in this section, which focus on two prominent service industries: the banking industry and the health care industry. Specifically, the study selected a multinational bank and a leading hospital in Taiwan for the demonstration of the proposed framework.

4.1 Banking services

4.1.1 The case bank

The case bank is a leading securities investment corporation located in Taipei, Taiwan (hereafter referred to as “BNK”). BNK is a company with more than 10,000 employees worldwide. At the time of this research, BNK had 215 service branches worldwide; among them, 142 were domestic. BNK had annual revenue of more than US$ 6 billion. With an outstanding balance of US$ 38 billion in deposits, BNK had assets of US$ 50 billion, surpassing all other private banks in Taiwan. BNK has been applying Six Sigma in service quality improvement since 2005 and is regarded as innovative in its application. Being one of the leading banks in Taiwan, BNK and its Six Sigma project deployment department anticipate to convert the bank’s DNA via relentless efforts in performing Six Sigma projects as its benchmark company GE Capital does. Nonetheless, BNK has not established a standard or paradigm in business strategy deployment and project breakdown framework even though hundreds of Six Sigma projects were executed these years. Furthermore, there was no commonly followed criterion in place to assess the project value and complexity during the project selection process. As a result, BNK performed the proposed approach noted above aiming to overcome these problems.

4.1.2 Implementation

4.1.2.1 Phase 1: initial project identification

After reviewing BNK’s organizational mission and vision statements the following strategic goals were identified: (1) to provide well-rounded products and premium services to clients in the Greater China market beyond Taiwan and Hong Kong, (2) to establish and put into practice a corporate philosophy that truly caters to what customers need; uphold BNK’s leading status in such critical areas as wealth management, syndicated loans, foreign exchange (i.e., forex), and forex derivatives, and (3) to become Taiwan’s foremost supplier of payment services and small personal unsecured loans. The above goals were then broken down into potential Six Sigma candidate projects for selection. Table 3 presents the deployment of BNK’s strategic goals. For example, the deployment aligns projects B1–B3 with the first strategic initiative to improve customer services. Likewise, projects B4 and B5 (B6 and B7) correspond to the second (third) strategic goal.

Table 3 BNK’s strategic goals deployment
Full size table
4.1.2.2 Phase 2: project value assessment

Each project’s value was appraised in this phase based on financial return, cost, and the impact on employee behavior. As suggested above, the financial return and cost were estimated based on the consensus method. The Six Sigma deployment office (including two external experts) convened the finance department and a master BB (MBB) to assess the financial return and cost for each project in a joint meeting. Additionally, each project’s impact on employee behavior was ranked and rated via pairwise comparisons by the team. The priority of each project was thus calculated using the AHP. Specifically, the Expert Choice software package, a multiattribute decision support software tool based on the AHP methodology, was chosen to perform the weighting of each project. The software uses the AHP methodology to model a decision problem and evaluates the relative desirability of alternatives. Finally, each project’s value was calculated. The resulting estimates for financial return, cost, employee behavior, and project value are shown in Table 4.

Table 4 BNK’s project value assessment
Full size table
4.1.2.3 Phase 3: project complexity assessment

In this phase, each project’s complexity was evaluated in terms of scope, data availability, and potential risk involved. The project scope includes deliverables and limits and exclusions of the project per se. Due to the fact that a bank’s operation must comply with specific governmental regulations in Taiwan, certain external factors/noises were introduced to the projects and thus made them less controllable. For example, to avoid the rampant fraudulent incidents, Taiwanese government started in 2005 to require all the banks to validate at least two personal IDs before an individual can open a new account. This constraint created numerous customer complaints for BNK, especially during the project of reducing complaints from new account openers. Likewise, the often cross-functional processes involved in a transactional project also extended the project scope. Regarding data availability, several information systems exist within BNK that were installed at different times—a typical problem for a bank. Specifically, the account management system and the customer information system may contain the same customer’s information in pieces but for different purposes. BNK noticed that it was ambitious to retrieve the data in the desirable format and combine the necessary information altogether, e.g., in the project of eliminating the possibility of erroneous data entry. Apparently, the above concerns drew substantial attention when the project team ranked the candidate projects via the AHP.

