Interaction of Service Attributes on Customer Satisfaction

Dissatisfied Satisfied 1 TONTINI, G.. Interaction of Service Attributes on Customer Satisfaction. In: 2011 IEEE International Conference on Quality and Reliability, 2011, Bangkok. Proceedings of..., 2011. p. 1-5 Interaction of Service Attributes on Customer Satisfaction Prof. Dr. Gérson Tontini Department of Business Management, University of Blumenau, Blumenau, Brazil (tontini@furb.br) Abstract - This paper has as objective to verify how service attributes, classified by Kano Model as must-be, one-dimensional or attractive, interact on customers The research was carried out interviewing 119 customers of pizzeria restaurants. The Kano Model category of four attributes is confirmed in the first part. In the second part, a full factorial research explores the interaction among a must-be (), a onedimensional () and two attractive attributes (Extra Products and Pizza Border). The results of this research show that a superior level of attractive and one-dimensional attributes only have a full impact on customer satisfaction if must-be attributes are fulfilled. No interaction was found among one-dimensional and attractive attributes. The managerial implication is that companies should identify and keep must-be attributes in adequate level of performance. Only in this way the attractive and one-dimensional attributes, which could bring differential in the market, have a full impact on customer Keywords - Customer Satisfaction, Kano Model, Quality Management, Service Quality. I. INTRODUCTION The impact of customer satisfaction on business success has been widely discussed in the literature. According to [1], an increase in customer loyalty by 5% can increase the profit of a business by 100% because loyal customers purchase the products or services of a company more often and in greater quantities. Service quality and customer satisfaction are the main antecedents of customer loyalty. Reference [2], studying companies that are part of the Swedish Customer Satisfaction Barometer, found that an increase of 1% in the customer satisfaction index was associated with 2.37% increase in the return over investment. On the other hand, a decrease of 1% in this index was associated with a decrease of 5.08% in the return over investment. These results show that while increasing customer satisfaction is important, avoiding customer dissatisfaction is critical. However, how can a company continuously satisfy its customers? Satisfaction is related to the fulfillment of implicit and explicit customer needs by the totality of service attributes. So, it becomes important to find out how attributes performance impact on customer Most of the traditional techniques aimed to identify attributes importance assume that there is a linear relationship between attribute performance and customer Kano Model of Customer Satisfaction proposes that the relationship between the existence or performance of attributes and customer satisfaction is non-linear, classifying the attributes as must-be, onedimensional and attractive factors (Fig. 1) [3][4][5]. Must-be attributes are related to the basic functions of the service. Generally, customers don t perceive the presence of these attributes but their absence brings strong dis For one-dimensional attributes, satisfaction is proportional to performance. Higher performance brings higher satisfaction and vice versa. Attractive attributes bring superior satisfaction if present, but they don t bring dissatisfaction if absent or insufficient. Two other attributes may be identified in the Kano Model: Neutral and Reverse attributes. Neutral attributes don't bring either satisfaction or dis Reverse attributes bring more satisfaction if absent than if present. It is very important to identify which attributes should be improved to keep company competitiveness. The most used method for identification of which attributes should be improved in services is the Importance Performance Analysis, or IPA [6][7]. Although largely used, IPA presents several limitations that have been criticized in the scientific literature [8][9][10][11][12]. One limitation is that IPA doesn t consider the non-linear relationship between attribute performance and customer Some papers have proposed methods to overcome this limitation by joint use or fusion of IPA and Kano Model [13][14][15][16]. Non-existent or low performance Must-be Attractive Fig. 1 Kano Model [3] One-dimensional Existent or high performance

