Integrating the Kano model into a robust design approach to enhance
customer satisfaction with product design
Chun-Chih Chen
, Ming-Chuen Chuang
Institute of Applied Arts, National Chiao Tung University, Hsinchu, Taiwan, ROC
a r t i c l e
i n f o
Article history:
Received 1 September 2006 Accepted 1 February 2008 Available online 1 April 2008 Keywords: Customer satisfaction Kano model Robust design Multiple-criteria optimization Kansei engineering Product design
a b s t r a c t
The aesthetic qualities of products are critical factors in achieving higher customer satisfaction. This study presents a robust design approach incorporating the Kano model to obtain the optimal combination of design form elements. This can effectively enhance customer satisfaction and aesthetic product qualities with multiple-criteria character-istics. The Kano model is used to better understand the relationship between performance criteria and customer satisfaction, and to resolve trade-off dilemma in multiple-criteria optimization by identifying the key criteria in customer satisfaction. The robust design approach combines grey relational analysis with the Taguchi method to optimize subjective quality with multiple-criteria characteristics. This simultaneously yields the optimal aesthetic performance and reduces the variations in customer evaluations. Based on Kano model analysis, a weight adjustment process determines the weight of each product criterion for achieving the desired customer satisfaction performance. This process guides the prioritizing of multiple criteria, leading to higher customer satisfaction. A mobile phone design experiment was conducted to verify the benefits of using the proposed integrative approach. Results show that the generated optimal mobile phone design can effectively enhance overall aesthetic performance and customer satisfaction. Although mobile phone designs are the examples of this study, the proposed method may be further used as a universal robust design approach for enhancing customer satisfaction and product quality with multiple-criteria characteristics.
&2008 Elsevier B.V. All rights reserved.
1. Introduction
Customer satisfaction is the major concern and prerequisite for competitiveness in today’s global market. Because of market equivalence in product quality, the subjective quality of aesthetics is a critical determinant of customer satisfaction. For example, Apple’s iMac was heralded as an ‘‘aesthetic revolution in computing’’. This indicates that the visual aesthetics of computers have become a factor in customer purchase decisions (Postrel,
2001). Related studies also concluded that the aesthetic quality of a design has a positive effect on customer satisfaction (Fynes and Bu´rca, 2005; Yamamoto and Lambert, 1994). Aesthetic design can enhance the desir-ability of a product and greatly influence customer satisfaction in terms of perceived product quality (Bloch, 1995). However, the relationship between subjective quality and customer satisfaction is seldom discussed (You et al., 2006; Yun et al., 2003). This study regards aesthetics as an aspect of quality and explores the impact of aesthetics on customer satisfaction.
Scientifically and efficiently enhancing the aesthetic quality of product design can be achieved by gauging customer responses to product aesthetics and correlating these perceptions to form elements. This enables re-searchers to modify designs and closer align them with Contents lists available atScienceDirect
journal homepage:www.elsevier.com/locate/ijpe
Int. J. Production Economics
0925-5273/$ - see front matter & 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpe.2008.02.015
Corresponding author. Department of Industrial Design, National Kaohsiung Normal University, No. 62, Shenzhong Road, Yanchao Shiang, Kaohsiung County, Taiwan, ROC. Tel.: +886 7 7172930x7813;
fax: +886 7 6051156.
customer needs (Coates, 2003). The customer-oriented Kansei engineering (Nagamachi, 2002) method is a tool for translating customer perceptions and feelings (Kansei in Japanese) into concrete form elements. This method has been successfully used to infer optimal product design (Chuang and Ma, 2001; Lai et al., 2006; Schu¨tte and Eklund, 2005;You et al., 2006). Previous studies on Kansei engineering and aesthetics used questionnaire-collected data to examine customer subjective evaluations based on a mean scale rating. However, the evaluation of aesthetics is subjective and highly individualistic. Aesthetics evalua-tion based solely on mean scale ratings, without con-sidering variation in customer evaluations, is not appropriate. Lai et al. (2005) presented a robust design approach to enhance quality perception by reducing the discrepancy between the actual customer feeling and the desired feeling and reducing ambiguity created by the highly individualized characteristics of the customers. The robust design approach focuses on bringing the mean closer to the desired target and simultaneously reducing quality variation. This design may be successfully used in subjective quality management.
Aesthetic experience has a multidimensional nature. Previous studies (Lavie and Tractinsky, 2004; Liu, 2003, Rashid et al., 2004; Schenkman and Jonsson, 2000) that used a one-dimensional construct (e.g., a semantic index ‘‘beautiful verse ugly’’ or a single aesthetic measure with Likert scale rating) to explain how users perceived subject quality are not appropriate. Optimizing aesthetic quality should be considered a multiple-criteria problem. Thus, multiple-criteria decision making is required. Usually these criteria are not equivalent, i.e., they make different contributions to the integral quality assessment. Some criteria are even competitive, i.e., an improvement in one criterion will inevitably lead to deterioration in another (Dimova et al., 2006; Chen et al., 2006). However, most studies (Bottani and Rizzi, 2008; Partovi, 2007; Wang et al., 2007) on multiple-criteria optimization employed weight determination methods to reflect how customers prioritize their wants without considering these features. The relationship between product criteria and customer satisfaction has mostly been assumed to be linear—the higher the perceived criteria quality, the higher the customer’s satisfaction and vice versa. However, from the viewpoint of current theory, this relationship may be non-linear. Continuous improvement in some criteria, without considering what customers actually desire, may not be sufficient to enhance satisfaction. Conventional weight determination methods may not be able to completely illustrate the relationship between quality criteria and customer satisfaction levels. Understanding the relationship between certain quality criteria and customer satisfaction is necessary to decide which criteria to offer.Kano et al. (1984)developed a two-dimensional (linear and non-linear) quality model to address linear quality model shortcomings. This two-dimensional model divides quality criteria into must-be quality, one-dimen-sional quality and attractive quality. These terms describe a product’s effect on customer satisfaction with or with-out a specific quality. The Kano model is an effective tool for categorizing product criteria and product
require-ments. Based on the Kano classification, the criterion with the greatest influence on customer satisfaction, i.e., the attractive quality, should be offered if two criteria cannot be promoted simultaneously due to technical or financial reasons. This method provides valuable guidance in trade-off situations during the product development stage (Conklin et al., 2004; Huiskonen and Pirttila¨, 1998; Matzler and Hinterhuber, 1998). Accordingly, a design team can determine which areas should be targeted to produce maximum benefits in customer satisfaction. This study investigates the possible integration between a robust design approach and the Kano model for achieving higher customer satisfaction and the effectively optimiz-ing multiple criteria.