Next, to assess the risk involved in each project, the corresponding overall risk score for each project was calculated based on its impact, occurrence, and detection. Unlike the projects in a manufacturing context where historical data are generally available for reference in determining the individual scores, BNK encountered challenges in estimating these scores as expected in a typical service organization. For example, in the project of reducing response delays, it was arduous to determine the impact in that the consequence may range from a mere oral complaint to the loss of the customer. After each score in impact, occurrence, and detection was estimated, the overall risk score for each project was calculated accordingly using FMEA. The project team was then able to determine the complexity for each project. Table 5 demonstrates these estimates for each project.

Table 5 BNK’s project complexity assessment
Full size table
4.1.2.4 Phase 4: project prioritization

In this final phase, all the projects were prioritized in the project prioritization matrix with the two dimensions of project value and complexity (see Fig. 2). Each diamond on the matrix corresponds to a particular project. In this case, the value of 0.5 (100) in project value (complexity) was adopted to separate the four quadrants. Additionally, a threshold value of project complexity equal to 150 served as the cut-off line for BB and NG projects. The cut-off line of 150 is a result of team agreement. The Six Sigma project deployment team, which comprises two external experts and three internal executives, reached the consensus mainly based on an overall assessment of what level of complexity could go beyond the capacity of the implementation team to finish the project in time. As can be seen in Fig. 2, the seven projects being evaluated were finally categorized into 1 BB, 3 GB, 1 just-do-it, and 2 NG projects.

Fig. 2
figure 2

BNK’s project prioritization matrix

Full size image

4.2 Health care services

4.2.1 The case hospital

The case hospital (hereafter referred to as “HC”) is a major medical center located in Kaohsiung, Taiwan with approximately 340 full-time physicians and over 1,600 patient beds. As an initiative to become a high-quality medical center, HC established in 2006 the Medical Quality Control Office (MQCO) to oversee and promote all quality-related activities such as training physicians, monitoring medical performance, launching campaigns for quality patient care, etc. An important mission for MQCO was to employ the Six Sigma philosophy in patient care quality improvement. Fulfillment of this mission, however, required HC to support its employees with new system of care. Particularly, the pressing need was to create a systematic approach to select Six Sigma projects for implementation. Thus, HC adopted the proposed framework in this study to fulfill the above need.

4.2.2 Implementation

4.2.2.1 Phase 1: initial project identification

A cross-functional Six Sigma core team in MQCO led by HC’s Vice President for Six Sigma was established to identify and supervise the projects. After reviewing HC’s organizational mission statements the following strategic goals were determined: (1) to avoid providing ineffective inpatient and outpatient services, (2) to provide appropriate and speedy support for medical decision-making, and (3) to fulfill social responsibility/governmental policy compliance. These goals were then broken down into potential Six Sigma candidate projects for selection. As Table 6 indicates, the deployment of HC’s strategic goals generated six projects. For example, the deployment aligned projects H1 and H2 with the first strategic initiative to improve patient services. Similarly, the second (third) strategic goal lead to projects H3 and H4 (H5 and H6).

Table 6 HC’s strategic goals deployment
Full size table
4.2.2.2 Phase 2: project value assessment

Each project’s value was appraised subsequently based on financial return, cost, and the impact on employee behavior. Calculating the financial return and cost was manageable for particular projects such as the project to increase nurse retention in intensive care units (ICUs), whereas it was baffling for certain projects (e.g., the project to improve patient satisfaction). Specifically, it was concluded that the project to improve patient satisfaction should potentially produce the most profit, even though quite a portion of the financial return resulted from the soft savings. Conversely, improving inpatient exclusive breastfeeding rate was considered least profitable. As noted above, each project’s impact on the employee behavior was ranked via the pairwise comparisons using AHP. The comparison results suggested that increasing nurse retention in ICUs was mostly related to the promotion of positive employee behavior. Hence, this project yielded the highest weight. Accordingly, each project’s value was assessed. Table 7 depicts the detailed estimates for financial return, cost, employee behavior, and project value.