The assumption behind most published papers is that customer satisfaction is an additive result of individual attributes. Then, a question remains: would it be possible that attributes effect on customer satisfaction is significantly affected by the interaction with other attributes? If the answer to this question is affirmative, the identification of interactions may lead managers to better improvement decisions to increase customers One paper appears in the literature studying how attributes with different Kano Model classification interact to affect customer Reference [17] studies only the interaction between must-be and attractive attributes. No work was found in the literature that studies how the three types of attribute (attractive, must-be and one-dimensional) interact to bring satisfaction to customers. The study of these interactions may lead to the development of more refined methods that point out not only the relevant attributes, but also their best combination. Thus, the study of these interactions is still an area to be addressed. The present work has as objective to study how attractive, one-dimensional and must-be attributes interact to affect customers II.. METHODOLOGY This work was carried out interviewing a sample of 135 undergraduate business students, customers of pizzeria restaurants. The sample is composed of students present in the classroom during data collection and that voluntarily consented to participate in the study. As the students come from different regions and social classes, and frequent different pizzerias, the sample was considered adequate for research generalization. A pretest with 25 subjects was carried out for questionnaire adjustments. Three versions with different questions' sequence were used to decrease the possible effect of questions sequence. After final data collection, 16 questionnaires with invalid or inconsistent answers were eliminated. Invalid or inconsistent are those that had several blank answers or the same answer for all questions. Then, a sample of 119 students was used in the research. Any service could be used in the research but, since the objective of this work is to study attributes interaction and not to make an exploratory study, pizzeria was chosen because it is of frequent use by all respondents. It is an all you can eat pizzeria where customers are continuously served from a wide variety of pizzas, pasta and other kinds of food. This kind of pizzeria is very popular among young people and families in the south region of the Country. In the first part of the research, the traditional Kano Model category of four attributes was identified: Restaurant perceived cleanness (Must-be), Waiters courtesy (One-dimensional), Choice of pasta besides pizza (Attractive) and Diversified filled border, i.e., filing the pizza border with the same pizza s topping (Attractive). This last attribute was an innovation, not offered in the market at the time of the research. From this point on, we call the choice of pasta as Extra Products and diversified filled border simply as Filled border. The Kano Model classification of each attribute (A Attractive, M Must-be, O One-dimensional, N Neutral, R Reversal) is identified by customers answering two questions. One question asks about customer feeling with the presence or sufficiency of the attribute. The other question asks about the feeling with its absence or insufficiency. The answers for these two questions are used to identify the Kano Model classification of each attribute for each customer (Fig. 1). The attributes of the current research are the same used by [12], but the classification of is different. In the current research, it is classified as a must-be attribute. In Reference [12] it was classified as one-dimensional. This difference may be either, due to time evolution of attribute s classification [18], due to sample or service differences, or due to symmetric answers because of questions' sequence (presence and absence). To prevent or decrease the possible symmetry, the sufficiency and insufficiency questions, to identify the Kano Model classification, were placed in random order. In the second part, the interaction among the four attributes is studied using full factorial research. The full factorial design verifies how the attractive, onedimensional and must-be attributes interaction affects customers Having 16 answers for each subject and 119 subjects, the total is 1904 cases (16 * 119). The missing data was treated using the SPSS leastwise procedure, resulting in 1,805 valid cases. Samples of five cases, randomly chosen, were averaged and used to form the final 361 cases. The SPSS stepwise regression procedure was used to identify the final equation. The effect of each attribute is calculated by the average of customer satisfaction when it is present less when it isn t. For example, the effect of courtesy is given by the average satisfaction of all combinations that it has a high level, less the satisfaction with all combinations that it has a low level. The logic behind the factorial research is that the effect of presence or absence of other attributes is canceled by the combination of their symmetric answers. One advantage of this method, called conjoint analysis [19], is that it measures attributes effect on satisfaction by forcing the respondent to make a tradeoff between attributes' combination. The last part asked information about frequency of going to pizzerias, age and gender. III. RESULTS About 50% of the respondents are male, most with age lower than 23 years old (undergraduate students), and 54.6% frequent pizzeria once or more per month. The number of pizzerias that respondents go more frequently is 24, with 50.4% of the students going mostly to three different ones.

