A study of consumer preferences for e-retailers’ attributes: an application of conjoint analysis

Technology Information and Science Management of Journal International The (IJMSIT) Publishers NAISIT Chief in Editor [email protected] Email: Portugal, Interior, Beira of University Ferreira, J. J. Editors Associate Portugal interior, Beira of University Ferreira, M. J. João Editor-in-Chief: Editors: Main USA Memphis, of University and Portugal Lisbon, of Institute University Ferreira, F. A. Fernando Spain Barcelona, of University Lindahl, Merigó M. José Editors: Assistant University, Portucalense and (UBI) Sciences Business in Unit -Research NECE at Reseacher Fernandes, Cristina Portugal UK Reading, of University Co, Jess Portugal Lisbon, of Institute University Jalali, S. Marjan Board: Advisory Editorial UK Management, of School Cardiff Lincoln, Adebimpe Israel College, Academic Netanya Tziner, Aharon USA Pennsylvania, University, Morris Robert Smith, D. Alan Spain Barcelona, of University Lafuente, G. Maria Ana Norway Management, of School Oslo Mariussen, Anastasia Spain Barcelona, de Autònoma Universitat Tarrés, i Serarols Christian UK university, City -Birmingham School Business Millman, Cindy Romania Bucharest, of University Gh, Popescu R. Cristina UK School, Business University Newcastle Irawati, Dessy Spain Valencia, of University Ribeiro, Domingo USA Business, of Schools Carayannis, G. Elias USA University, Technological Michigan Oliveira, Emanuel Spain Seville, of University Liñán, Francisco UK University, City Birmingham Matlay, Harry Romania Studies, Economic of University Bucharest The Purcarea, Irina HK University, Polytechnic Kong Hong The Choi, Jason Spain Valencia, of University Vila, Jose Canada Montréal, HEC Filion, Jacques Louis Italy II, Federico Naples of University Landoli, Luca Brazil Paulo, Säo de Universidade at Researcher Sakuda, Ojima Luiz Portugal Interior, Beira of University Raposo, L. Mário Spain València, de Politècnica Universitat Peris-Ortiz, Marta Zealand New Waikato, of University The Akoorie, Michele Canada Trois-Rivières, à Québec du Université Julien, Pierre-André Jordan University, Hashemite The Karabsheh, Radwan Zimbabwe Technology, and Science of University National Mhlanga, Richard Brazil – Vargas Getulio Fundação Bandeira-de-Mello, Rodrigo Netherlands The - University Tilberg Rutten, Roel Verde Cabo Empresariais, e Económicas Ciências de Superior Instituto Cruz, Rosa Netherlands The Rotterdam, University Erasmus Thurik, Roy India Delhi, Technology of Institute Indian Jain, K. Sudhir Portugal Interior, Beira of University Azevedo, G. Susana Denmark University, Business Copenhagen Hollensen, Svend Austria Vienna, of University Frisch, Walter USA University, State Colorado Byrne, S. Zinta Board Review Editorial Turkey Turkey, University Selçuk Ögüt, Adem Greece Athens, of University Agricultural Sideridis, B. Alexander Netherlands The Amsterdam, University VU Sharpanskykh, Alexei USA York, -York, University State Pennsylvania Kara, Ali Brazil Rio, Grande Universidade Freitas, Angilberto Portugal Interior, Beira of University Paço, do Arminda Finland Jyväskylä, of University Ojala, Arto Portugal Douro, Alto e Tras-os-Montes of University Marques, Carla Turkey University, Çukurova Tanova, Cem Brazil Catarina, Santa de Federal Universidade Tolfo, Cristiano Portugal Branco, Castelo of Institute Polytechnic Estevão, S. Cristina Croatia Split, of University Miocevic, Dario Zealand New School, Business Auckland of University The Askarany, Davood USA Washington, of University Revere, Debra USA Ohio, Cincinnati, of University Gormley, Kolesar Denise Kong Hong Technology, and Science of University Kong Hong Chiu, K.W. Dickson Spain Navarra, of University Melé, Domènec Brazil School, Business FUCAPE Mainardes, Emerson USA University, Arizona Northern Otenyo, E. Eric USA University, Illinois Southern Watson, W. George Brazil Maria, Santa de Federal Universidade Moura, de Luiz Gilnei China University, Psychology,Zhejiang of Department Zhong, An Jian Portugal Lisbon, University, Catholic Portuguese Sciences, Human of Faculty Pinto, Carneiro Joana Spain Valencia, of University Alegre, Joaquín