Table 7 HC’s project value assessment
Full size table
4.2.2.3 Phase 3: project complexity assessment

The implementation of mandatory National Health Insurance (NHI) program in Taiwan played a crucial role in the assessment of project complexity. Launched in 1995, NHI represented the realization of a primary social and health care policy of the Taiwanese government. The ultimate goal of the program is the universal coverage for all Taiwan citizens. It is estimated that NHI currently covers more than 99% of the population. In 2009, national health care expenditures in Taiwan totaled US$ 13.73 billion, representing 3.51% of Taiwan’s GDP. Recognizing the rapid growth of health insurance expenditures and a slow increase in premium collection, the NHI Bureau has taken initiatives to maintain the balance of income and expenditures to reduce deficits and to be financially self-sustained. These campaigns include, e.g., increasing cost-sharing from the beneficiaries to prevent the patients from abusing medical services, expanding the case-payment system, and providing patients alternative insurance plans.

Inevitably, being a participating hospital in the NHI program, HC was impeded to some extent when performing Six Sigma projects. For example, only a limited number of HC’s wards were available for NHI subscribers without co-payment. This resulted in an over usage for those wards while an under usage for other types of wards that need extra co-payment made by the patients. This external factor apparently drove patients to wait longer for available wards. After the cross-functional Six Sigma core team in MQCO examined each project’s scope, data availability, and potential risk involved, it was concluded that improving patient satisfaction involved the largest scope. Unexpectedly, the data needed for reducing the delay of joint consultation in emergency rooms was considered least available as there was no mechanism in place to effectively monitor the process. Even though each physician was required to log into the system when he/she arrived at the emergency room, the time mark could be fraudulently altered. Thus, the data integrity was dubitable.

Next the project team estimated the scores of impact, occurrence, and detection for each project to determine the overall perceived risk. It was noteworthy that the project to reduce start time delays in operating rooms engendered the highest risk. The project team attributed the high risk to the involvement of several key patient care participants, namely, the anesthesiologist, nurse, patient, and doctor in the operating room. In practice, anesthesiologists and nurses need to cooperate seamlessly in setting up the patient for surgery in advance to avoid the delay. Table 8 shows the detailed calculation for project complexity for each project.

Table 8 HC’s project complexity assessment
Full size table
4.2.2.4 Phase 4: project prioritization

In this final phase, each project was positioned on the project prioritization matrix (see Fig. 3). In this case, the value of 0.5 (100) in project value (complexity) served to separate the four quadrants, whereas the threshold value of 150 in project complexity was used to differentiate between BB and NG projects. The six projects being evaluated turned out to be 3 BB, 2 GB, 1 just-do-it, and no NG projects.

Fig. 3
figure 3

HC’s project prioritization matrix

Full size image

5 Conclusions

This study proposes a framework for the selection of service Six Sigma projects and demonstrates its application with two case studies in banking and health care services. The project selection process consists of four phases, namely, Phase 1: initial project identification, Phase 2: project value assessment, Phase 3: project complexity assessment, and Phase 4: project prioritization. Specifically, the study highlights projects’ strategic link to business goals in the phase of initial project identification. The study also develops two pivotal criteria to measure project value and complexity. These measures are evaluated in Phase 3 and Phase 4, respectively. In particular, the study evaluates project value based on financial return, cost, and the impact on employee behavior. Likewise, the study appraises project complexity according to scope, data availability, and the potential risk involved. In the final phase, the study establishes a project prioritization matrix to facilitate the categorization of different types of projects including BB, GB, just-do-it, and NG projects. Furthermore, the study employs tools such as AHP and FMEA to rank and rate the potential projects.

Although this study contributes to the extant literature by developing a viable and systematic approach to facilitate service Six Sigma project selection, several limitations are of note drawing directions for future research. Particularly, differences remain across service contexts in implementation. For example, the measurement of risk, especially on the occurrence and detection, is more difficult in health care than in banking services due to greater human involvement (Jenkins 2006). Furthermore, data availability varies significantly across service settings, depending on the level of “industrialization” of the service operations (e.g., banking versus education). Likewise, financial return is generally limited in non-profit organizations (e.g., government agencies and hospitals) versus literally unrestricted in for-profit organizations (e.g., financial services). These issues arising from context differences apparently affect the determination of threshold values in the project prioritization matrix. Finally, the enforcement of government laws and regulations plays a critical role in the cases demonstrated by this study, which merits further cross-national research on the generalizability of the results.