SERVICE CLASSIFICATION AND MANAGEMENT CHALLENGES Rohit Verma Kenneth K. Boyer DePaul University Chicago,IL Abstract Coincident with the increasing importance of services as a primary compo- nent ofthe economies in developed countries, a number oftheoretically derived service typologies have been developed, yet there has been virtually no empiri- cal validation ofthe proposed ideas. We conducted a survey of273 managers in four service industries (Fast Food, Auto Repair, Retail Sales, Legal Services) to test how management challenges differ across different sendee industries. We also empirically tested a widely accepted service typology developed by Schmenner (1986) within the context of management challenges. Discriminant analysis was utilized to test the degree to which companies can be classified into groupings similar to the Schmenner's service process matrix according to em- pirical data rather than anecdotal observations. Our findings indicate that the while the service process matrix can be partially validated using empirical data, the distinctions between various industries are much ''fuzzier'' in practice. Different service industries can be classified according to empirical data, but misclassifications do occur. In particular, misclassifications are most prevalent where two service industries share a common characteristic as described by the service process matrix. Introduction As the post-industrial economy evolves, the service sector continues to in~ crease in importance, both in terms of its contribution to the gross domestic product (GDP) of all advanced economies and in terms of the percentage of workforce employed in services. For example, it has been predicted that the service sector will account for more than 88% of the workforce in the United States by the year 2001 (Fitzsimmons & Fitzsimmons, 1994). Accordingly, the last decade has witnessed an increased emphasis on teaching and research of service management issues by business schools and professional organizations. In response to the increased importance of services, numerous articles on issues related to the effective management of service operations have appeared in hoth academic and practitioner based publications (for example, Chase & Hayes; 1991, Karmarker & Pitbladdo, 1995; Kellogg & Nie, 1995; Lovelock, 1992; Roth & Van Der Velde, 1991). Several of these articles present typologies 6 Journal of Business Strategies Vol. 17, No.1 of services and provide directions for improving quality, productivity and oper- ating efficiency, however relatively little has been done to empirically test the proposed ideas. This article presents an empirical assessment of the management challenges proposed by one of the widely accepted service typologies - the service pro- cess matrix (SPM) developed by Schmenner (1986). We gathered data relating to management challenges experienced by managers of four different types of service industries (Fast Food, Auto Repair, Retail Sales and Legal Services). These services were chosen because they differ in terms of various attributes of service delivery systems as suggested by the SPM. Based on empirical data collected from 273 managers, we show how management challenges differ across four types of services. In addition, we provide an empirical test of how well the four types of industries described by Schemnner (1986) can be classi- fied using empirical data. The remainder of the article is divided into four sections. First, we present a review of various service typologies; Next we describe the research methods used in the study; Third, we present the results of our analysis; and Finally, we present a discussion of the implications of the findings from this research. Service Typologies This section offers a review of various service classifications schemes that have been developed, as well as a discussion of their relative strengths and weaknesses (Table 1). This review is provided in order to illustrate that while a variety of insightful conceptual typologies have been developed, there is a need to provide empirical validation in order to identify whether these typologies accurately model reality, as well as identify any shortcomings. The diversity of the service sector makes it difficult to come up with useful generalizations concerning the management of service organizations. There- fore, a considerable of amount of research has been focused on developing service classification schemes. For example, Judd (1964) classified services according to three categories: rented goods, owned goods and non-goods ser- vices. Similarly, Rathmell (1974) categorized services according to: type of buyer, buyer moti ves, buying practices, type of seller, and degree of regulation. Even though these classifications show how some services are different from the others, they do not provide much useful insight into the design and management of service processes from an operational perspective. More recent classification schemes have explored the complex nature of service delivery systems with the goal of identifying differentiating characteris- tics, which affect quality and process improvement, as well as service design. For example, Shostack (1977) and Sasser, Olsen and Wyckoff (1978) developed the concept of "product-service package" based on the tangible versus intan- gible nature of services. Based on similar ideas, Levitt (1972, 1976) suggested Spring 2000 Verma & Boyer: Service Classification 7 that services are commonly thought of in humanistic terms and manufacturing is thought of in technocratic terms. Accordingly, manufacturing is seen as efficient and forward~looking, whereas services are viewed as primitive and inefficient. Table 1 Service Typologies Author(s) Judd, R.c. (1964) Rathmell, J.M. (1974) Shostack, G.L. (1977) Sasser, W.E. Jr., Olsen, R.P., & Wyckoff, D.D. (1978) Hill, T.P. (1977) Chase, R.B. (1978, 1981) Kotler, P. (1980) Lovelock, C.R (1980) Schmenner, R.W. (1986) Mersha, T. (1990) Chase, R.B. & Hayes (1991) Kellogg & Nie (1995) Categories/Groups • Rented Goods Services • Owned Goods Services • Non~goods Services • Types of seller • Types of buyer • Buying motives • Buying practice • Degree of regulation Proportion of physical goods and intangible services contained in each "product~service package." • Services affecting people vs. those affecting goods • Permanent vs. temporary effects of service • Reversibility vs. non reversibility of service • Physical vs. mental effects of service • Individual vs. collective services Degree of customer contact • People vs. equipment based • Extent of customers' presence • Public ~ Private vs. For~profit ~ Non~profit Five two-by-two classification matrices based on the following ideas: • nature of service act • relationship between service provider and customer • customization • demand and supply • service delivery Service Process Matrix based on two dimensions: • Customer contact and customization • Labor intensity Degree of customer contact. Definition of customer contact expanded to include active and passive contact Based on competitive stage Service Product - Service Process Matrix 8 Journal of Business Strategies Vol. 17, No.1 More recent researchers however believe in an integrated approach to service management. For example, Thomas (1978) argues that a large part of manufac- turing experience is irrelevant to the management of service operations because services are very different from manufacturing. Sullivan (1981) and Bowen and Cummings (1990) also advocate an integrated approach to service management and suggest that operations management researchers must include organiza- tional behavior and marketing constructs and techniques to address service operations problems adequately. In addition to the above-cited studies, a number of articles and books empha- size the multi-functional nature of services. Therefore, recent service typologies build on managerially useful service delivery system attributes. For example, Lovelock (1992) classifies services in five different two-by-two matrices and examines how the specific nature of services in a particular class affects opera- tions and marketing. Lovelock's (1992) framework addresses the following questions: (1) What is the nature of the service act? (2) What type of relationship does the service organization have with its customers? (3) How much room is there for customization and judgment on the part of the service provider? (4) What is the nature of demand and supply for the service? and (5) How is the service delivered? Lovelock (1992) proposed that his classification scheme addressing the above five questions can help managers develop a better under- standing of their business. Chase (1978, 1981) proposed that if there is less direct customer contact in the service system, then the service system is more likely to operate at its peak efficiency. Conversely, the system is less likely to operate at its peak potential with high direct customer contact. Mersha (1990) proposed a broadened defini- tion of customer contact and differentiated between active and passive contact. Based on these distinctions, Mersha (1990) extended the customer contact model and addressed several earlier concerns about this classification scheme. Building on the customer contact approach to services, Schmenner (1986) proposed the Service Process Matrix (SPM), based on three characteristics of service delivery systems. The SPM expands the customer-contact model and categorizes services on two dimensions: (1) labor intensity, and (2) Customer contact and service customization. Labor intensity is defined as the ratio of the labor cost incurred to the value of the plant and equipment. A high labor intensity business involves relatively small plant and equipment investment relative to a considerable amount of worker time, effort, and cost. The second dimension in the classification scheme combines two distinct concepts: customer interaction and customization. A service with a high level of interaction is one in which the customer can actively intervene in the service process. A service with high customization will work to satisfy an individual's particular preferences. The joint measure has a high value when a service exhibits both a high level of interaction and a high level of customization for the customers. Schmenner (1986) proposed a two-by-two service process matrix that classifies services as Spring 2000 Verma & Boyer: Service Classification 9 service factory, service shop, mass service, and professional service. Figure I shows the service process matrix and the corresponding management challenges associated with each type of service. A brief description of each of the four service types is provided below. Figure 1 Service Process Matrix (Schmenner, 1986) Capital Decisions Technological Advances Managing Demand Scheduling Service Delivery i i Customer contactlCustomization Fighting Cost Increases• Maintaining Quality Marketing t+-- I r---- Reacting to Customer Making Service "Warm" il Service I Service Intervention Attention to Physical Factory I Shop Managing Career Surroundings Advancements Managing Fairly Rigid ------+------ Managing Aat Hierarchy I+-- Mass I Professional r---- Hierarchy I Gaining Employee Service I Service Loyalty 1 1 Hiring Training Methods Development & Control Employee Welfare Scheduling Workforce Control of Far-flung Operations Startup of New Operations Managing Growth Service Factory. Services with both low customer contactlcustomization and a low degree of labor intensity are classified as Service Factories. Similar to line type processes in manufacturing, the facilities and equipment account for a large fraction of costs. Much of the transportation industry (airlines, trucking compa- nies), hotels and fast-food establishments can be classified as Service Factories because of low customer contact, customization and low labor intensity. Service Shop. Services with low labor intensity but high customer contact! customization are classified as Service Shops. Similar to a Job-Shop type of operation in manufacturing industry, Service Shops can provide various types of customized services for their customers. Hospitals, auto and other repair ser- vices are excellent examples of Service Shops. 