CHAPTER 5 REFINING KANO’S MODEL
Moreover, through the Kano’s model refined by IPA, this study could use the benefits of integrating the Kano’s model and IPA to resolve the related issues surrounding the traditional Kano’s model and IPA.
Section 2 Methodology
This study uses the classification of quality attributes in the revised analytical Kano model as a basis, and the determination method of Matzler, et al. (2004a) on the category of quality attributes is used to integrate the concepts of importance and performance into the classification of quality attributes. Simultaneously, regression analysis with dummy variables is used to estimate the impact of attribute performance on overall satisfaction to substitute the indicators of the vertical axis (SII) and the horizontal axis (DDI) of the revised analytical Kano model. Then, the definition of the eight categories of quality attributes of the refined Kano’s model (Yang, 2005) are used as a reference to further categorize the quality attributes to obtain valuable information for improvements. Finally, the results of the categorizing the quality attributes, the attribute performance, and the asymmetric impact of attribute performance on overall satisfaction are combined to suggest the priority of improvements. The classification of quality attributes of the refined Kano’s model using IPA is shown in Figure 18.
Figure18 The Kano’s model refined by IPA
Highly Attractive
High value-added
Critical Indifferent
0 1
1
I
−I
+ LessAttractive
Low value-added
Necessary
0 I I
Iv
In Figure 18, the I+ refers to the impact on overall satisfaction when attribute performance is high, and theI−denotes the impact on overall satisfaction when attribute performance is low. The asymmetric impact of attribute performance on overall satisfaction can be represented as a vectorIv
, and the magnitude of the vector denotes the overall impact of the quality attribute on customer satisfaction. Therefore, the magnitude of the vector Iv
can be considered the importance index.I0is an indifference threshold that is used to differentiate the important quality attributes from the less important ones. If the radius is smaller thanI0, it is considered an indifferent attribute. I is the mean of the overall impact of quality attributes on customer satisfaction. According to the quality attribute categories redefined by Yang (2005), usingI can enable the revised analytical Kano model divided into three categories (namely, attractive, one-dimensional, and must-be) to be divided into six categories (namely, highly attractive and less attractive, high value-added and low value-added, and critical and necessary). Indifferent represents an attribute whose presence or absence does not cause satisfaction or dissatisfaction to customers; therefore, this study does not follow Yang (2005) for further categorization.
Subsequently, the Kano’s model refined by IPA comprises seven categories of quality attributes: highly attractive and less attractive, high value-added and low value-added, critical and necessary, and indifferent. The methodology steps of the refined Kano’s model using IPA are as follows:
Step1 Design a performance questionnaire. Performance with single quality elements as well as overall satisfaction with the service is evaluated using a 10-point scale, and each question is closed.
Step2 Collect data through questionnaires and acquire theI+ and I− of each quality element using regression analysis with dummy variables. One set of dummy variables is created and used to quantify attractive quality attributes, and another set is created to quantify must-be quality attributes. Must-be quality attributes and attractive quality attributes are expressed in scale units of the dependent variable (overall satisfaction). To conduct the analysis, attribute satisfaction ratings should be recoded. Performance ratings are recoded to form the dummy variables; “low performance” is coded (0,1), “high performance” (1,0), and “average performance”
(0,0). Based on this coding scheme, multiple regression analysis is conducted.
ε
β + + +
= 0 I+dummy1 I−dummy2
y i i (36) where y = Overall satisfaction
dummy1=Dummy set indicating the highest performance level dummy2 =Dummy set indicating the lowest performance level
+
Ii = Impact of the ith quality attribute on the overall satisfaction associated with a high performance level (i = 1, 2,…, n)
−
Ii = Impact of the ith quality attribute on the overall satisfaction associated with a low performance level (i = 1, 2,…, n)
ε = Random error Step3 Calculate the magnitude of Ivi
. The overall impact of the ith quality attribute on customer satisfaction can be represented as a vector Ivi
, and the magnitude of Ivi
is Ivi
. The formula is as follows:
2
2 −
+ +
= i i
i I I
Iv
, 0≤ Ivi ≤ 2 (37) Step4 Calculate theI and theI0.I is the mean of the overall impact of quality attributes on customer satisfaction.I0is an indifference threshold that is used to differentiate the important quality attributes from the less important ones. This study believes that if the Ivi
is smaller than the average of theI+and I−, the ith quality attribute is not significant. The formulae are as follows:
∑
==
n
i
Ii
I n
1
1 v
(38)
∑
=− + +
=
n
i
i
i I
n I I
1
0 2
1
(39) Step5 CalculateP and plot the refined Kano’s model using IPA. P is the mean of the
performance of quality attributes. Through theP , the degrees of performance are classified into two categories, “high” performance if the degree of performance is greater than the P , and “low” if below theP . According to theI+,I−,I0,I ,
andP , plot the refined Kano’s model using IPA.
Step6 Summarize and categorize the results according to the refined Kano’s model using IPA.
