Statistical Process Control for Controlling Service Quality - Essay Example

? Statistical Process Control whilst primarily a manufacturing quality technique can be usefully applied in service industries Introduction- ResearchBackground “The consumer compares his expectations with the service he perceives he has received, i.e. he puts the perceived service against the expected service. The result of this process will be the perceived quality of service” - (Gronroos, 1984, p. 37) The above lines are not only reflecting arguments of Gronroos (1984) but also shedding light on of the biggest challenge for service provider regarding maintenance of service quality and fulfilling service gap. According to above lines, delivery of service is being compared in context to expectation of customers and divergence of expected service quality from delivered quality creates the gap. Ladhari (2009) stated that four characteristics of service like intangibility, heterogeneity, perishability and inseparability make it different from manufacturing offering. . Markovic (2006) argued that manufacturing sector should not be compared with service sector because customers might act as co-producer in service delivery process while customer involvement is negligent in manufacturing process. In such context, Khan (2003) stated that intangibility and inseparability make it difficult to control service quality while there are statistical procedures available to manage quality of manufacturing process. In such context, Chakrabarty and Tan (2007) found that unlike the manufacturing sector, it took time for service sector to realize the importance of Statistical Process Control (SPC) in managing quality. Sulek (2004) argued that most of the common statistical control mechanism can also be used in service sector to manage quality but little bit recalibration of the statistical model is needed in order to utilize it accurate manner in service environment. Discussion Six Sigma & Control Chart Antony (2006) defined the term Sigma as the deviation from service performance characteristics mean while objective of deploying Six Sigma in service sector is to reduce the scope of variation and subsequently improve quality. In order to control variation in the service performance, specific control limit is being assigned (SLupper). Aim of the service performance would be not to cross the upper control limit or the maximum tolerance zone (Yilmaz and Chatterjee, 2000). In case of Six Sigma, distance between SLupper and service process mean is equal to six standard deviations and in this way term “Six Sigma” has been arrived. In case of six sigma process, deviation in service performance caused by external uncontrollable influences would not exceed the limit of 3.4 parts per million or 3.4 times the service process will show defect out of 1 million times (Antony, 2006). Antony (2006) and Hoerl (2001) stated that Six Sigma process can be applied to service processes like order entry, invoicing, shipping, baggage handling, payroll processing etc. On the other hand, Yilmaz and Chatterjee (2000) measured that defect rate in service sector is less than 3.5 sigma quality level which means 23,000 times the service process will show defect out of 1 million times. In such context, applying Six Sigma as SPC would improve the service performance level to 99.38 per cent. Hoerl and Snee (2002 and 2003) identified benefits of deploying Six Sigma in service sector as 1- decrease in service defect rate which would automatically increase cost efficiency in the service process, 2- management decision would be guided by data driven statistical analysis which would decrease the scope of assumption bases errors and 3- decrease in service defect would significantly decrease customer complaints. Some practical examples can be cited in order to highlight usefulness of Six Sigma model in service sectors. Table 1: Practical Evidences of Implementation of Six Sigma in Service Organizations Organization Benefits J P Morgan Chase (Global Investment Banking) Applying Six Sigma model has helped the company to reduce flaws in service delivery processes like payment handling, account opening, cheque-book ordering by 30%. Citibank Applying Six Sigma model has helped the bank to reduce credit decision cycle from 3 days to 1 day (67% reduction). (Source: Antony, 2006) Taguchi loss function Taguchi (1987) gave the concept of loss function in order show how financial loss is related to predetermine performance parameters. Bergman and Klefsjo (2003) stated that in Taguchi loss function, every deviation from predicted mean value would increase financial incurred in the process. Figure 1: Taguchi Loss Function (Source: Gregorio and Cronemyr, 2011, p. 167) Gregorio and Cronemyr (2011) stated that Taguchi Loss Function is generally used in manufacturing sector but there were some exceptions where the function has been used in service scenario. Interesting fact is that Vargo and Lusch (2004) argued that service sector is much more competitive in comparison to product driven market. In such context, applicability of Taguchi Loss Function in service sector should be reconsidered in order to use the statistical model in effective way. Taguchi (1987) identified four categories such as higher-the-better (e.g. performance of the machine), lower-the-better (e.g. waiting time in service queue), nominal-is-best (e.g. scheduling of