Another purpose of this study is to explore aesthetic criteria characteristics and apply the Kano model to investigate the different impacts of criteria quality on customer satisfaction. An integrative method combining the Kano model with the robust design approach is proposed to enhance the subjective quality of aesthetics and customer satisfaction. The robust design approach combines grey relational analysis (GRA) and the Taguchi method (TM) into a grey-based TM (Lin and Lin, 2002; Tarng et al., 2002;Wang and Tong, 2004). We adopted this method to explore the relationship between design para-meters and quality performance with multiple-criteria considerations. It also determines the optimal combination of design parameters to maximize quality performance and minimize quality variation. We adopted the Kano model to explore the relationship between multiple aesthetic criteria and customer satisfaction, and to identify the key factors that enhance satisfaction. The Kano classification results determined which aesthetic criteria should be emphasized to achieve higher satisfaction and optimize trade-offs between multiple criteria. Each criterion’s effect on customer satisfaction was considered in the grey-based TM to effectively optimize aesthetic quality and customer satisfaction. We conducted an experimental study on mobile phone design to illustrate how the Kano model can be integrated into the robust design approach and to verify the effectiveness of the proposed method.
2. Theoretical background
2.1. Robust design approach for multiple-criteria optimization
Robust design is a quality improvement engineering method that seeks the lowest cost solution to product design specifications based on customer requirements. The TM is the conventional approach to achieve robust-ness (Cabrera-Rios et al., 2002). The primary tools of the TM are orthogonal arrays (OAs) and the signal-to-noise (S/N) ratio. The former substantially reduces the number of required experiments and the latter simultaneously finds the most robust combination and the best possible performance (Taguchi and Clausing, 1990). The TM defines a loss function to calculate the deviation between the experimental value and the desired value. The value of the loss function is further transformed into a S/N ratio. S/N
ratio analysis usually considers three performance char-acteristic categories: lower-the-better, higher-the-better and nominal-the-better. The S/N ratio for each design parameter level is computed based on S/N analysis. Regardless of the performance characteristic category, a larger S/N ratio corresponds to a better performance characteristic. Therefore, the optimal design parameter level is the level with the highest S/N ratio. In most industrial applications, the TM is used to solve problems with a single performance characteristic. In the real world, however, most products require more than one quality characteristic to be considered simultaneously, i.e., most of the problems that customers encounter involve multi-ple criteria. Optimizing multimulti-ple performance character-istics is much more complicated than optimizing a single performance characteristic (Korpela et al., 2007; Nearch-ou, 2006). A design with a higher S/N ratio for one performance characteristic may produce a lower S/N ratio for another performance characteristic. As a result, an overall evaluation of S/N ratios is required to optimize multiple performance characteristics. This study adopts a grey-based TM (Lin and Lin, 2002;Wang and Tong, 2004), which combines GRA with the TM, to solve this problem. The grey system theory developed byDeng (1982)is an effective mathematical means to deal with system analysis characterized by incomplete information. The GRA method, in grey system theory, measures the relationship between factors based on their degree of similarity in development trends (Deng, 1982). More precisely, during the system development process, if the trend for the change between two factors is consistent, it produces a higher grey relational grade (GRG). The GRA method can effectively solve complicated inter-relation-ships between multiple performance characteristics.
The GRA calculation process in the grey-based TM is briefly reviewed. In GRA, data preprocessing is first performed to normalize the raw data for analysis. A linear normalization of the S/N ratio is performed in the range between zero and unity, which is also called grey relational generating (Deng, 1989). The data are trans-formed into the normalized data in the following three situations (Lin and Lin, 2002):
1. Measuring the effectiveness of the lower-better xij¼
maxjZijZij
maxjZijminjZij
(1)
2. Measuring the effectiveness of the higher-better xij¼
ZijminjZij
maxjZijminjZij
(2)
3. Measuring the effectiveness of the nominal-better xij¼
jZijZobj
maxfmaxjZijZob; ZobminjZijg
(3) where Zijis the S/N ratio for the performance of the
ith criterion in the jth experiment (design) and xijis the
normalized S/N ratio; Zob is the target value and
minjZijpZobpmaxjZij.
The grey relational coefficient is calculated from the normalized S/N ratio to express the relationship between the desired and the actual normalized S/N ratio. The grey relational coefficient xijfor the ith performance
character-istic in the jth experiment (design) can be expressed as xij¼
miniminjjx0i xijj þzmaximaxjjx0i xijj
jx0
i xijj þzmaximaxjjx0i xijj
(4) where x0
i is the ideal normalized S/N ratio for the ith
performance characteristic and zA[0,1] is a distinguishing coefficient for controlling the resolution scale, usually assigned with the value of 0.5.
A weighting method is then used to integrate the grey relational coefficients of each criterion into the GRG to reflect the importance of each criterion. Many methods can be used to determine weights (Hwang and Yoon, 1989), including the eigenvector method, entropy method, analytic hierarchy process, etc. The overall evaluation of the S/N ratio of multiple performance characteristics (i.e., the GRG) is based on the following equation:
gj¼
Xn i¼1
Wixij (5)
where gj is the GRG for the jth experiment, Wi is the
weight for the ith criterion, i ¼ 1,2,y,n, n is the number of criteria and wi¼[0,1] satisfiesPni¼1Wi¼1.
As a result, optimizing complicated multiple perfor-mance characteristics merely requires optimizing a single GRG. The optimal combination of design parameters is the one with the highest GRG. Furthermore, a statistical analysis of variance (ANOVA) or a quantitative theory type I analysis can specify which design attributes are statistically significant in affecting the GRG. With GRA, ANOVA or quantitative theory type I analysis, the optimal combination of design attributes can be predicted. The TM combined with GRA can greatly simplify optimization procedures for determining optimal parameters for multi-ple performance characteristics.
To resolve the complicated problem of multiple-criteria optimization, a weighting method is usually used to determine the importance of each criterion on affecting perceived quality in the grey-based TM. However, criteria weights only reflect how customers prioritize their wants, and cannot illustrate the relationship between customer satisfaction and product criteria. To effectively achieve the desired customer satisfaction, a design team should know what the customer wants most and also understand how much attention to pay to each product criterion. The authors of this study propose a process model to adjust criteria weights by incorporating the Kano model to achieve the highest customer satisfaction.
2.2. Kano model of customer satisfaction
Customers evaluate the quality of a product using several factors and dimensions. Therefore it is important to identify which product criteria or attributes create more satisfaction than others. Kano et al. (1984) developed a two-dimensional model to explain the different relation-ship between customer satisfaction and product criterion
performance. The Kano model classifies product criteria into three distinct categories, as shown in Fig. 1. Each quality category affects customers in a different way. The three different types of qualities are explained as follows:
1. The must-be or basic quality: Here, customers become dissatisfied when the performance of this product criterion is low or the product attribute is absent. However, customer satisfaction does not rise above neutral with a high-performance product criterion. 2. One-dimensional or performance quality: Here, customer
satisfaction is a linear function of a product criterion performance. High attribute performance leads to high customer satisfaction and vice versa.
3. The attractive or excitement quality: Here, customer satisfaction increases superlinearly with increasing attribute performance. There is not, however, a corresponding decrease in customer satisfaction with a decrease in criterion performance.
Besides these three, two more quality types can be identified: the indifference and reversal qualities (to be precise, they should call them characteristics because they are not really a customer need). For the indifference quality, customer satisfaction will not be affected by the performance of a product criterion. For the reversal quality, customers will be more dissatisfied with the increase of a criterion performance.