10 Journal of Business Strategies Vol. 17, No.1 Mass Service. Mass Services have low customer contact!customization in combination with high labor intensity. Retail companies, wholesaling and schools are examples of Mass Service. Professional Service. These services have both high customer contact! customization and a high degree of labor intensity. Services provided by doc- tors, lawyers, accountant and architect all have a very high labor costs due to the large amount of education associated with these professions. In addition, these services tend to be highly customized according to the particular situation/need of each customer. The single common characteristic of all the cited studies is that they are primarily conceptual or theoretical in nature. Each of the cited studies presents a typology of either ideal service management or theoretically derived differences between ser- vices. As a result, each of the studies provides important insights into important characteristics of services. Yet, although many of these studies are based on anec- dotal or actual experience with a representative sample of companies. none have been empirically validated using a broad sample of quantitative data rather than subjective inferences. For example, even though the customer contact model was first proposed over fifteen years ago and is widely cited in business management literature, only recently has an empirically derived measure for customer contact been developed (Kellogg & Chase, 1995). The development of empirical models or taxonomies to quantitatively measure differentiating characteristics of services is an area sorely deprived of adequate study. Empirical studies are needed not only to validate largely conceptual models, but also to highlight areas where such models fall short of providing a perfect explanation of important relationships in a service environment. In particular, empirical studies provide an opportunity to iden- tify companies or industries, which are a little bit off the beaten path and don't fit the assumptions of the conceptual model well. These companies often are trendsetters that develop new techniques for providing a product that are then copied by competitors as an industry evolves. For example, Amazon.com (http://www.amazon.com) is changing the nature of book re- tailing through its use of the Internet to minimize customer contact while cutting costs and increasing choices. Competitors such as Barnes and Noble are also implementing online sales services in an effort to remain competi- tive. Empirical studies thus present an opportunity to develop or validate models that classify services. with much of the value in these models lying in their ability to identify outliers such as Amazon.com. This study presents an empirical assessment of one of the more widely ac- cepted service typologies. As described earlier, we use the service process matrix because it not only expands the customer contact model but also suggests how management challenges differ across different types of services (see Figure 1). We also seek to test the predictive accuracy of the service process matrix by means of discriminant analysis. Spring 2000 Verma & Boyer: Service Classification Research Methods 11 The data for this study was collected from service managers in four different types of industries: Fast Food, Automobile Repair, Retail Sales and Legal Ser- vices. A random sample of 70 firms from each of these industries was selected from the yellow pages phone directory of a large metropolitan area in the western United States. Each service firm that was contacted received a cover letter from the lead re- searcher, a forwarding letter from the chairperson of the management department of the sponsoring university and a two-page questionnaire. The questionnaire asked the managers to rate twenty-two management challenges (identified by Schmenner, 1986) on a five point Likert scale (1 = not a challenge; 3 = average challenge; 5 = big challenge). The items addressing management challenges included in the question- naire are shown in Table 3 and Figure 1. In addition, the instrument contained four demographic questions regarding the gender, age, work experience and education level of the respondent. The length of the questionnaire was intentionally kept to less than two pages so that the total time needed to respond to survey was less than 15 minutes. In order to increase the response rate, we offered to send a summary of the results to the managers. The respondents also had an opportunity to participate in a raffle and win a $200 cash prize. Each of these techniques has been shown to encourage participation in the data collection effort (Linsky, 1972). The survey instrument was hand delivered to the managers of each of the 280 service firms sampled. The managers were asked to complete the questionnaire immediately (if possible) and told that it would take less than 15 minutes of their time. Almost 75% of the managers completed it immediately. The rest of the managers agreed to complete the survey within a week. A return visit was made to these companies after 4-8 days with another copy of the survey instrument. After the completion of data collection, only six firms chose not to participate in the study. One questionnaire was not complete and was therefore excluded from further analysis, The resulting final sample size was 273, or an effective re- sponse rate of 97.5%. Results and Discussion Table 2 provides demographic information for the sample. Several differ- ences in education levels of the respondents can be readily identified. The Legal Service managers are most highly educated (91 % have college degrees). Man- agers in Retail Sales also are highly educated, with 40% holding a college degree. In contrast, less than 20% of the managers in the Fast Food and Auto Repair industries have college degrees. Another interesting demographic pattern was observed with respect to the gender of the respondents. There were no female respondents from the Auto Repair industry and only 7.25% of