time) and asymmetric. According to Gregorio (2008), simple descriptive statistics such as mean, median, variance or standard deviation are being used in order to calculate Taguchi Loss Function in service industry. Despite wide use in manufacturing sector and economic studies, Gregorio and Cronemyr (2011) found that Taguchi Loss Function might not work in correct manner in service functions due to issues like intangibility and heterogeneity. In such context, following evidence can be used to denote the use of Taguchi Loss Function in service sector. Table 2: Practical Evidences of Implementation of Taguchi loss function in Service Organizations Organization Benefits Service Division of Siemens Industrial Turbomachinery AB (Located in Finspong, Sweden) Loss function has been identified at the 45th day of operation and critical service areas were identified which were increasing financial loss of the company. (Source: Gregorio and Cronemyr, 2011) Least Absolute Value Regression (LAV) In case of service sector, use of ordinary least squares (OLS) regression to develop control chart is not a new phenomenon but use of least absolute value (LAV) regression control quality of service delivery has been started recently (Utley and May, 2009). In case of LAV model, forecasting parameters are being selected as independent variable in the control panel and then regression is being done in order to adjust the sensitivity of residuals control chart (Alwan, 2000). However, Hensley and Dobie (2005) and Sulek, Marucheck and Lind (2006) showed doubt over the integration of a fully functional statistical model in controlling service quality because possibility of existence of autocorrelation between predictors. Taking help of the work of Alwan (2000), it can be said that LAV model minimizes the sum of modulus value the residuals and LAV regression is free from outlier Problems. Using LAV model can help the service companies to determine absolute value of deviation in the process and on the basis of the result, control measures can be taken in order to reduce residue statistics. There can be several algorithms for calculating LAV model parameters and following equation can be used as the basic regression model. ^ y = b + mx xi= predictors in the service process y= Performance quality of the service process b= regression constant m= slope in regression line pi= positive deviation from process quality mean; yi- (b+mxi) ni= negative deviation from process quality mean; (b+mxi) - yi LAV= ? (pi + ni); b+ mxi + pi + ni=yi, i= 1,.......n pi, ni > 0 while b, m unrestricted It is evident from the above mathematical expression that LAV regression only focuses on residue statistics and same model can be used in service sector in order to identify defects in service delivery process. Using statistical techniques, upper control limit and lower control limit can be derived in the following manner. Zi= yi - ^ yi Zi= centreline value at it h observation, in such context, mean process value can be derived by the following equation; Zmean= ?ni=1(Zi)/n Therefore, Upper control limit (UCL) = Zmean + 3? and Lower Control Limit (LCL) = Zmean - 3? ?= Standard deviation Using the upper control limit and lower control limit assumptions, zone of tolerance can be created in case of service processes. Following evidence can be used to understand the use of Least Absolute Value Regression (LAV) in case of service organizations to control the quality of service process. Table 3: Practical Evidences of Implementation of LAV Regression in Service Organizations Organization Benefits AT&T The telecom giant has used the LAV regression model in order to monitor relationship between key performance measures like total revenue/month and customer call frequency in every month. Data collected for a length of quarter and on the basis of the result, the company prepared control chart. Finally, using LAV regression helped AT&T to identify gap in customer query handling service and in future the company used the result to develop strategy to improve service quality. (Source: Utley and May, 2009) Kolmogorov-Smirnov goodness-of-Fit Test Probably one of the most complex but equally effective Statistical Process Control (SPC) techniques which are very frequently used in manufacturing sector (Leyer and Moormann, 2012) but the question is that whether this SPC is good for measuring quality in service sector? Gronroos (2007) and Sampson and Froehle (2006) argued that service companies like financial institutions, travel companies, telecommunication sectors consider high productivity as the primary success factor in case of service delivery process. However, Zomerdijk and de Vries (2007) criticized such assumptions and stated that automated or marginal customer integration should be considered as important performance predictor for service companies. In such context, question may arise that how quality of service can be controlled where customers play important role throughout the service life cycle? Using Kolmogorov-Smirnov goodness-of-Fit Test might help in answering the question. Let’s try to understand how customers are integrated in the service delivery process. P [CIPi] = percentage of customer input at particular service process O [CIPi] = total number of occurrences per customer Aj[R] = requesting activities for each customer O (Aj[R]) = at particular event log, total occurrence of service requesting