A simple way of identifying different Kano categories, one-dimensional, attractive and must-be qualities, is to use a Kano questionnaire (Kano et al., 1984). In this questionnaire, customers indicate if they feel satisfied or dissatisfied with a given situation. First, a situation supposes the quality (criterion) is present or sufficient. The customer must choose one of the following answers to express his feelings:
a. Satisfied
b. It should be that way c. I am indifferent d. I can live with it e. Dissatisfied.
A second situation assumes the quality is absent or insufficient. Again, the customer must choose one of the above-mentioned feeling responses. By combining the two answers in the Kano evaluation table (Table 1), the product criterion can be identified as attractive, must-be, one-dimensional, indifference or reversal.
Matzler and Hinterhuber (1998) showed that the advantage of using the Kano model to classify customer requirements is a better understating of product require-ments. This permits designers to focus on priorities for product development. It is, for example, not very useful to invest in improving must-be qualities that have already reached a satisfactory level. It would be better to improve one-dimensional or attractive qualities because they have a greater impact on perceived quality, and consequently, on customer satisfaction. Kano classification also provides valuable help in case of a trade-off situation in multiple-criteria decision making. If two product multiple-criteria cannot be promoted simultaneously due to technical or financial reasons, the criterion with the greater influence on customer satisfaction should be enhanced first. The Kano model can also be used to establish the importance (weight) of individual product criterion in multiple-criteria decision making, and thus create the optimal perquisite for product development activities. Tan and Pawitra (2001)presented an integrative approach invol-ving the Kano model and quality function deployment (QFD). The Kano model adjusts the improvement ratio for re-prioritizing attributes in the QFD method. The inte-grative approach provides a basis for deciding the relative priority of improving product attributes based on Kano categories. However, it cannot illustrate how to design and improve product quality to meet customer requirements. In the study, the Kano model is integrated into the robust design approach to infer the optimal design parameters for achieving the highest customer satisfaction.
3. A proposed integrative approach for customer satisfaction
An integrative approach is proposed to better under-stand the relationship between customer satisfaction and
product criteria, and to obtain the optimal design attribute combination of the multiple-criteria optimiza-tion in this study.Fig. 2outlines the proposed approach using the Kano model and the grey-based TM.
Multiple criteria affecting perceived product quality should first be identified. The conventional weighting method is used to obtain the raw weights of criteria. The Kano model differentiates how well the criteria are able to affect customer satisfaction. A Kano questionnaire helps categorize criteria related to consumer satisfaction into different types of qualities and indicates how much attention should be paid to each product criterion to achieve the desired customer satisfaction. Logically, the criteria in the attractive category should receive attention first. Criteria in the one-dimensional and must-be cate-gories should receive successively lower priorities. A weight adjustment method based on the Kano classifica-tion is used to re-prioritize the criteria in this study. The process of adjusting criteria weights is also shown inFig. 2. The first step of the weight adjustment process is to identify the proper Kano category for each criterion
related to consumer satisfaction based on the Kano questionnaire results. The purpose of this is to magnify the weights of higher-return criteria in increasing custo-mer satisfaction and resolving the trade-off situation of multiple-criteria optimization. Raw weights obtained in the conventional weighting method were then adjusted by multiplying with the adjustment coefficient (K) for each Kano category. Values of ‘‘4’’, ‘‘2’’, ‘‘1’’ and ‘‘0’’ were assigned to the attractive, one-dimensional, must-be and indifference categories, respectively. The weight adjust-ment can be expressed as
wi_adj¼
WiKi
Pn i¼1WiKi
(6) where wi_adj is the final adjusted weight for the ith
performance criterion, Wi is the raw weight for the
ith performance criterion, i ¼ 1,2, y, n, and Ki is the
adjustment coefficient according to its Kano quality classification.
The key difference between the conventional weight-ing method and the method based on the Kano model is
Table 1
Kano evaluation table
Product criteria/attributes Insufficiency
Satisfied It should be that way I am indifferent I can live with it Dissatisfied
Sufficiency Satisfied Q A A A O
It should be that way R I I I M
I am indifferent R I I I M
I can live with it R I I I M
Dissatisfied R R R R Q
A—attractive, O—one-dimensional, M—must-be, I—indifference, R—reversal, Q—questionable.
that the former represents the importance of customer requirements while the latter represents the importance of customer satisfaction. The final adjusted weights are then used as criteria priorities in the grey-based TM to better understand the customer needs and desires and effectively achieve customer satisfaction.
The grey-based TM was conducted to infer the optimization of design attributes. This includes the following steps:
Step 1: Identify the design attributes and setting levels for the Taguchi experiment design.
Step 2: Select an appropriate Taguchi’s OA and assign the design attribute parameters to the OA. Then generate experimental samples based on the OA.
Step 3: Conduct the evaluation experiment for each sample on the identified criteria.
Step 4: From the experimental data, calculate the S/N ratio for each criterion performance.
Step 5: Perform GRA by combining the final adjusted weights based on the Kano model.
Step 6: Analyse the experimental results using the GRG and statistical analysis method.
Step 7: Select the optimal level of design attributes to obtain the optimization design and identify the significant attributes.
Step 8: Perform a verification experiment to confirm the design.
4. The proposed integrative method to optimize the aesthetics satisfaction of mobile phone design
This section presents a mobile phone case example to illustrate how the proposed approach can be used to optimize aesthetics satisfaction robustly.
4.1. Determining the aesthetic criteria of product design Customer satisfaction with product aesthetics involves multiple criteria. It is important to identify the most important and representative aesthetic criteria to ensure efficiency. An appropriate set of criteria for aesthetics satisfaction was first collected through literature reviews (Liu, 2003; Rashid et al., 2004; Schenkman and Jonsson, 2000;Talia and Noam, 2004). Six experts, senior designers with an average design experience of more than 10 years in the product design field, participated in focus groups (Nielsen, 1993) to identify the proper aesthetic criteria of product design. These experts identified aesthetic criteria including originality, unity, completeness, pleasure, sim-plicity and satisfaction of form.
4.2. Determining the design attributes of a mobile phone The product form attributes that elicit customers’ aesthetic perception were defined as the control factors in the current Taguchi experiments. Related literature on Kansei engineering for mobile phone design (Chuang et al., 2001; Lai et al., 2005; Yun et al., 2003) was first reviewed to identify the appropriate design attributes of mobile phones. Then, existing mobile phones were
collected to identify commonly used important design attributes. The six experts were then asked to review the information and extract design attributes using morpho-logical analysis (Zwick, 1967). The design attributes that were thought most likely to influence aesthetics were identified. The identified design attributes were grouped into two types: form features and feature composition, and the compositional relationship among these features. Table 2presents the 10 identified design attributes and their corresponding level setting. Design attributes A, B, C, D, E, F, G describe the form features of the mobile phone designs; design attributes H, I, J describe the feature composition. Related studies (Shao et al., 2000;Wu and Chuang, 2003) suggest that the screen, function button style and speaker receiver were not significant factors in aesthetics evaluation. These features were kept fixed and excluded from the design attributes setting.