Section 3 Practical example
The subject for this case study is the mobile telecommunication company CT, one of the three largest mobile telecommunication companies in Taiwan. This study refers to the definition by Lee, et al. (2010) of the telecommunication industry’s service quality elements the basis for Step1 -- design a standardized questionnaire to measure the attribute performance and overall satisfaction. A scale from 1 (extremely low) to 10 (extremely high) was used. The questionnaire used in this study comprised 13 questions regarding service quality elements and 1 question regarding overall satisfaction of the telecommunications company, as shown in Table 9.
The questionnaire was distributed to 1800 customers randomly, and 203 valid questionnaires were returned. According to Step 2, the results of the questionnaire were calculated. Attribute importance was measured using multiple regression analysis, with overall satisfaction as the dependent variable, and attribute performance as the independent variable; the results are shown in Table 12 (see the second column of Table 12). Then, using regression analysis with dummy variables, this study estimated the asymmetric impact of attribute performance on overall satisfaction. For each variable, two regression coefficients were obtained, one to measure the impact when performance is low, the other when performance is high. The results are shown in Table 13.
According to Steps 3 and 4, the Ivi
,I andI0were calculated; the results are shown in Table 13. According to Step 5, attribute performance is classified into two categories (see the fourth column of Table 12). Based on the implicit importance and attribute performance, the IPA strategy is determined (see the fifth column of Table 12). The results of Tables 12and 13depict the refined Kano’s model using IPA, as shown in Figure 19.
Additionally, the quality attribute categories are listed in the sixth column of Table 12.
Table12
Summarized results of service quality elements in telecommunications
Item Implicit importance
Attribute
performance Category in performance
IPA strategy
Category in the Kano’s model refined by IPA
Q1 .022(ns) 7.374 High P High value-added
Q2 .187*** 7.172 High K High value-added
Q3 .021(ns) 6.631 Low L Critical
Q4 .180*** 7.158 High K Necessary
Q5 .090** 6.483 Low L Low value-added
Q6 .069* 7.167 High P Critical
Q7 .025(ns) 7.956 High P Low value-added
Q8 .068* 7.399 High P Low value-added
Q9 .296*** 6.217 Low C High value-added
Q10 .116** 6.394 Low C High value-added
Q11 .143*** 5.980 Low C Low value-added
Q12 .089*** 4.483 Low L Less attractive
Q13 .139*** 6.212 Low C Low value-added
mean .111 P = 6.664
Notes1: Implicit importanceR2=.918. ns = not significant
* p<.10 ** p<.05 *** p<.001
Notes2: K: Keep Up the Good Work, P: Possible Overkill, L: Low Priority, and C: Concentrate Here
Table13
Dummy variable regression results
Dummy variable regression coefficients Item
Low performanceI− High performanceI+
Ivi
Q1 -.585*** .350*** .682 Q2 -.474*** .415*** .630 Q3 -.627*** .257*** .678 Q4 -.509*** .144** .529 Q5 -.393*** .437*** .588 Q6 -.704*** .141** .718 Q7 -.387*** .445*** .590 Q8 -.460*** .262*** .529 Q9 -.411*** .449*** .609 Q10 -.499*** .390*** .633 Q11 -.419*** .373*** .561 Q12 -.155** .484*** .508 Q13 -.493*** .253*** .554
I0= .404 I = .601
** p<.05 *** p<.001
Section 4 Discussion
Regarding analysis based on Tables 12 and 13 and Figure 19, the refined Kano’s model using IPA and compared with an IPA model are described below. From the results shown in Table 13 and Figure 19, Q3: professional knowledge of the service center clerk, Q4: online service provided by the telecommunications company, and Q6: quality of the mobile phone signal are must-be quality attributes. Q4 is the quality element with a smaller Ivi
and I0(0.404)< Ivi <I(0.601)
, thus Q4 is considered a necessary quality attribute. Q3 and Q6 are the quality elements with a larger Ivi
and Ivi ≥I(0.601)
, thus Q3 and Q6 are considered critical quality attributes. Q1: successful connection to service center, Q2: attitude of the service center clerk, Q5: service provided in the shop, Q7:
correctness of the mobile phone bill, Q8: e-bill service, Q9: premium programs for mobile
0 0.2 0.4 0.6
0 0.2 0.4 0.6
I
−I+
Q6 Q1
•
Q3Q4 Q13
•
Q8
•
Q10 Q2 Q11•
•
Q12
•
Q9 Q7Q5•
Figure19 The display of the refined Kano’s model using IPA
Notes: the symbol • denotes the low performance the symbol о denotes the high performance
phones, Q10: value-added service for mobile phones, Q11: internet connection service for mobile phones, and Q13: service rate for mobile phones are one-dimensional quality attributes. Q5, Q7, Q8, Q11, and Q13 are the quality elements with a smaller Ivi
and )
601 . 0 ( )
404 . 0
0( I I
I < vi <
; thus, Q5, Q7, Q8, Q11, and Q13 are considered low value-added quality attributes. Q1, Q2, Q9, and Q10 are the quality elements with a larger Ivi
and Ivi ≥I(0.601)
; thus, Q1, Q2, Q9, and Q10 are considered high value-added quality attributes. Q12: 3G visual service for mobile phones is an attractive quality attribute with a I0(0.404)< Ivi <I(0.601)
; thus, Q12 is considered a less attractive quality attribute.