activities P [CIPi] = O [CIPi]/?j O(Aj[R])*Aj[R]; i=1........n According to the above model, greater the value of P [CIPi] would indicate higher involvement of customers in service process. It is evident from the research works of Schau, Hemetsberger, and Kozinets (2008) and Sheikh (2003) that marketers should consider the customer involvement in service delivery as the sign of quality. In such context, it can be said service organizations can use P [CIPi] as SPC to set customer driven quality parameters in service environment. So what is the use of Kolmogorov-Smirnov goodness-of-Fit Test to control quality in service process? In service environment, Kolmogorov–Smirnov (K–S) goodness-of-fit test compares hypothetical quality based cumulative distribution function (cdf) ^F(x) with actual quality cdf Fn(x) and on the basis of comparison, fitness of quality is being tested. Fn(x) = proportion of observations ranging from X1……..Xn which might be less than or equal to mean process value (Drew, Glen and Leemis, 2000; Rigdon and Basu, 2000). Fn(x) = I(x)/n…….[n= total number of customer encounter in the service process & I(x)= proportion of service quality value less than or equal to process mean]. In such context, Dn can be considered as the gap in the service quality or the vertical distance between Fn(x) and ^F(x). Dn (Service quality gap) = supremum (x) (sup) {|Fn(x)- ^F(x)|} Taking help of the Law and Kelton’s (2000) statistical explanation, Dn (Service quality gap) can be derived in the following manner; Dn+ (greater than the perceived quality of service) = Maxi=1,2….n{i/n- ^F(xi) and Dn- (lower than the perceived quality of service)= Maxi=1,2….n {^F(xi) – i-1/n} i= no. of occasions and Xi is i th order statistics. On the basis of the above mathematical explanation, following service quality control chart can be created. Figure 2: Goodness-of-Fit Plot for Service Quality (Source: Evans, Drew and Leemis, 2008, p. 1398) It is evident from the quality diagram that service organizations need to control the above mentioned service gaps in order to control the quality of customer driven service process like restaurant food service, banking service, medical services etc. Table 4: Practical Evidences of Implementation of Goodness-of-Fit Test in Service Organizations Organization Benefits A renowned Financial Organization based in Frankfurt, Germany The company used Goodness-of-Fit Test in order to solve problems regarding customer grievances for loan processing speed. On the basis of test results, the company understood the customer’s expected speed of loan processing. At the end, the financial company took necessary steps to increase the loan processing speed and subsequently enhanced customer driven quality of the service process. (Source: Leyer and Moormann, 2012) Considering the research works of Yazici and Yolacan (2007) reveals the fact that any integration of Statistical Process Control (SPC) in service environment can be normalized and due to such aspect, controlling stochastic variables in service process is very important. Throughout the paper, the researcher has discussed four different SPC that can be used in order to control quality in the service process. Soyuer, Kocamaz and Kazancoglu (2007) argued that job scheduling can play vital role in predicting service quality and there are mathematical algorithms that can be used to address scheduling problem. In the next section, the researcher will compare all the four different SPC in order to reach a definitive conclusion. Table 5: Comparative Analysis SPC Advantage Disadvantage Probability of Use in Service Organization Six Sigma & Control Chart Easy to develop the control chart without compromising precision It is difficult to identify the control variable High Taguchi loss function Quality is directly related to financial result hence controlling service quality would increase profitability of the company. Statistical parameters are not rigidly defined. Moderate Least Absolute Value Regression (LAV) Provide the absolute measure of the quality gaps in the service process. The process does not cover outlier or extreme value problem. Low Kolmogorov-Smirnov goodness-of-Fit Test Works very well in measuring the quality gap in service processes where customer involvement is very high. Complex mathematical analysis creates problem for frontline employees to understand the implication. Very Low It is evident from the above comparative analysis that Six Sigma & Control Chart is the probably most popular SPC among service organizations in order to control quality of operation. Spring and Araujo (2009) argued that service organizations need to redesign its service offering in order to reduce amount of errors. Although, at present, use of Least Absolute Value Regression (LAV) and Kolmogorov-Smirnov goodness-of-Fit Test are least popular among service organizations but in future, these tools have the potential to become vital Statistical Process Control mechanism to control service quality. Conclusion It is evident from the above mentioned explanations, theoretical arguments and evidences provided by various research scholars that controlling service quality is a complex issue and using Statistical Process Control (SPC) to control service quality is still a matter of argument among scholars (Maglio, Kieliszewski and Spohrer, 2010; Prajogo, 2006). In such context, this essay tried to pick four different SPC mechanisms that are being used in service organizations and compared their effectiveness. Therefore, it can be concluded that although Statistical Process Control whilst primarily a manufacturing quality technique but it can be usefully applied in service industries. Reference List Alwan, L., 2000. Statistical Process Control. Boston, MA: McGraw-Hill. Antony, J., 2006. Six sigma for service processes. Business Process Management Journal, 12(2), pp. 234-248. Bergman, B. and Klefsjo, B., 2003. Quality from customer needs to customer satisfaction. 