the managers in Legal Service are female. By 12 Journal of Business Strategies Vol. 17, No.1 comparison, 39% of the Retail Sales managers and 29% of the Fast Food managers were female. Table 2 also shows the average ages, average work experience and the sample sizes for the four industries. It is interesting to note that managers in Auto Repair and Legal Service tend to be older and more experienced than managers in either Fast Food or Retail Sales. Table 2 Sample Demographics Fast Auto Retail Legal Food Repair Sales Service (n =66) (n =69) (n =69) (n =69) Education (%) High School 48.5 39.4 26.1 0.0 Associate Degree/Some College 36.4 40.9 34.8 4.4 Four Year College Degree 13.6 19.7 29.0 4.4 Masters 1.5 0.0 10.1 31.9 Doctorate/JD 0.0 0.0 0.0 59.4 Total 100.0 100.00 100.0 100.0 Female Respondents (%) 28.8 0.0 39.1 73 Average Age (years) 28.2 38.7 31.3 44.5 Average Work Experience (years) 6.8 18.5 9.8 17.8 Scale Development The service process matrix proposes that management challenges differ across different types of services (see Figure 1). These management challenges have not been empirically assessed in prior research. Instead, they are based on anecdotal evidence derived from limited samples. We therefore seek to develop scales to assess management challenges, which are more readily generalizable in a variety of situations. Therefore, we examine the twenty-two management challenges shown in Appendix I with the objective of developing reliable and valid scales. Many of the individual management challenges appear to be components of a larger, underlying construct. For example, employee hiring, training and welfare could be considered as individual management challenges which comprise a larger construct (employee management). In order to identify the underlying factors within the 22 management challenges, as well as reduce the size of the data set to facilitate further analysis, we conduct an exploratory factor analysis. An exploratory factor analysis of the twenty-two management challenges was conducted and the results are summarized in Table 3. Seven eigenvalues ex- ceeded the generally accepted cutoff value of 1.0 (Kim & Mueller, 1978) and are therefore retained in the analysis. The seven factors explain a total of 60.6% of Spring 2000 Verma & Boyer: Service Classification 13 the variance in the data. In order to increase the interpretability, a Varimax rotation was performed on the principal components. Items were then assigned to the factor on which they had the highest loadings. Only items which had loadings of at least 0.40 on at least one factor were retained in the analysis. As a result of this cutoff, one item (QIO. Managing Growth) did not load on any factor and was therefore removed from further analysis. Table 3 shows the results of the Varimax factor rotation. Table 3 Factor Analysis of Service Management Challenges Factor I 2 3 4 5 6 7 Eigenvalue 5.20 1.93 1.52 1.42 1.13 1.08 1.05 Cumulative Percentage of Variance 23.6% 32.4% 39.3% 45.8% 50.9% 55.9% 60.6% EMPLOYEE a:::: 0.77 Q5. Employee Hiring 0.77 0.21 -0.04 0.03 -0.Q7 0.07 0.12 Q6. Employee Training 0.80 0.13 0.Q7 -0.03 0.19 0.02 0.08 Q7. Employee Welfare 0.70 0.04 0.15 0.10 0.09 0.23 -0.12 Q9. Gaining Employee Loyalty 0.59 0.29 0.Q2 0.08 0.17 0.15 -0.03 SERVICE QUALITY a:::: 0.70 Q I. Physical Surroundings 0.06 0.71 0.02 0.09 -0.04 0.05 0.04 Q 16. Quality of Service 0.33 0.63 0.18 0.14 0.15 O.ll -0.13 Q17. Making Service Warm 0.29 0.77 -0.Q2 0.16 0.14 0,11 -0.03 OPERATIONS a:::: 0.62 Q3. Capital Decisions 0.00 0.02 0.83 -0.03 0.00 -0.13 0.14 Q4. Developing Work and Control Methods 0.26 0.25 0.59 -0.06 0.20 -0.09 0.08 Q8. Fighting Cost Increases 0.26 -0.12 0.44 0.20 0.32 0.03 -0.Q7 QIL Technological Advances -0.15 -0.02 0.68 0.24 0.02 0.28 -0.14 MARKET Q12. Managing Demand 0.03 0.19 0.15 0.56 0.04 0.29 -0.06 Q 18. Marketing 0.02 0.11 -0.02 0.79 0.15 -0.19 0.10 SCHEDULING a=0.65 Q19. Reacting to Customer Intervention 0.13 0.34 0.02 0,45 0.49 0.15 0.12 Q20. Scheduling Service Delivery 0.02 0.00 0.22 0.05 0.81 0.12 -0.04 Q21. Scheduling Workforce 0.41 0.23 -0.05 0.12 0.61 0.11 0.09 ORGANIZATION a= 0.57 Q13. Managing Career Advancements 0.37 -0.15 0.09 0.38 -0.11 0.51 0.15 Q14. Managing Flat Hierarchy 0.18 0.09 -0.07 0.07 0.1l 0.74 0.07 Q15. Managing Rigid Hierarchy 0.10 0.32 0.04 -0.09 0.28 0.62 0.14 CONTROL Q2. Controlling Work for Far Locations -0.11 0.06 0.01 0.08 -0.10 0.19 0.80 Q22. Startup of New Operations 0.20 -0.13 0.08 0.26 0.17 0.01 0.70 NOTE: The highest loading for each item is indicated in bold. 14 Journal of Business Strategies Vol. 17, No. I We have labeled each of the seven factors shown in Table 3 according to the items which loaded on that factor. The EMPLOYEE factor is comprised of items relating to the hiring and training of employees, as wen as employee welfare and gaining employee loyalty. The items loading on the SERVICE QUALITY factor pertain to the challenges associated with physical surroundings, quality of ser- vice and making service warm. The third factor, OPERATIONS, includes ques- tions regarding capital decisions, work and control methods, cost increases and technological advances. The MARKET factor includes two items: managing demand and marketing. SCHEDULING is comprised of challenges relating to scheduling both the workforce and service delivery, as well as reacting to customer intervention. The ORGANIZATION factor includes items relating to managing career advancements, managing a flat hierarchy and managing rigid hierarchy. The final factor, CONTROL, is comprised of items regarding control- ling work in far locations and the startup of new operations. Scale Reliability Assessment Since the scales developed via factor analysis are new, we must take care to assess the inter-item reliability of the questions comprising each scale (Flynn et aL, 1990). Cronbach's coefficient alpha is used to assess inter-item reliability, with alpha values of 0.70 or higher considered to indicate acceptable reliability for established scales and 0.60 being acceptable for new scales (Churchill, 1979; Nunnally, 1978). As shown in Table 