4.3. Design of Taguchi experiments
The control factor array chosen for this case study had to accommodate 10 control factors (A–J in Table 2), including nine factors with four levels and one with two levels. The full factorial of this combination would have required up to 524,288 (2 49) samples. The use of an OA
can effectively reduce the number of experiments neces-sary to determine the optimal design attributes combina-tion in a product design. The experiment layout using a Taguchi’s L032 (2149) OA, as shown inTable 3, was used
to design the Taguchi experiment in this study. The data of each experimental run in the orthogonal table were then converted to a computer image of mobile phone design for aesthetic criteria evaluation. The 32 experimental samples are shown inFig. 3.
4.4. Experimentation
A total of 35 male and 25 female subjects ranging from age 18 to 24 were recruited for the evaluation experiment. The subjects were asked to evaluate the 32 experimental samples on each of the six aesthetic criteria with a 7-point Likert scale (from 1 ¼ not at all to 7 ¼ intensely so). A Kano questionnaire survey was then conducted to classify the aesthetic criteria into the Kano categories. Table 4shows the experimental results, the mean and the standard deviation (S.) of the evaluation of each sample on the six aesthetic criteria. The higher-the-better perfor-mance characteristic was assumed for these aesthetic criteria. The higher-the-better S/N ratio of each criterion for each sample was calculated for each criterion evalua-tion basis, i.e.,
S=N ðHigher-the-betterÞZij¼ 10 log 1 n Xn i¼1 1 y2 ijk ! ðdBÞ (7) where Zijis the S/N ratio of the ith performance criterion
in the jth experiment (sample), n is the total number of subjects and yijk is the evaluation value of the ith
criteria in the jth sample by the kth subject.Table 4shows the S/N ratios.
4.5. The Kano classification
Table 5 shows the results of the Kano questionnaire, which provided information for classifying criteria. The criteria of originality, pleasure and satisfaction of form can be considered attractive requirements; completeness can be considered a one-dimensional quality. On the other hand, simplicity can be classified as must-be quality that customers take for granted. Unity is not related to customer satisfaction, and it was classified as an indifference quality. As mentioned before, efforts should be directed toward the
attractive and one-dimensional criteria. The Kano classifica-tion corresponding to each criterion was then integrated to adjust the criteria weights by multiplying the adjustment coefficient (K) with each Kano category in the next step.
4.6. Determining adjusted criteria weights using the Kano model
Assigning criteria weights is usually based on expert opinions, and may cause a subjective bias. This study
Table 2
adopted the entropy weighting method (Zeleny, 1982), which rests on the bases of the criteria evaluating only, to objectively determine the raw weights of aesthetic criteria. The relative entropies of criteria regarded as a measurement of structural similarity determine the relative importance, i.e. the criteria weights. The criteria weight processing steps can be summarized as follows:
Step 1: Allow dijto be the S/N ratios for evaluation of
the ith criteria in the jth experiment samples, i ¼ 1,2, y, n; j ¼ 1,2, y, m, then D ¼ ðdijÞnm is the sample evaluation
matrix. Assume that zijis the converted value via dij; it can
be defined as zij¼ dij Pm j¼1dij (8) Step 2: The entropy of the ith criteria eican be measured
using the following equation, where c ¼ 1/ln(m) and ei40:
ei¼ c
Xm j¼1
zijln zij (9)
Step 3: Compute the objective weight of the criteria, where E ¼Pni¼1ei and the weight of ith criterion can be
calculated as
wi¼ ð1 eiÞ=ðn EÞ (10)
Since unity was identified as an indifference quality based on the Kano classification results, it was excluded from this case for economic and efficiency considerations. The S/N ratios of the other five aesthetic criteria for each experimental sample formed the raw data for the entropy weighting calculation. The raw criteria weights were obtained through data processing using Eqs. (8)–(10). The Kano model was then integrated to adjust each criterion weight according to its Kano category. The final adjusted weights for aesthetic criteria, explaining prioritization related to customer satisfaction, were computed using Eq. (6). Table 6 shows results of the adjusted weight incorporating with the Kano model. These results imply that attractive qualities should be emphasized, such as originality, pleasure and satisfaction of form. Simplicity, which was categorized as a must-be quality, should be given a lower priority. The entropy weighting method, on the other hand, indicated that simplicity should be prioritized first and originality last.
4.7. Grey relation analysis
In GRA, data preprocessing was first performed using Eq. (2) to normalize the S/N ratios of the subjects’ criteria evaluation. Next, the grey relational coefficient was
Table 3
Experimental layout using L320(2149) OA
Experiment no. Level of form attribute
A B C D E F G H I J 1 1 1 1 1 1 1 1 1 1 1 2 1 2 2 2 2 2 2 2 2 1 3 1 3 3 3 3 3 3 3 3 1 4 1 4 4 4 4 4 4 4 4 1 5 2 1 1 4 2 2 3 3 4 1 6 2 2 2 3 1 1 4 4 3 1 7 2 3 3 2 4 4 1 1 2 1 8 2 4 4 1 3 3 2 2 1 1 9 3 1 2 4 3 4 2 1 3 1 10 3 2 1 3 4 3 1 2 4 1 11 3 3 4 2 1 2 4 3 1 1 12 3 4 3 1 2 1 3 4 2 1 13 4 1 2 1 4 3 4 3 2 1 14 4 2 1 2 3 4 3 4 1 1 15 4 3 4 3 2 1 2 1 4 1 16 4 4 3 4 1 2 1 2 3 1 17 1 1 4 3 1 4 3 2 2 2 18 1 2 3 4 2 3 4 1 1 2 19 1 3 2 1 3 2 1 4 4 2 20 1 4 1 2 4 1 2 3 3 2 21 2 1 4 2 2 3 1 4 3 2 22 2 2 3 1 1 4 2 3 4 2 23 2 3 2 4 4 1 3 2 1 2 24 2 4 1 3 3 2 4 1 2 2 25 3 1 3 2 3 1 4 2 4 2 26 3 2 4 1 4 2 3 1 3 2 27 3 3 1 4 1 3 2 4 2 2 28 3 4 2 3 2 4 1 3 1 2 29 4 1 3 3 4 2 2 4 1 2 30 4 2 4 4 3 1 1 3 2 2 31 4 3 1 1 2 4 4 2 3 2 32 4 4 2 2 1 3 3 1 4 2
1 2 3 4 5 6 7 8
9 10 11 12 13 14 15 16
17 18 19 20 21 22 23 24
25 26 27 28 29 30 31 32
Fig. 3. Thirty-two mobile phones used in the current Taguchi experiment.