DI - Dissatisfaction Index TABLE 1 KANO CLASSIFICATION OF THE RESEARCHED ATTRIBUTES Extra Products Filled Border Attractive 0 31 56 70 One-dimensional 32 61 5 13 Must-be 82 15 9 6 Neutral 3 4 37 22 Reverse 0 0 1 0 Questionable? 2 8 11 8 Total answers 119 119 119 119 Customer Satisfaction Index (CSI) SI 0.27 0.83 0.57 0.75 DI 0.97 0.68 0.13 0.17 Table 1 shows the results of Kano Model classification for all respondents. As it can be seen, most customers classified as must-be, as one-dimensional, and Extra Products and Filled Border as attractive attributes. Customer Satisfaction Index (CSI) [4] is a method to identify the attributes classification. It is the rate of customers who become satisfied with attributes presence or sufficiency (SI Satisfaction Index) and the rate of customers who become dissatisfied with attributes absence or insufficiency (DI Dissatisfaction Index). These indexes are used in a dispersion graph that confirms the classification of the researched attributes (Fig. 2). Table 1 and Fig. 2 confirm Kano Model classification of the researched attributes. To study the attribute s interaction, the initial model included as independent variables the four studied attributes and all possible interactions. The residuals of the final regression equation follow a normal distribution (Skewness = 0.223, Std. Error = 0.122; Kurtosis = 0.122, Std. Error = 0.244; Jarque-Bera = 4.05 < 5.99), indicating that the regression equation is valid [20]. 1,0 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 Attractive Variety Neutral Filled Border Border Extra Products 0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 SI - Satisfaction Index Performance Must-be TABLE 2 REGRESSION EQUATION OF SATISFACTION ABOUT EXISTENCE OF THE ATTRIBUTES AND THEIR INTERACTIONS. Model No std. Coefficients Std. B Error Std. Coeff. (Beta) T p-value (Constant) -4.20 0.14-30.63 0.000 2.98 0.19 0.50 15.33 0.000 X 2.41 0.19 0.35 12.42 0.000 0.85 0.14 0.14 6.18 0.000 Filled Border 0.41 0.14 0.07 2.99 0.003 X Extra Products 0.75 0.19 0.11 3.88 0.000 X Filled Border 0.89 0.19 0.13 4.57 0.000 Extra Products 0.28 0.14 0.05 2.05 0.041 R 2 = 0.897 R 2 adj.= 0.895 Durbin Watson = 2.079 N = 361 Table 2 shows the results of the regression model. It shows that there is an interaction between cleanness (must-be) and all other attributes. The result demonstrates that the effect of the one-dimensional and attractive attributes in customer satisfaction is very reduced if the must-be attribute is not fulfilled. The impact of the attractive attributes on customer satisfaction decreased 68% for the Filled Border and 73% for Extra Products ( B coefficients, table 2). (onedimensional) has an impact reduction of 74%. It shows the importance of must-be attributes fulfillment in order to increase customer These findings are the same of [17] that studied only the interaction between must-be and attractive attributes. The authors found that the effect of attractive attributes on customer satisfaction becomes much less intensive if any must-be attribute is unfulfilled. Table 3 shows the evolution of the stated satisfaction due to attributes presence. The first column presents an index of attributes combination. The following four columns indicate if the attributes are present or not in each combination. The number 1 means that the attribute is present or has high performance, while 0 means that it is not present or has low performance. Each line represents a combination of all attributes. The sixth column shows the average of stated satisfaction for each combination. The seventh column presents the increase in stated satisfaction in relation to the preceding combination. Column eight presents p-value of the satisfaction increase (probability of the satisfaction increase being neutral). Fig. 2 Customer Satisfaction Index Graph