3, four of the seven scales possess alphas which exceed the 0.60 threshold and are therefore considered to exhibit accept- able reliability. Two of the remaining three scales have only two items, while the final scale (ORGANIZATION) does not meet the minimum threshold or reli- ability (Ct. = 0.57). Based on these results, we therefore discard these three scales and remove them from further analysis. Scores for the remaining scales are developed by taking the average of the items, which had their highest loading on that scale in the factor analysis shown in Table 3. Industry Comparison Dess, Ireland, and Hitt (1990) suggest that the variables of interest in a particular study should be examined in order to ascertain their sensitivity to industry condi- tions. Similarly, Ward et al. (1995) find between industry differences in manufactur- ing strategies with regard to a sample of Singapore-based manufacturing firms. We therefore test for industry level differences in the four remaining management challenges scales across the four service industries studied. Table 4 contains the means for the four management challenges scales for both the entire sample and for each of the four service industries. The four scales are ranked in order of decreasing importance based upon the average response for the entire sample. SERVICE QUALITY is considered to be the most impor- Spring 2000 Verma & Boyer: Service Classification 15 tant challenge (mean = 3.64), while SCHEDULING is considered to be the least important (mean = 3.07). Table 4 also provides a more intriguing analysis by breaking the sample down based on industry membership. If we examine the two industries with low customer contact (Fast Food and Retail Sales) separately from the two industries with high customer contactlcustomization (Auto Repair and Legal Services), several interesting differences appear. First, the low cus- tomer contact industries consider SERVICE QUALITY and EMPLOYEE chal- lenges to be most important. In contrast. Auto Repair shops consider OPERA- TIONS to be their primary challenge. Table 4 Group Means and One-Way ANOVA Results Industry 2 3 4 Entire Fast Auto Retail Legal Sample Food Repair Sales Service F= 2.58 Service Quality 3.64 3.74 3.48 3.83 3.52 P =0.05 (4) (4) (1.3) F= 6.09 Employee 3.56 3.78 3.49 3.74 3.23 p < 0.01 (2) (1,3,4) (2) (2) F= 17.36 Operations 3.25 2.98 3.79 3.06 3.18 p < 0.01 (4) (4) (2,3) F = 4.44 Scheduling 3.07 3.10 3.23 3.22 2.73 P < 0.01 NOTE:Numbers in parentheses indicate the group numbers from which this group was significantly different at the p < 0.05 level according to the Scheffe pairwise comparison procedure. F statistics and associated p-values are derived from one-way ANOVAs. The industry with the highest mean value for each management challenge scale is shown in bold. Next, a series of one-way ANOVAs was conducted to test ifthe four manage- ment challenge scales differ across the four types of services identified by the service process matrix. The overall F-test indicates that there are significant differences across the fouf industries for each of the four management challenge scales (used as the dependent variable). In addition, a Scheffe pairwise compari- son was conducted to test for differences between individual pairs of industries. No significant pairwise differences were found for SERVICE QUALITY. but the other three management challenge scales did exhibit dramatic differences. First, EMPLOYEE issues were much more important for the Fast Food and Retail Sales industries (means of3.78 and 3.74) than they were for Legal Service (mean of 3.23). In addition. the importance of OPERATIONS was dramatically higher for Auto Repair (mean of 3.79) than for any of the other three industries. 16 Journal of Business Strategies Vol. 17, No.1 Finally, Scheduling was of relatively little importance (mean of 2.73) for the Legal Services industries, but was of significantly greater importance for the Auto Repair and Retail Sales industries (means of 3.23 and 3.22, respectively). Manaeement Challenees and Service Process Matrix One of the primary objectives of this empirical study is to detennine whether the service process matrix typology developed by Schmenner (1986) can be validated and confinned using empirical data. We therefore employ the four management challenge scales developed in section 4.1 in a discriminant analysis to test whether these scales can be used to differentiate and classify the four service types posited by SPM. Discriminant analysis is used because it pennits examination of the differ- ences between two or more groups with respect to multiple discriminating variables simultaneously (Klecka, 1980). The management challenges scales are used as the discriminating variables to predict the actual industries of the respondents. The data are randomly divided into two samples, a calibration sample (n=180) and a validation sample (n=90). It is a common practice to use a calibration or training sample to derive the discriminant functions which best classify the data into groups. The coefficients derived from the calibration sample are then applied to the validation sample to test how well the discriminant functions actually classify a set of independent data into groups (Johnson & Wichern, 1988). If the discriminant functions derived from the calibration sample perfonn well in predicting the indus- try of companies in the independent validation sample, then the predictive ability is not merely an artifact of the set of companies contained in the calibration sample, thus demonstrating a base level of generalizability. Calibration Sample. The discriminant model is developed by applying a stepwise procedure in SPSS to the 180 companies included in the calibration sample. The independent variables are the four management challenge scales with high inter-item reliabil- ity which were retained in section 4.2 (EMPLOYEE, SERVICE QUALITY, OPERATIONS and SCHEDULING) and the dependent variable is the industry of the company (Fast Food, Auto Repair, Retail Sales or Legal Services). The stepwise procedure is analogous to the stepwise