Table 4
Result of criteria evaluation in Taguchi experiments
No. Simplicity Originality Completeness Pleasure Sat. of form Unity
Mean S. S/N Mean S. S/N Mean S. S/N Mean S. S/N Mean S. S/N Mean S. S/N 1 4.98 1.49 10.60 2.63 1.20 4.76 3.73 1.56 7.31 2.98 1.42 5.62 2.92 1.43 5.35 3.52 1.67 6.65 2 3.88 0.93 10.97 3.35 1.25 7.10 3.65 1.20 8.69 3.04 1.26 6.30 2.77 1.21 5.49 3.40 1.30 7.05 3 4.27 0.91 11.68 3.94 1.34 8.56 3.88 1.25 9.08 3.54 1.47 7.14 3.42 1.40 6.73 3.94 1.33 9.03 4 2.88 1.03 6.98 4.25 1.39 9.55 3.10 1.26 7.21 2.44 1.21 4.46 2.29 1.14 4.01 2.98 1.27 6.41 5 3.71 1.47 8.15 4.25 1.59 8.03 2.92 1.22 5.52 2.48 1.24 4.36 2.50 1.29 4.24 2.79 1.22 5.53 6 4.17 1.11 10.42 3.81 1.24 8.53 3.60 1.15 8.72 3.29 1.27 7.20 3.15 1.31 6.65 3.65 1.28 7.55 7 3.29 0.98 8.74 4.17 1.33 10.09 3.54 1.17 8.51 3.21 1.49 5.91 2.94 1.48 5.32 3.46 1.22 7.56 8 4.25 0.88 11.86 3.90 1.31 8.94 3.98 1.20 8.83 3.42 1.44 6.70 3.42 1.43 7.53 4.10 1.07 10.36 9 3.54 1.00 8.78 3.60 1.59 6.80 2.81 1.36 4.90 2.29 1.19 3.85 2.23 1.18 3.53 2.88 1.47 4.91 10 3.48 1.37 7.14 4.02 1.35 8.25 3.25 1.49 6.61 2.75 1.18 5.32 2.71 1.38 4.44 3.35 1.44 6.88 11 3.44 1.35 6.80 4.06 1.57 7.68 3.13 1.35 5.88 2.85 1.50 4.65 2.71 1.46 4.19 3.15 1.57 5.13 12 3.85 1.21 8.66 3.69 1.40 7.66 3.25 1.36 6.77 2.92 1.43 5.35 2.67 1.28 5.08 3.23 1.31 6.80 13 3.10 1.25 6.43 3.58 1.47 7.74 3.33 1.28 7.24 3.04 1.38 5.91 2.90 1.39 5.12 3.23 1.39 6.67 14 3.94 1.31 9.10 3.21 1.40 6.61 3.44 1.55 5.86 3.15 1.41 6.28 3.08 1.55 5.25 3.58 1.54 6.51 15 4.38 1.05 11.08 3.35 1.35 6.95 3.63 1.30 7.47 3.54 1.50 6.79 3.19 1.55 5.57 3.75 1.30 8.01 16 3.33 1.25 7.67 3.40 1.24 7.52 3.44 1.35 7.07 3.10 1.19 6.28 2.90 1.23 5.85 3.40 1.19 7.83 17 4.23 1.40 9.01 3.08 1.38 5.60 4.06 1.68 7.24 3.19 1.48 5.60 3.27 1.44 6.02 4.08 1.59 8.13 18 4.27 1.15 9.93 3.60 1.40 7.32 3.85 1.14 9.13 3.50 1.32 7.92 3.42 1.34 7.09 3.75 1.35 7.94 19 2.94 1.23 5.73 3.44 1.46 6.67 2.88 1.30 5.25 2.56 1.31 4.29 2.42 1.26 3.99 2.81 1.30 5.35 20 3.33 1.12 8.16 3.65 1.28 8.52 3.06 1.38 5.79 2.48 1.24 4.08 2.50 1.32 3.79 2.94 1.25 5.56 21 3.33 1.26 6.78 3.27 1.32 6.88 3.13 1.32 5.94 2.73 1.22 5.26 2.73 1.25 5.18 3.04 1.24 6.37 22 3.21 1.35 6.51 4.15 1.41 8.50 3.40 1.35 6.52 3.23 1.46 6.19 3.15 1.43 5.87 3.35 1.45 6.14 23 4.31 1.24 9.96 3.38 1.35 6.76 3.73 1.51 7.05 3.04 1.50 5.06 3.02 1.48 5.04 3.90 1.53 6.95 24 3.27 1.20 7.04 3.10 1.28 6.13 3.08 1.29 6.11 2.79 1.29 5.27 2.56 1.17 5.02 2.96 1.04 6.67 25 3.23 1.34 6.57 3.69 1.60 6.90 2.94 1.34 5.26 2.71 1.47 4.53 2.65 1.51 4.12 2.85 1.21 5.84 26 4.06 1.30 9.33 3.17 1.37 6.64 3.50 1.53 6.49 2.92 1.50 5.08 2.85 1.51 4.67 3.42 1.54 6.32 27 3.17 1.42 6.14 4.35 1.51 8.92 3.19 1.41 5.94 2.85 1.54 4.49 2.96 1.53 4.61 3.10 1.45 5.77 28 3.15 1.31 6.21 3.29 1.15 7.41 3.27 1.41 6.04 3.02 1.38 5.72 2.92 1.26 5.70 3.54 1.35 7.28 29 4.06 1.20 10.13 3.63 1.32 8.58 3.83 1.39 8.36 3.25 1.23 7.54 3.23 1.39 7.03 3.75 1.42 7.76
calculated by Eq. (4), with the distinguishing coefficient z ¼0.5, to express the relationship between the ideal (best) and actual normalized S/N ratio of each design. Then, according to Eq. (5), the GRG (which was defined as the aesthetics index) was calculated by summing up the grey relational coefficients multiplied by the adjustment weight corresponding to each aesthetic criterion. An experimental sample with higher GRG has better performance in aesthetics satisfaction. Table 7 shows the grey relational coefficient and GRG for each experimental sample (design) using the L032 (2149) OA. It shows that sample no. 8 has
the best performance with multiple-criteria characteristics among the 32 samples because it has the highest GRG. In other words, optimizing complicated multiple performance characteristics can be converted into optimizing a single performance index, the GRG.
4.8. Determining the optimal combination of design attributes
The response of each form attribute level belonging to a design and the overall effect of each form attribute upon
the performance index (GRG) was investigated using quantitative theory type I analysis. Quantitative theory type I (Nagamachi, 1989) is a multiple regression analysis technique for deducing the relationship between a quantitative variable (a dependent variable) and qualita-tive (nominal) variables (independent variables). Here, the dependent variable is the GRG of each experimental sample (phone design). The 10 independent form attribute variables (A–J) were represented one mobile phone design. Table 8 shows the quantitative theory type I analysis results.