USA Jersey, New Business, of School Anisfield Rakotobe, Thierry Joel USA , FL Sanford, Florida, Central of University Matusitz, Jonathan India Kharagpur, Technology of Institute Indian Srivastava, L. B. Kailash Netherlands Twente,The of University Sanders, Karin Germany Koblenz-Landau, of University Troitzsch, G. Klaus China Nanjing, Technology, of University Nanjing Shi, Kuiran Portugal ISLA, Faria, Costa da Liliana Canada Ontario, Western of University Capretz, Fernando Luiz USA Business, of College Godkin, Lynn Canada Winnipeg, of University Liu, Chunhui Maggie Belgium Liège, of University Ausloos, Marcel USA Texas, University,Denton, Woman's Texas Benham-Hutchins, Marge Spain Granada, of University Pérez-Aróstegui, Nieves María Italy Udine, of University Cagnina, Rosita Maria University,Taiwan Hwa Dong National Tabata, Mayumi Portugal University, Lusíada and University Portucalense Pinho, Micaela Italy Basilicata, of University Renna, Paolo Portugal Coimbra, of University Cunha, Rupino Paulo Germany University, Saarland Loos, Peter Spain Vigo, de Empresas de Administración e Economia de F. García, Piñero Pilar Romania Bucharest, Studies, Economic of University Bucharest Gheorghe, N. Popescu Economic of University Bucharest The and Satu-Mare of Academy Commercial The Adriana, Veronica Popescu Romania Bucharest, Studies, India Technology, and Management of Institute Singh, Ramanjeet Portugal of University Catholic Morais, Ricardo Spain Rioja, of University Ortiz, Fernández Ruben Canada Manitoba, of University Thulasiram, K. Ruppa USA NJ, University,Montclair, State Montclair Kim, Soo Taiwan University, Yat-Sem Sun National Chiou, Wen-Bin USA GA, ,Augusta, College Paine Lawless, Willaim Singapore University, Management Singapore Koh, T.H. Winston Table of Contents This is one paper of The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 38 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 A STUDY OF CONSUMER PREFERENCES FOR E-RETAILERS’ ATTRIBUTES: AN APPLICATION OF CONJOINT ANALYSIS. BEGOÑA PERAL PERAL JOAQUINA RODRÍGUEZ-BOBADA REY ANGEL FRANCISCO VILLAREJO RAMOS UNIVERSIDAD DE SEVILLA The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 39 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 ABSTRACT The aim of this work is to determine and analyse consumer preferences regarding the profiles of an e-retailer’s web page. Two types of products are examined, a pleasure trip and a laptop computer, to test whether there are differences in the individuals’ preferences. There are two reasons for this choice: these two products are purchased the most over the Internet in Spain and the different motives for buying them hedonic-pleasurable and utilitarian. We conducted an initial study, from which we identified the principal attributes valued by the participants in the survey. These attributes were then used to design the profiles for the conjoint analysis. The variables that are most relevant to the shopping task are those which receive a higher response frequency. In both products, the attributes that are most valued by the participants are the virtual store’s security and privacy policy. However, for a laptop computer, consumers also emphasize the importance of the provision of the technical details of the product and the fact that the supplier also has a physical store. We recommend that e-retailers’ web pages need to clarify and facilitate access to the most relevant variables to the shopping task. Likewise, public institutions and e-retailers need to continue to work towards minimising non-buyers’ rejection of online purchasing and their fears regarding security on the Web. Firms with both physical and online outlets have an important competitive advantage over pure-players, for certain products at least. Keywords e-retailers, consumer behaviour, conjoint analysis. INTRODUCTION Electronic commerce has made rapid and radical changes to the way we make our purchases today. Proof of this is the €2.322 million -worth of goods and services purchased over the Internet in Spain in the second trimester (CMT, 2011). However, firms should not focus solely on increasing their Internet sales, but rather, as Rust