Satisfaction TABLE 3 EVOLUTION OF STATED SATISFACTION DUE TO THE PRESENCE OF THE ATTRIBUTES Attributes Evolution 1 y Courte- Extra Filled Stated Satisf. p- Products Border Satisf. Increase value 1 0 0 0 0-4.40 2 1 0 0 0-1.13 3.28 0.000 3 1 1 0 0 1.91 3.04 0.000 4 1 1 1 0 3.14 1.23 0.000 5 1 1 1 1 4.46 1.33 0.000 Attributes Evolution 2 Products Border Satisf. Increase value Extra Filled Stated Satisf. p- 1 0 0 0 0-4.40 2 0 0 0 1-3.80 0.60 0.001 3 0 0 1 1-3.40 0.40 0.001 4 0 1 1 1-2.78 0.62 0.005 5 1 1 1 1 4.46 7.25 0.000 The graph of satisfaction with attributes presence (Evolution 1), showed in Fig. 3, presents the following sequence: Low level or not presence of all attributes (1) => + (2) => + (3) => + Extra Products (4) => + Filled Border (5). Only the introduction of good cleanness, a must-be attribute, brings a decrease in expected dissatisfaction of 3.28 (Table 3). The introduction sequence of the other attributes shows that the one-dimensional attribute () has a stronger effect than the attractive ones (Extra products and Filled border). This effect agrees with the Kano Model assumption that absence or low level of onedimensional attributes brings customer dissatisfaction, and their presence or high level brings superior In addition, Fig. 3 shows that, together with the must-be attribute, a high level of the onedimensional attribute may be enough to bring customer The lower line (Evolution 2) of Fig. 3 shows the increase of stated satisfaction when the attributes are introduced in the following order: Low level or not presence of all attributes => + Filled Border (1) => + Extra Products (2) => + (3) => + (4). Table 3 shows that the increase in satisfaction ranges from 0.4 to 0.6 when each one-dimensional and attractive attribute is improved or offered. 5 4 3 2 1 0-1 -2-3 -4-5 -4.40 Satisfaction Evolution -1.13-3.80 1.91-3.40 3.14-2.78 1 2 3 4 5 Evolution 2 Evolution 1 4.46 Fig. 3 - Evolution of Stated Satisfaction Due to Attributes Presence The effect of the attractive attributes (Extra products and Filled border) is only 32% and 42% of what they can cause when (must-be) is in superior level. The influence of (one-dimensional) is much more affected. Its effect in customer satisfaction is about 20% of the effect it has when the must-be attribute () is in superior level. IV. CONCLUSION The fulfillment of customers needs is very important to the survival of any company in a competitive market. It is important because satisfied customers tend to repeat consumption and are less sensitive to prices. This fulfillment depends on performance of the service s attributes. So, the identification of the relationship between attributes performance and customer satisfaction becomes a key issue to market success. The non-linear relationship between attributes performance and customer satisfaction brings a very important issue when a company wants to find out how to improve customer However, the identification and independent improvement of attractive, must-be and one-dimensional attributes don t guarantee the desired effect in customer Although published papers identify the Kano Model categories of the attributes, and use this classification together with the Importance Performance Analysis to take improvement decisions, they do not consider attributes interaction. This work is at the beginning of researches about how attributes interaction affects customer The findings in this research, in consonance with those of [17], show that must-be attributes have strong interaction with one-dimensional and attractive ones. The non fulfillment of must-be attributes reduces the effect of increasing or offering one-dimensional and attractive attributes. This finding shows that low performance of must-be attributes can t be compensated by superior performance or presence of other attributes. Customers will remain dissatisfied. The managerial implication is that companies should identify and keep must-be attributes in adequate performance level. Only in this way the "attractive" and "one-dimensional" attributes, which could bring differential in the market, have full effect on customer A limitation of the current research is that it included only one must-be and one one-dimensional attribute. Similar to what was found about the interaction between must-be and attractive attributes by [17], a question remains: how is the interaction between more than one must-be and one-dimensional attributes? Furthermore, since low performance of onedimensional attributes can cause dissatisfaction to customers, why no interaction with attractive attributes was found? These questions remain and should be further studied.