In the last two rows of Table 8, R represents the correlation between the observed and predicted values of the dependent variable, and ranges from 0 to 1. The coefficient of multiple determination is R2. This explains
the linear relation between the independent variables (10 form attributes) and the dependent variable (GRG). The higher the R2 value, the better the linearity between the
dependent and independent variables. The partial correla-tion coefficients (PCC) indicate the relative importance of each of the 10 product variables (A–J) to overall aesthetics satisfaction. For example, the variable with the highest PCC is the ‘‘top shape’’ (PCC ¼ 0.790), meaning that the top
Table 4 (continued )
No. Simplicity Originality Completeness Pleasure Sat. of form Unity
Mean S. S/N Mean S. S/N Mean S. S/N Mean S. S/N Mean S. S/N Mean S. S/N 30 4.02 1.51 8.16 3.60 1.37 7.65 3.77 1.52 7.45 3.48 1.59 6.41 3.54 1.58 6.45 3.81 1.56 7.10 31 3.44 1.35 7.42 3.67 1.49 7.73 3.35 1.45 6.12 2.81 1.42 4.67 2.81 1.45 4.80 3.42 1.51 6.20 32 3.81 1.38 8.45 3.58 1.34 8.05 3.60 1.55 7.16 3.10 1.46 6.31 3.17 1.53 6.08 3.52 1.51 7.06 MAX 4.98 – 11.86 4.35 – 10.09 4.06 – 9.13 3.54 – 7.92 3.54 – 7.53 4.1 – 10.36
Table 5
The Kano classification of aesthetic criteria
Criteria A O M N R Total (%) Kano category
Originality 50 20 14 16 0 100 A Unity 30 8 20 38 4 100 I Completeness 28 46 8 16 2 100 O Pleasure 52 30 14 4 0 100 A Satisfaction of form 64 22 10 2 2 100 A Simplicity 28 32 38 2 0 100 M
A—attractive, O—one-dimensional, M—must-be, I—indifference, R—reversal.
Table 6
Result of the determined criteria weights
Simplicity Originality Completeness Pleasure Sat. of form Unity The adjusted criteria weights using Kano model
Kano category M A O A A N
K 1 4 2 4 4 0
Raw weight 0.271 0.122 0.240 0.177 0.190 0 Adjusted weight 0.100 0.180 0.177 0.261 0.281 0
Conventional weighting method
shape of a mobile phone has the greatest influence on the perceived mobile phone aesthetics. The ‘‘corner type’’ (PCC ¼ 0.078) is the least significant variable. According to the analysis, the ‘‘top shape’’ (PCC ¼ 0.790), ‘‘body shape style’’ (PCC ¼ 0.752), ‘‘outline division style’’ (PCC ¼ 0.697) and ‘‘outline of number button’’ (PCC ¼ 0.677) are the most significant design attributes affecting the aesthetics of a mobile phone. The category grades of a level indicate the effect of each form attribute on each level for aesthetics satisfaction. A positive grade indicates that this form attribute level can increase the perceived aesthetics of a mobile phone, while a negative grade should be avoided in product aesthetics. Based on the analysis, the optimal combination of form attributes, i.e. A2, B2, C3, D3, E2, F3, G2, H1, I1 and J1, is summarized inTable 9. A computer image of this optimized mobile phone design with the highest aesthetics satisfaction was constructed, as shown in no. 7 ofFig. 4.
4.9. Verification of improvement
A confirmation test was then conducted to verify the performance of the optimized design generated using the
proposed method. Six product design experts chose five mobile phones for competition. These phones are cur-rently available or ready to enter the market. To provide an identical baseline for this evaluation, each sample (e.g. a real product) was transformed into a 2D image according to its specified form attribute level setting. The mobile phone form attributes including screen, function button style and speaker receiver were controlled to be the same as the samples used in our Taguchi experiment. The compared designs are shown inFig. 4. An optimized design using the grey-based TM without applying the Kano model was also generated for compar-ison. In this aesthetics optimization case, the experimen-tal data for the six aesthetic criteria and the weights for these six criteria were determined using the entropy weighting method, shown in Table 6. The optimal combination of form attributes for mobile phone design was found using GRA and the quantitative theory type I analysis, i.e. A4, B2, C3, D3, E1, F3, G3, H2, I1 and J1. The constructed image of this optimized design is shown in no. 6 ofFig. 4.
To conduct the competition evaluation, the same 60 subjects were asked to rank the seven samples on each of six aesthetic criteria and overall satisfaction with
Table 7
GRA aesthetics optimization results (z ¼ 0.5)
Simplicity Originality Completeness Pleasure Sat. of form GRG Weight 0.100 0.180 0.177 0.261 0.281 Ideal 1 1 1 1 1 1 1 0.715 0.335 0.572 0.478 0.478 0.493 2 0.779 0.473 0.847 0.565 0.495 0.600 3 0.946 0.637 0.979 0.731 0.714 0.775 4 0.393 0.833 0.558 0.379 0.363 0.489 5 0.459 0.566 0.402 0.372 0.378 0.423 6 0.686 0.633 0.856 0.746 0.694 0.724 7 0.503 1.000 0.797 0.512 0.475 0.639 8 1.000 0.699 0.890 0.633 1.000 0.830 9 0.506 0.449 0.364 0.341 0.333 0.379 10 0.401 0.593 0.491 0.447 0.393 0.461 11 0.384 0.527 0.427 0.392 0.375 0.417 12 0.496 0.525 0.507 0.450 0.449 0.478 13 0.367 0.533 0.562 0.512 0.454 0.494 14 0.533 0.435 0.426 0.562 0.467 0.485 15 0.801 0.461 0.594 0.651 0.505 0.580 16 0.429 0.510 0.540 0.562 0.543 0.530 17 0.525 0.374 0.563 0.476 0.570 0.504 18 0.620 0.492 1.000 1.000 0.820 0.820 19 0.340 0.440 0.385 0.367 0.361 0.379 20 0.460 0.631 0.420 0.354 0.348 0.425 21 0.383 0.455 0.432 0.442 0.460 0.442 22 0.370 0.628 0.482 0.550 0.546 0.533 23 0.624 0.446 0.538 0.424 0.445 0.474 24 0.395 0.404 0.445 0.443 0.444 0.432 25 0.373 0.457 0.386 0.383 0.370 0.392 26 0.555 0.437 0.479 0.426 0.412 0.446 27 0.355 0.696 0.432 0.380 0.406 0.451 28 0.358 0.500 0.440 0.489 0.522 0.478 29 0.645 0.640 0.759 0.847 0.799 0.760 30 0.460 0.524 0.590 0.582 0.650 0.580 31 0.415 0.532 0.446 0.393 0.423 0.438 32 0.480 0.568 0.552 0.567 0.580 0.559
a 7-point Likert scale (1 ¼ not at all to 7 ¼ intensely so). Confirmation test results were obtained for the means, standard deviations and S/N ratios of evaluations on each
criterion and overall satisfaction. The optimal designs, nos. 6 and 7, exhibit smaller standard deviations for each of the criterion evaluation based on different weighting
Table 8
Quantitative theory type I results
Form attribute Level Category grade PCC Form attribute Level Category grade PCC
A A1 0.032 0.752 F F1 0.010 0.697 A2 0.034 F2 0.030 A3 0.091 F3 0.076 A4 0.025 F4 0.035 B B1 0.043 0.598 G G1 0.028 0.487 B2 0.053 G2 0.041 B3 0.009 G3 0.010 B4 0.001 G4 0.003 C C1 0.078 0.790 H H1 0.015 0.213 C2 0.017 H2 0.001 C3 0.087 H3 0.013 C4 0.008 H4 0.002 D D1 0.017 0.619 I I1 0.066 0.677 D2 0.034 I2 0.006 D3 0.061 I3 0.009 D4 0.010 I4 0.051 E E1 0.002 0.078 J J1 0.021 0.422 E2 0.004 J2 0.021 E3 0.003 Constant 0.528 E4 0.005 R ¼ 0.931 R2 ¼0.867 Table 9
Optimal form elements of the mobile phone design for enhancing aesthetics
methods adopted in the robust design approach of the TM. Other non-optimized designs had significantly higher standard deviations. It can be concluded that the optimized design using the robust design approach consistently maintained a lower variation between con-sumer evaluations.