2nd ed. Lund: Student literature. Chakrabarty, A. and Tan, K., 2007. The current state of six sigma application in services. Managing Service Quality, 17(2), pp. 191-208. Drew, J. H., Glen, A. G. and Leemis, L. M., 2000. Computing the cumulative distribution function of the Kolmogorov–Smirnov statistic. Computational Statistics and Data Analysis, 34, pp. 1–15. Evans, D. L., Drew, J. H. and Leemis, L. M., 2008. The Distribution of the Kolmogorov–Smirnov, Cramer–von Mises, and Anderson–Darling Test Statistics for Exponential Populations with Estimated Parameters. Communications in Statistics—Simulation and Computation, 37, pp. 1396–1421. Gregorio, R. and Cronemyr, P., 2011. From expectations and needs of service customers to control chart specification limits. The TQM Journal, 23(2), pp. 164-178. Gregorio, R., 2008. Breaking the customer code. A model to translate customer expectations into specification limits. Linko?ping: Linko?ping Institute of Technology. Gronroos, C., 2007. Service management and marketing: Customer management in service competition. Hoboken, NJ: Wiley. Gronroos, C., 1984. A service quality model and its marketing implications. European Journal of Marketing, 18(4), pp. 36-44. Hensley, R. and Dobie, K., 2005. Assessing readiness for six sigma in a service setting. Managing Service Quality, 15(1), pp. 82-101. Hoerl, R. W. and Snee, R. D., 2002. Statistical thinking – improving business performance. Belmont, CA: Thomas Learning, Hoerl, R. W. and Snee, R. D., 2003. Leading six sigma. Englewood Cliffs, NJ: Prentice-Hall. Hoerl, R. W., 2001. Six sigma black belts: What do they need to know? Journal of Quality Technology, 33(4), pp. 391-435. Khan, M., 2003. ECOSERV: Ecotourists’ quality expectations. Annals of Tourism Research, 30(1), pp. 109-24. Ladhari, R., 2009. A review of twenty years of SERVQUAL research. International Journal of Quality and Service Sciences, 1(2), pp. 172-198. Law, A. M. and Kelton, W. D., 2000. Simulation modeling and analysis. 3rd ed. New York: McGraw-Hill. Leyer, M. and Moormann, J., 2012. A method for matching customer integration with operational control of service processes. Management Research Review, 35(11), pp. 1046-1069. Maglio, P. P., Kieliszewski, C. A. and Spohrer, J. C., 2010. Handbook of service science. Berlin: Springer. Markovic, S., 2006. Expected service quality measurement in tourism higher education. Nase Gospodarstvo, 52(1/2), pp. 86-95. Prajogo, D., 2006. The implementation of operations management techniques in service organizations. International Journal of Operations & Production Management, 26(12), pp. 1374-90. Rigdon, S. and Basu, A. P., 2000. Statistical methods for the reliability of repairable systems. New York: John Wiley & Sons. Sampson, S. E. and Froehle, C. M., 2006. Foundations and implications of a proposed unified services theory. Production and Operations Management, 15(2), pp. 329-43. Schau, H. J., Hemetsberger, A. and Kozinets, R. V., 2008. The wisdom of consumer crowds: Collective innovation in the age of networked marketing. Journal of Macromarketing, 28(4), pp. 339-54. Sheikh, K., 2003. Manufacturing resource planning (MRP II): With an introduction to ERP, SCM, and CRM. New York, NY: McGraw-Hill. Soyuer, H., Kocamaz, M. and Kazancoglu, Y., 2007. Scheduling jobs through multiple parallel channels using an expert system. Production Planning & Control, 18(1), pp. 35-43. Spring, M. and Araujo, L., 2009. Service, services and products: rethinking operations strategy. International Journal of Operations & Production Management, 29(5), pp. 444-67. Sulek, J., 2004. Statistical process control in services. International Journal of Services Technology and Management, 5(5/6), pp. 522-31. Sulek, J., Marucheck, A. and Lind, M., 2006. Measuring performance in multi-stage operations: An application of cause-selecting control charts. Journal of Operations Management, 24, pp. 711-27. Taguchi, G., 1987. System of experimental design: Engineering methods to optimize quality and minimize costs. Vols 1 & 2. White Plains, NY: UNIPUB/Kraus International Publications. Utley, J. S. and May, J. G., 2009. Monitoring service quality with residuals control charts. Managing Service Quality, 19(2), pp. 162-178. Vargo, S. L. and Lusch, R. F., 2004. Evolving to a new dominant logic for marketing. Journal of Marketing, 68(1), pp. 1-17. Yazici, B. and Yolacan, S., 2007. A comparison of various tests of normality. Journal of Statistical Computation and Simulation, 77(2), pp. 175-83. Yilmaz, M. R. and Chatterjee, S., 2000. Six sigma beyond manufacturing – a concept for robust management. IEEE Engineering Management Review, 28(4), pp. 73-80. Zomerdijk, L. G. and de Vries, J., 2007. Structuring front office and back office work in service delivery systems. International Journal of Operations & Production Management, 27(1), pp. 108-31. Read More