REFERENCES [1] F. F. Reichfeld, and W. E. Sasser, Zero-defections: quality comes to services, Harvard Business Review, Sept./Oct., pp. 105 111. 1990. [2] E. W.Anderson, and V. Mittal, Strengthening the satisfaction-profit chain, Journal of Service Research, vol. 3, no. 2, pp. 107-120, 2000. [3] N. Kano, N. Seraku, F. Takahashi, and S.Tsuji, Attractive quality vs must be quality, Journal of the Japanese Society for Quality Control, vol. 14, no. 2, pp. 39-48, 1984. [4] C. Berger, R. Blauth, D. Boger, et al, Kano`s methods for understanding customer-defined quality, Center for Quality Management Journal, vol 4, pp. 3-36, Fall 1993. [5] K. Matzler, H. H. Hinterhuber, F. Bailon, and E. Sauerwein, How to delight your customers, Journal of Product & Brand Management, vol. 5, no. 2, pp. 6-18, 1996. [6] J. A. Martilla, and J. C. James, Importanceperformance analysis, Journal of Marketing, vol. 41, pp. 77-79, 1977. [7] N. Slack, The Importance-Performance Matrix as a Determinant of Improvement Priority, International Journal of Operations & Production Management, vol. 14, no 5, pp. 59-75, 1994. [8] H. Oh, Revisiting importance-performance analysis, Tourism Management, vol. 22, pp. 617-27, 2001. [9] S. C. Ting, and C. N. Chen, The asymmetrical and non-linear effects of store quality attributes on customer satisfaction, Total Quality Management, vol. 13, no 4, pp. 547-569, 2002. [10] M. S. Garver, Best practices in identifying customer-driven improvement opportunities, Industrial Marketing Management, vol. 32, pp.455-66, 2003. [11] K. Matzler, and E. Sauerwein, E. (2002), The factor structure of customer satisfaction: an empirical test of the importance grid and the penalty-rewardcontrast analysis, International Journal of Service Industry Management, vol. 13, no. 4, pp. 314-332. [12] G. Tontini, and A. Silveira, Identification of satisfaction attributes using Competitive Analysis of the Improvement Gap, International Journal of Operations & Production Management, vol. 27, no. 5, pp. 482-500, 2007. [13] C. C. Yang, The refined Kano s model and its application, Total Quality Management, vol. 16, no. 10, pp. 1127-1137, 2005. [14] K. Matzler, M. Fuchs, and A. K. Schubert, Employee satisfaction: does Kano s model apply? Total Quality Management, vol. 15, no. 9-10, pp. 1179-1198, 2004. [15] J. Mikulić, and D. Prebezӑc, Prioritizing improvement of service attributes using impact range-performance analysis and impact-asymmetry analysis, Managing Service Quality, vol. 18, no. 6, pp. 559-576, 2008. [16] G. Tontini, J. D. Picolo, Improvement Gap Analysis, Managing Service Quality, vol. 20, no. 6, pp. 565-584, 2010. [17] G. Tontini, and A. J. Sant Ana, A..J. Interaction of a attractive and must-be attributes of a service in customer satisfaction (in Portuguese), Revista Produção, vol. 18, no. 1, pp. 112-125, 2008. [18] L. Nilson-Witell, and A. Fundin, Dynamics of service attributes: a test of Kano s theory of attractive quality, International Journal of Service Industry Management, vol. 16, no. 2, pp. 152-168, 2005. [19] P. E. Green, A. M. Krieger, and Y. J. Wind, Thirty years of conjoint analysis: Reflections and prospects, Interfaces, pp. 56-73, Mai/Jun 2001. [20] C. M. Jarque, and A. K. Bera, Efficient tests for normality, homoscedasticity and serial independence of regression residual, Economic Letters, vol. 24, no. 3, pp. 255-259, 1980.