Table 10shows that the performance of sample no. 7 (the optimized design generated by the proposed inte-grative method) in overall satisfaction and most of the aesthetic criteria were better than the compared designs. It obtained the highest evaluations on the attractive qualities of originality, pleasure and satisfaction of form. The obtained evaluations on completeness, simplicity and unity were moderate. The best performance in each criterion in the Taguchi experiment (shown in the second last row of Table 10) was also used as a reference for comparison with the optimized design. This result verified that the optimal combination of form attributes can effectively improve aesthetic quality with multiple-criter-ia characteristics, leading to higher consumer satisfaction. Furthermore, GRA was conducted to compare the efficiency of these two optimized designs (nos. 6 and 7) on improving the overall S/N ratios. The GRG for each sample of the confirmation test was computed using the adjusted weights based on the Kano model and the entropy weights, as shown in Table 6. Table 10 shows the results of this analysis, and sample no. 7 was better than the compared designs because it obtained the highest GRG. The optimal design (no. 7), generated by
the proposed approach, effectively enhanced attractive and one-dimensional criteria and properly offered the must-be criterion. It generated higher customer satisfac-tion than the optimal design computed without the Kano model. This confirmed the benefits of using the weight adjustment incorporated with Kano analysis to more accurately re-prioritize criteria improvement and resolve trade-offs in the multiple-criteria optimization problem.
A performance evaluation based on the ranking score (Feng and Wang, 2000) was also utilized to further investigate the validity of the proposed integrative method. Table 10 shows that the performance of each design in each criterion was ranked based on S/N ratio magnitude. The ranking of 1st, 2nd, 3rd, 4th, 5th, 6th and 7th then were scored with points 7, 6, 5, 4, 3, 2 and 1, respectively. The ranking score of each design in each criterion is presented inTable 11. Sample no. 7 was ranked 1st three times and 3rd three times, and its total score was 36 (7 3+5 3 ¼ 36). Total score results for each design were computed, and are shown inTable 11. The optimized design no. 7 obtained higher scores than the compared designs and the other optimized design no. 6. It can be concluded that the proposed method of integrating the Kano model into the robust design approach is effective in simultaneously achieving aesthetic quality and overall satisfaction. Furthermore, a computer-aided design (CAD) system can use the optimization results to build a 3D model for facilitating the mobile phone design process, as shown inFig. 5.
Table 10
Confirmation test results
No. S/N ratio Overall
sat.
GRG Simplicity Originality Completeness Pleasure Sat. of
form
Unity The adjusted weights The entropy weights 1 10.55 10.41 10.94 11.73 12.22 12.39 11.36 0.70 0.74 2 12.13 7.23 11.43 10.33 8.22 9.37 9.67 0.53 0.59 3 11.25 9.91 12.83 11.67 10.18 11.77 11.95 0.67 0.73 4 12.92 7.07 10.46 7.78 7.34 9.34 9.67 0.47 0.57 5 11.50 11.98 9.73 10.70 8.60 10.97 10.75 0.55 0.59 6 12.72 11.71 12.44 11.55 11.98 12.22 12.72 0.78 0.84 7 12.36 14.37 12.33 12.95 12.52 12.01 13.07 0.96 0.93 The best performance in Taguchi
experiment
11.86 10.09 9.13 7.92 7.53 10.36 – – – Improvement in S/N ratio 0.50 4.28 3.20 5.03 4.99 1.65 – – –
Table 11
Result of the performance evaluations by the ranking score
No. Performance order of each design in each criterion Total scores Simplicity Originality Completeness Pleasure Sat. of form Unity
1 1 4 3 6 6 7 27 2 4 2 4 2 2 2 16 3 2 3 7 5 4 4 25 4 7 1 2 1 1 1 13 5 3 6 1 3 3 3 19 6 6 5 6 4 5 6 32 7 5 7 5 7 7 5 36
5. Conclusion
This study uses a robust design approach that integrates the Kano model to optimize quality with multiple-criteria characteristics to achieve aesthetic satisfaction. The proposed robust design approach can be applied to objective and subjective quality, especially for multiple-criteria optimization. Hence, using the TM experimental design method requires only a small number of experiments, which saves time and money. Using the Kano model helps to differentiate between multiple criteria affecting customer satisfaction. It can also re-prioritize criteria to resolve the trade-off dilemma in multiple-criteria optimization. The application of this method was demonstrated with a case study on optimiz-ing the aesthetics of mobile phones. It was verified that the Kano model presented advantages to better under-stand customer requirements, to identify the critical and high-return factors of customer satisfaction, and to resolve the trade-off dilemma in multiple-criteria decision making. This method improves the accuracy of criteria priorities determination. It can also be used to resolve existing problems in the current research on subjective quality, such as the Kansei engineering approach ( Naga-machi, 1995; Nagamachi, 2002). The proposed method deals with the complicated inter-relationship between multiple criteria, reduces the variations existing between different customer evaluations and ensures accuracy with an economical and effective experimental design method. The results from this study provide useful insights for designing a mobile phone with optimal design attributes for enhancing aesthetics and overall customer satisfac-tion. The proposed robust design method integrated with the Kano model may be used as a universal method to simultaneously enhance customer satisfaction and pro-duct quality despite multiple-criteria characteristics.
A recommendation for future research in this integra-tive method is to define a means for more accurately representing and quantifying the information provided by the Kano model. The purpose would be to reduce ambiguity in criteria that straddle two categories. In addition to relationship analysis (linear and non-linear) between customer satisfaction and criteria performance, non-linear techniques such as neural networks could also be used in further studies.
Acknowledgements
The authors are grateful to the editor and referees for their valuable comments and advice in this paper. This research was supported in part by the National Science Council of Taiwan, ROC under Grant no. NSC 92-2213-E-009-071.
References
Bloch, P.H., 1995. Seeking the ideal form: Product design and consumer response. Journal of Marketing 59, 16–29.
Bottani, E., Rizzi, A., 2008. An adapted multi-criteria approach to suppliers and products selection—An application oriented to lead-time reduction. International Journal of Production Economics 111, 763–781.
Cabrera-Rios, M., Mount-Campbell, C.A., Irani, S.A., 2002. An approach to the design of a manufacturing cell under economic considerations. International Journal of Production Economics 78, 223–237. Chen, C.T., Lin, C.T., Huang, S.F., 2006. A fuzzy approach for supplier
evaluation and selection in supply chain management. International Journal of Production Economics 102, 289–301.