procedure for multiple regres- sion, i.e. variables which explain a significant amount of variance (p < 0.05) are selected to enter the model in order of decreasing effect and variables can be removed at later stages if they are no longer significant due to the addition of other variables at earlier stages. The stepwise procedure selected two of the four management challenge scales, EMPLOYEE and OPERATIONS, for inclusion in the model. Table SA shows the coefficients for each of the two discriminant functions, as well as Wilk's lambda and the industry means for each of the two functions for each of the four service industries. Spring 2000 Verma & Boyer: Service Classification Table 5 Discriminant Analysis Results 17 A. Standardized Canonical Discriminant Function Coefficients and Industry Means Scale Industry Means for Discriminant Functions Function 1 Function 2 EMPLOYEE -0.78 0.73 OPERATIONS 0.96 0.47 Wilks' Lambda 0.75 0.94 P < 0.01 p < 0.01 Fast Food -0.59 0.02 Auto Repair 0.65 0.28 Retail Sales -0.37 0.09 Legal Service 0.32 -0.39 B. Classification Results for Calibration Sample Predicted Group Membership Fast Auto Retail Legal Food Repair Sales Service Total Fast Food 21 2 14 7 44 47.7% 4.5% 31.8% 15.9% 100.0% Auto Repair 4 26 5 9 44 Actual 9.1% 59.1% 11.4% 20.5% 100.0% Group Retail Sales 16 9 13 8 46 34.8% 19.6% 28.3% 17.4% 100.0% Legal Service 6 16 2 22 46 13.0% 34.8% 4.3% 47.8% 100.0% Total 47 53 34 46 180 As shown in Table 5A, both of the functions are statistically significant based upon Wilk's lambda (p < 0.01). In addition, the group centroids (in- dustry means) for each of the four industries differ substantially. Discrimi- nant function scores are standardized so that the entire sample has a mean of 0.00 with a standard deviation of 1.00. This allows easy comparisons be- tween the groups being classified. For example, the mean for the Fast Food industry on function 1 is -0.59, while the mean for Auto Repair is 0.65, or a difference of 1.24. This substantial difference (more than one standard de- viation) indicates that Auto Repair places a much greater premium on OP- ERATIONS than on EMPLOYEE concerns relative to Fast Food companies 18 Journal of Business Strategies Vol. 17, No.1 (based on the negative coefficient for EMPLOYEE and the positive coeffi- cient for OPERATIONS for function 1). Although it is important to have statistically significant functions, it is at least equally important that the discriminant functions perform well in classifying com- panies into groups. Therefore, Table SB presents the classification results for the calibration sample based on the two functions shown in Table SA. The rows of Table 5B show the actual group/industry of which a company was a member, while the columns show the predicted group based on the discriminant functions. The compa- nies in the diagonal have correct predictions (shown in bold), while companies off the diagonal have incorrect predictions. Without prior information, we could expect to guess group membership correctly in approximately 25% of the cases by guessing that all of the companies were in the Auto Repair industry (we simply pick the group with the most members and guess that each company will be a member of this group). Lacking prior information, this is the best guess we can make. In compari- son, the discriminant model accurately predicts 45.56% (82/180) of the company's industry membership accurately. Table 5B indicates that our discriminant model accurately classifies a sub- stantially higher percentage of companies (45.6% versus 25%) than we could expect based on random guessing. The model works particularly well for the Auto Repair industry (59.1%), but not so well for Retail Sales (28.3%). The model clearly provides an increase in predictive power over random guessing and does provide support for the conceptually derived service process matrix. Validation Sample. Table 6 shows the classification results when the disciminant model shown in Table 5 is applied to the independent validation or holdout sample of 90 compa- nies. The classification accuracy for the validation sample is consistent (41.1 % or 37/90) with that of the calibration sample. Once again, the Auto Repair industry has the highest percentage of correct classifications (68.2%). In sum- mary, the fact that the discriminant model works well on the validation sample in addition to the calibration sample indicates that the classification power of the model is not dependent on the data used to estimate and develop the model. Conclusions and Future Research Directions The objective of this research was to empirically assess the differences in management challenges experienced by managers in different industries using service process matrix (SPM) as the framework. Data collected from the manag- ers of four different industries from separate cells of the SPM provide several interesting insights about the management of service operations. The results also serve to partially validate SPM while simultaneously illustrating the "fuzzy nature" of services. Spring 2000 Verma & Boyer: Service Classification 19 Table 6 Classification Results for Validation Sample Predicted Group Membership Fast Auto Retail Legal Food Repair Sales Service Total Fast Food 8 5 4 5 22 36.4% 22.7% 18.2% 22.7% 100.0% Auto Repair 1 15 2 4 22 Actual 4.5% 68.2% 9.1% 18.2% 100.0% Group Retail Sales 7 4 6 6 23 30.4% 17.4% 26.1% 26.1% 100.0% Legal Service 6 6 3 8 23 26.1% 26.1% 13.0% 34.8% 100.0% Total 22 30 15 23 90 The exploratory factor analysis of the management challenge questions revealed seven underlying factors. These factors were than compared across the four distinct service industries, and the results yielded important insights. For example, it is interesting to note that service quality was identified as the top management chal- lenge for all the respondents. This result reaffirms the importance of quality in service businesses as proposed by a number of researches (for example. Parasuraman, Zeithaml, & Berry, 1985). The importance attributed to service quality was rela- tively consistent across the four industries