By contrast, regarding performance, Q1: successful connection to service center, Q2:
attitude of the service center clerk, Q4: online service provided by the telecommunications company, Q6: quality of the mobile phone signal, Q7: correctness of the mobile phone bill, and Q8: e-bill service are the quality elements with a performance greater than theP(6.664); thus, Q1, Q2, Q4, Q6, Q7, and Q8 are considered high performance. Q3:
professional knowledge of the service center clerk, Q5: service provided in the shop, Q9:
premium programs for mobile phones, Q10: value-added service for mobile phones, Q11:
internet connection service for mobile phones, Q12: 3G visual service for mobile phones, and Q13: service rate for mobile phones are the quality elements with a performance below theP ; thus, Q3, Q5, Q9, Q10, Q11, Q12, and Q13 are considered low performance and should be improved. Q3, Q9, and Q10 are the quality elements that are more important from a customer perspective. Thus, according to Yang (2005), Q3, Q9, and Q10 should be prioritized for improvement, followed by Q5, Q11, Q12, and Q13.
According to IPA strategy, Q2: attitude of the service center clerk, and Q4: online service provided by the telecommunications company, are classified as “Keep Up the Good Work”. Q1: successful connection to service center, Q6: quality of the mobile phone signal, Q7: correctness of the mobile phone bill, and Q8: e-bill service, are classified as “Possible Overkill”. Q3: professional knowledge of the service center clerk, Q5: service provided in the shop, and Q12: 3G visual service for mobile phones are classified as “Low Priority”.
Q9: premium programs for mobile phones, Q10: value-added service for mobile phones, Q11: internet connection service for mobile phones, and Q13: service rate for mobile phones are classified as “Concentrate Here”. Therefore, Q9, Q10, Q11, and Q13 should be
listed first on the improvement list, followed by Q3, Q5, and Q12.
A comparison of the quality elements in the IPA strategy and the Kano’s model refined by IPA reveals that the results of the three elements differed. Q11: internet connection service for mobile phones, for instance, the result of the IPA strategy is prioritized improvement, whereas the result of the Kano’s model refined by IPA is secondary priority. Developing Internet connection services for mobile phones is indeed a growing trend, but current basic infrastructure in Taiwan is still inadequate, meaning that mobile Internet connections cannot yet be provided throughout the entire country.
Therefore, the improvement of the quality element must be considered under mid- to long-term planning. Because consumers who use mobile Internet are still comparatively in the minority, considering other quality elements that must be improved, it is reasonable that Q11 is regarded as an improvement of secondary priority. Q13: service rate for mobile phones, the result of the IPA strategy is prioritized improvement, whereas the result of the Kano’s model refined by IPA is that it is a secondary improvement priority. For consumers, the lower the service rate is the better. However, one must consider the company’s operating costs and price competition with other telecommunication companies; thus, advantages gained by improvements are limited. Each company, however, is constrained by limitations on the resources they have available. Therefore, how scarce resources are best deployed must be decided to achieve the highest level of satisfaction. Therefore, it is more practical to consider Q13 a secondary improvement priority. Q3: professional knowledge of the service center clerk, the result of the IPA strategy is that it is a secondary improvement priority, whereas the result of the Kano’s model refined by IPA prioritized its improvement. The professional knowledge of the service center clerk forms the foundation of customer service. Personnel training should be the core concern of the company’s human resource management. In addition, the improvements to the quality element are the responsibility of the company’s administration department, meaning that more direct control would produce better results. Hence, Q3 should be a prioritized improvement.
Finally, using Table 13 and Figure 19, the order for improving the quality elements can be identified. According to Berger, et al. (1993), the impact of the quality attribute category on the products or services is: M (Must-be) ＞ O (One-dimensional) ＞ A (Attractive)＞I (Indifferent). Thus, Q3: professional knowledge of the service center clerk, must be improved first, followed by Q5: service provided in the shop, Q9: premium programs for mobile phones, Q10: value-added service for mobile phones, Q11: internet
connection service for mobile phones, and Q13: service rate for mobile phones. The final improvement is Q12: 3G visual service for mobile phones. Additionally, using the judgment of Ivi
, we can determine that Iv10 > Iv9 > Iv5 > Iv11 > Iv13 ; thus, the order for improvement is Q10 ≻ Q9 ≻ Q5 ≻ Q11 ≻ Q13. Therefore, the order for improvement of the quality elements is Q3 ≻ Q10 ≻ Q9 ≻ Q5 ≻ Q11 ≻ Q13 ≻ Q12.