Chuang, M.C., Ma, Y.C., 2001. Expressing the expected product images in product design of micro-electronic products. International Journal of Industrial Ergonomics 27, 233–245.
Chuang, M.C., Chang, C.C., Hsu, S.H., 2001. Perceptual factors underlying user preferences toward product form of mobile phones. Interna-tional Journal of Industrial Ergonomics 27, 247–258.
Coates, D., 2003. Watches Tell More Than Time: Product Design, Information and the Quest for Elegance. McGraw-Hill, UK. Conklin, M., Powaga, K., Lipovetsky, S., 2004. Customer satisfaction
analysis: Identification of key drivers. European Journal of Opera-tional Research 154, 819–827.
Deng, J., 1982. Control problems of grey systems. System and Control Letters 5, 288–294.
Deng, J., 1989. Introduction to grey system. The Journal of Grey System 1 (1), 1–24.
Dimova, L., Sevastianov, P., Sevastianov, D., 2006. MCDM in a fuzzy setting: Investment projects assessment application. International Journal of Production Economics 100, 10–29.
Feng, C.M., Wang, R.T., 2000. Performance evaluation for airlines including the consideration of financial ratios. Journal of Air Transport Management 6, 133–142.
Fynes, B., Bu´rca, S.D., 2005. The effects of design quality on quality performance. International Journal of Production Economics 96, 1–14. Huiskonen, J., Pirttila¨, T., 1998. Sharpening logistics customer service strategy planning by applying Kano’s quality element classification. International Journal of Production Economics 56–57, 253–260. Hwang, C.L., Yoon, K., 1989. Multiple Attributes Decision Making.
Springer, Berlin.
Kano, K.H., Hinterhuber, H.H., Bailon, F., Sauerwein, E., 1984. How to delight your customers. Journal of Product and Brand Management 5 (2), 6–17.
Korpela, J., Lehmusvaara, A., Nisonen, J., 2007. Warehouse operator selection by combining AHP and DEA methodologies. International Journal of Production Economics 108, 135–142.
Lai, H.H., Chang, Y.M., Chang, H.C., 2005. A robust design approach for enhancing the feeling quality of a product: A car profile case study. International Journal of Industrial Ergonomics 35, 445–460. Lai, H.H., Lin, Y.C., Yeh, C.H., Wei, C.H., 2006. User-oriented design for the
optimal combination on product design. International Journal of Production Economics 100, 253–267.
Lavie, T., Tractinsky, N., 2004. Assessing dimensions of perceived visual aesthetics of web sites. International Journal of Human–Computer Studies 60, 269–298.
Lin, J.L., Lin, C.L., 2002. The use of the orthogonal array with grey relational analysis to optimize the electrical discharge machining process with multiple performance characteristics. International Journal of Machine Tools & Manufacture 42, 237–244.
Liu, Y., 2003. Engineering aesthetics and aesthetic ergonomics: Theore-tical foundations and a dual-process research methodology. Ergo-nomics 46 (13/14), 1273–1292.
Matzler, K., Hinterhuber, H.H., 1998. How to make product development projects more successful by integrating Kano’s model of customer satisfaction into quality function deployment. Technovation 18 (1), 25–38.
Nagamachi, M., 1989. Kansei Engineering. Kaibundo Publication Com-pany, Tokyo.
Nagamachi, M., 1995. Kansei Engineering: A new ergonomic consumer-oriented technology for product development. International Journal of Industrial Ergonomics 15 (1), 3–11.
Nagamachi, M., 2002. Kansei engineering as a powerful consumer-oriented technology for product development. Applied Ergonomics 33, 289–294.
Nearchou, A.C., 2006. Meta-heuristics from nature for the loop layout design problem. International Journal of Production Economics 101, 312–328.
Nielsen, J., 1993. Usability Engineering. Academic Press Ltd., UK. Partovi, F.Y., 2007. An analytical model of process choice in the chemical
industry. International Journal of Production Economics 105, 213–227.
Postrel, V., 2001. Can good looks guarantee a product’s success? The New York Times July 12.
Rashid, A., Mac Donald, B.J., Hashmi, M.S.J., 2004. Evaluation of the aesthetics of products and integration of the finding in a proposed intelligent design system. Journal of Material Processing Technology 153–154, 380–385.
Schenkman, B.N., Jonsson, F.U., 2000. Aesthetics and preferences of web pages. Behavior and Information Technology 19 (5), 367–377. Schu¨tte, S., Eklund, J., 2005. Design of rocker switches for
work-vehicles—An application of Kansei engineering. Applied Ergonomics 36, 557–567.
Shao, C.J., Chen, C.C., Tung, T.C., Chen, K., Guan, S.S., Deng, Y.S., Chang, Y.M., 2000. Exploring the relationship between users’ Kansei evaluation and mobile phone designs. Industrial Design 28 (2), 154–159 (in Chinese).
Taguchi, G., Clausing, D., 1990. Robust Quality. Harvard Business Review, Boston.
Talia, L., Noam, T., 2004. Assessing dimensions of perceived visual aesthetics of web sites. International Journal of Human–Computer Studies 60, 269–298.
Tan, K.C., Pawitra, T.A., 2001. Integrating SERVQUAL and Kano’s model into QFD for service excellence development. Managing Service Quality 11 (6), 418–430.
Tarng, Y.S., Juang, S.C., Chang, C.H., 2002. The use of grey-based Taguchi methods to determine submerged arc welding process parameters in hard facing. Journal of Materials Processing Techno-logy 128, 1–6.
Wang, C.H., Tong, L.I., 2004. Optimization of dynamic multi-response problems using grey multiple attribute decision making. Quality Engineering 17 (1), 1–9.
Wang, L., Chu, J., Wu, J., 2007. Selection of optimum maintenance strategies based on a fuzzy analytic hierarchy process. International Journal of Production Economics 107, 151–163.
Wu, K.C., Chuang, M.C., 2003. Quality model of product Kansei—Using mobile phone as examples. Master’s Thesis, Institute of Applied Arts, National Chiao Tung University, Taiwan (in Chinese).
Yamamoto, M., Lambert, D.R., 1994. The impact of product aesthetics on the evaluation of industrial products. Journal of Product Innovation Management 11, 309–324.
You, H., Ryu, T., Oh, K., Yun, M.H., Kim, K.J., 2006. Development of customer satisfaction models for automotive interior materials. International Journal of Industrial Ergonomics 36, 323–330. Yun, M.H., Han, S.H., Hong, S.W., Kim, J., 2003. Incorporating user
satisfaction into the look-and-feel of mobile phone design. Ergo-nomics 46 (13/14), 1423–1440.
Zeleny, M., 1982. Multiple Criteria Decision. Singapore.
Zwick, F., 1967. The morphological approach to discovery, invention, research and construction. In: New Method of Thought and Procedure: Symposium on Methodologies, Pasadena, May, pp. 316–317.