studied. Employee issues were the second ranked concern for the entire sample, and were of the greatest importance to the fast food industry. It is interesting to note that employee issues were of significantly less importance in legal services than for either fast food or retail sales. Another intrigu- ing difference concerned the importance of operations issues. These were the fore- most concern of auto repair shops, but were of significantly less concern to the other three industries. This result is consistent with the service process matrix, since auto repair is considered to be high customer contactlcustomization, yet the fact that operations issues were relatively unimportant to legal services (the other high customer contactlcustomization industry studied) underscores the difficulties inher- ent in clearly differentiating service types. Perhaps operations issues are less impor- tant for legal services that for auto repair because of the higher level of education and professionalism. Lawyers are known for working long hours to make partner - as a result, short term fluctuations in demand may be more easily handled by simply encouraging young associate attorneys to work long hours. In contrast, extensive overtime in auto repair is more likely to be very expensive because workers are paid on an hourly basis rather than a fixed salary. The results of the discriminant analysis of management challenges serve to validate the descriptive power of the service process matrix, yet also demonstrate 20 Journal of Business Strategies Vol. 17, No.1 that services are inherently "fuzzier" than manufacturing and more difficult to differentiate cleanly. The results show partial support for the SPM classification scheme but also reveal the limitations of theoretical classification schemes. While theoretical typologies provide an important intuitive model of the basic differences among disparate groups, there are distinct limitations to their ability to capture all (or most) differences among service firms. Theoretical service typologies are to an extent similar to classifying individuals based on demographic characteristics. Such classifications do have intuitive appeal and can broadly provide general guidelines for the groups, yet do not take into the account the individual differences among the group members. This is not to suggest that conceptual models and typologies are not valued. Instead, such typologies serve to focus our thoughts and provide an easily understood description of complex relationships. A primary cause of many misclassifications is likely to be the relati ve similar- ity of different industry groups along one dimension of the service process matrix. For example, according to the service process matrix (Figure 1), both Fast Food and Retail Sales are characterized as low customer contact/ customization, yet they differ in terms of labor intensity. Similarly, Fast Food and Auto Repair are both characterized by low labor intensity, but differ in terms of customer contact. In contrast, there is in no commonality between Fast Food and Legal Services, or between Retail Sales and Auto Repair. These two pairs of services exist in opposite corners of the service process matrix. It is not entirely unreasonable to assume that service industries in adjoining cells of the service process matrix might be misclassified as a member of the adjacent industry. On the other hand, we would expect it to be relatively rare that service industries in non-adjoining cells be mis-classified. In order to test this proposition, we break the classifications shown in Tables 5B and 6 into three groups: (l) correct classifications [1191270 =44.1 %], (2) mis-classifications along adjoining cells ofthe service process matrix [1071270 = 39.6%], and (3) complete mis-classifi- cations of cells in non-adjoining cells [16.3%]. Table 7 presents the details. This analysis provides further support for the discriminant validity of the service process matrix. The discriminant model not only predicts a greater percentage of industries than could be expected due to random guessing (44.1 % versus 25%), but the mis-classifications which do occur are more likely to be along adjoining cells in the service process matrix than in non-adjoining cells. The role of empirical analysis is to test the extent to which such typologies fully represent reality and to suggest shortcomings, which lead to further research and refinement. Toward that end, further research should seek to add to the current findings by examining additional quantitative measures in an effort to develop a more accurate classification model. In particular, further analysis should seek to better differentiate fast food and retail sales, since these two groups were misclassified disproportionately by the dis- criminant model. Spring 2000 Verma & Boyer: Service Classification Table 7 Classification Accuracy A. Calibration Sample 21 Sample Correct Partially Correct Incorrect Actual Industry Size Classification Classification (.) Classification (*) I. Fast Food 44 21 16 7 47.7% 36.7% 15.9% 2. Auto Repair 44 26 13 5 59.1% 29.5% 11.4% 3. Retail Sales 46 13 24 9 28.3% 52.2% 19.6% 4. Legal Service 46 22 18 6 47.8% 39.1% 13.0% TOTAL 180 82 71 27 45.6% 39.4% 15% B. Validation Sample Sample Correct Partially Correct Incorrect Actual Industry Size Classification Classification (.) Classification (*) 1. Fast Food 22 8 9 5 36.4% 40.9% 22.7% 2. Auto Repair 22 15 5 2 68.2% 22.7% 9.1% 3. Retail Sales 23 6 13 4 26.1% 56.5% 17.4% 4. Legal Service 23 8 9 6 34.8% 39.1% 26.1% TOTAL 90 37 18 17 41.1% 20.0% 18.9% • represents the following misclassification: 1 classified as 2 or 3 3 classified as I or 4 2 c1assi fled as 1or 4 4 classified as 2 or 3 '" represents the following misclassification: I classified as 4 3 classified as 2 2 c1assi tied as 3 4 classified as 1 22 Journal of Business Strategies References Vol. 17, No.1 Bowen, D,E., & Cummings, T.G. (1990). Suppose we took service seriously. In~ Manafi.ement Effectiveness, New York: Jossey-Bass Publishers, 1-4. Chase, R.B. (1978). Where does the customer fit in a service operation. Harvard Busi- ness Review. 56 (6), 137-142. Chase R.B. (1981). The customer contact approach to services: Theoretical bases and practical extensions. Operations Research. 29 (4),698-700. Chase, R.B., & Hayes, RH. (1991). Beefing-up operations in service firms . .sJ.Q.im Management Review. 33 (l), 15-26. Churchill, G.A. (1979). A paradigm for developing better measures of marketing con- structs. Journal of Marketinfi. Research. 16 (2), 64-73. Dess, G.G., Ireland, RD., & Hitt, M.A. (1990). Industry effects and strategic manage- ment research. Journal of Management. 16 (1), 7-27. Fitzsimmons, J.A., & Fitzsimmons, M.J. (1994). Service manafi,ement (or competitive advanta&e, New York: McGraw-Hill. Flynn, B.B., Sakakibara, S., Schroeder, RG., Bates, K.A., & Flynn, E.J. (1990). Empiri- cal research methods in operations management. Journal of Operations Manage- ment. 9 (2), 250-284. Hill, T.P. (1977). On goods and services. Review ofIncome and Wealth. 23, 12. Johnson, RA., & Wichern, D.W. (1988). Applied multivariate statistical analysis, 2nd ed. Englewood Cliffs, NJ: Prentice Hall. Judd, R.c' (1964). The case for redefining services. Journal of Marketin&. 28, 1. Karmarkar, U.S., & Pitbladdo, R. (1995). Service markets and competition.1Qurnal of Operations Mana&ement. 12 (4), 397-411. Kellogg, D.L., & Chase, RB. (1995). Constructing an empirically derived measure for customer contact. Management Science. 41 (II), 1734-1749. Kellogg, D .L. & Nie, W. (1995). A framework for strategic service management. .lmll:niU of Operations Manafi,ement. 13, 323-337. Kotler, P. (1980). Principles of marketing. Englewood Cliffs. NJ: Prentice Hall. Kim, J .• & Mueller, C.W. (1978). Factor analysis: Statistical methods and practical ~, Newbury Park, CA: SAGE Publications. Spring 2000 Verma & Boyer: Service Classification 23 Klecka, w.R. (1980). Discriminant analysis. Newbury Park, CA: SAGE Publications. Levitt, T. (1972). Production·line approach to service. Harvard Business Review. 50 (4), 41-52. Levitt. T. (1976). The industrialization of service. Harvard Business Review. 54 (5), 63- 74. Linsky, A.S. (1972). Stimulating responses to mailed questionnaires: A review. fl!J2!k Opinion Ouarterly. 39. 82·101. Lovelock, C.H. (1983). Classifying services to gain strategic marketing insights. Journal of Marketini. 41 (3), 9·20. Lovelock. C.H. (1992). A basic toolkit for service management. In Mana~in!: services: marketing. operations. and human resources 2nd ed. Englewood Cliffs. NJ: Prentice Hall. Meredith. J.R., Raturi, A., Amoako·Gyampah. K., & Kaplan, B. (1989). Alternative research paradigms in operations. Journal of Operations Management. 8 (4). 297· 326. Mersha. T. (1990). Enhancing the customer contact model. Journal of Operations Man- aument. ..2 (3),391·405. Nunnally, J. (1978). Psychometric theory, New York: McGraw·Hill. Parasuraman, A., Zeithaml, V., & Berry. L.L. (1988). SERVQUAL: A multiple-stem scale for measuring consumer perceptions of service quality, Journal of Retailin!:, 39-48. Rathmell. J.M. (1974). Marketing in the service sector, Cambridge, MA: Winthrop. Roth. A. V., & Van Der Velde, M. (1991). Operations as marketing: A competitive service strategy. Journal of Operations Management. 10 (3), 303-328. Sasser, W.E., Jr., Olsen, R.P.• & Wyckoff. D.D. (1978). Manaiement of service opera· tions: Text and cases, Boston: Allyn and Bacon. Schmenner. R.W. (1986). How can service businesses survive and prosper. Sloan Man- agement Review. 27 (3), 21·32. Shostack. G.L. (1977). Breaking free from product marketing. Journal QfMarketini. 41 (4). Sullivan, R.S. (l98l). The service sector: Challenges and imperatives for research in operations management. Journal of Operations Manat;ement, 2(4),211-214. 24 Journal of Business Strategies Vol. 17, No.1 Thomas, D.R.E (1978). Strategy is different in service businesses. Haryard Business Review. 56 (4), 158-165. Ward, P, Duray, R., Leong, G.K., & Sum, C. (1995). Business environment, operations strategy, and performance: An empirical study of Singapore manufacturers. Journal of Operations Management. 13, 99-115. Robit Verma is Assistant Professor of Operations Management in the Department of Management at the DePaul University. He holds a Ph.D. in Operations Manage- ment from the David Eccles School of Business, University of Utah, an M.S. from the University of Utah and a B. Tech. from the Indian Institute ofTechnology, Kanpur, India. His research interests include Operations Strategy, Product/Service Design, Operations-Marketing interdisciplinary issues, and business education. His research is published in International Journal of Service Industry Management, Interna- tional Journal of Operations and Production Management, Journal ofOperations Management, Journal ofProduct Innovation Management, Omega, Production and Operations Management, Journal ofService Research and other journals. His cur- rent and research has been funded by First ChicagolNBD Corporation, APICS Edu- cation & Research Foundation, Marketing Science Institute, NCR Knowledge Labo- ratory, United States Forest Service and other organizations. Kenneth K. Boyer is Assistant Professor of Operations Management in the De- partment of Management at the DePaul University. He earned a B.S. in mechanical engineering from Brown University, and an M.A. and Ph.D. in Business Admin- istration from the Ohio State University. Dr. Boyer worked as a project engineer with General Dynamics Electric Boat Division in Groton, Connecticut. Dr. Boyer's research interests focus on the strategic management of operations and technol- ogy management, including e-commerce. He has published articles in Manage- ment Science, Business Horizons, Production and Operations Management, De- cision Sciences, Journal of Operations Management, Omega and the Interna- tional Journal of Operations and Production Management. Ken is an Associate Editor for the Journal of Operations Management and is a member of the Acad~ emy of Management, Decision Sciences Institute and the Society of Manufactur- ing Engineers. Service Classification and Management Challenges