et al. (2001) indicate, consider the potential sales they are missing out on because they are not offering what the consumer wants. Previous studies have analysed the relative importance of the attributes of a website, understood as “those factors both functional and psychological The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 40 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 that exist in an online store” (Lim and Dubinsky, 2004, p.501). Palmer (2002) and Yun and Good (2007) found that not all of a web page’s attributes will receive the same favourable response from online consumers. It is essential therefore to understand how customers evaluate different attributes when they decide to make an online purchase [Iqbal et al. (2003); Bart et al. (2005); Shun and Yunjie (2008)]. In other words, electronic commerce needs to identify and focus on developing attributes which increase value for the customer [Han and Han (2002); Su (2007)]. The specific aim of this research is to identify and analyse consumers’ preferences for the different attributes of e-retailers’ web pages. Prior studies have focused on understanding whether a consumer’s preference for online shopping changes with different types of products (Korgaonkar et al., 2006). In this investigation, the products we have chosen are pleasure trips and laptop computers because these are the items that are purchased the most over the Internet in Spain. In 2010, 52.4% and 42.93% of Internet shoppers bought travel tickets and booked accommodation respectively; and electronic products in general were purchased by one in four Internet shoppers in Spain (ONTSI, 2011). Another reason for this choice is the different motives for buying these two products, respectively being, hedonic-pleasurable and utilitarian. This distinction between the type of product analysed allows us to carry out a conjoint analysis to test whether there are differences in the importance given to different attributes. Following a literature review, we explain the methodology used and the reasons for our choice. We then set out the results of our conjoint analysis and propose a number of arguments and implications for their development. Lastly, the study’s limitations and future lines of research will be discussed. CONCEPTUAL BACKGROUND To understand a consumer’s choice of e-retailer, we must consider the relative importance that consumers give to its attributes at the time of purchase (Lim and Dubinsky, 2004). Prior studies have analysed the attributes of the online store as predictors of the consumers’ intention to buy [Bart et al., (2005); Su, (2007)], their satisfaction, their acceptance of new technology (Song and Zinkhan, 2003), their attitude towards online purchases (Lim and Dubinsky, 2004) and customer loyalty [Zeithaml et al., (2002); Yun and Good, (2007)]. We now discuss the dimensions proposed in previous investigations, cited below: merchandise, convenience, interactivity, navigation, reliability, promotions and design. The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 47 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 3. Results of initial study We received 140 valid questionnaires, from students enrolled in two degree courses, of whom 65% were women and 92% of the sample was under 27 years of age, 65% of the participants had made an online purchase in the previous year and 78% intended to do so over the coming year. The participants indicated the most important attributes of a virtual store (Table 2) some of which are common to both products, while others show statisticallysignificant differences according to the productii. Table 2. Initial study: Percentages of choice of attributes Pleasure trip Laptop computer Attributesiii % of choice % of choice Price of the product/ service 77.857 80.714 Product guarantee and returns policy 64.286 77.857 Data security and privacy policies 61.429 52.857 Payment information 55.000 40.714 Product’s images 46.429 53.571 Existence of alternative payment 45.714 46.429 Make phone or e-mail contact 40.714 38.571 Option to reserve products 39.286 17.857 Company reputation 38.571 36.429 Posting customer reviews 37.857 30.714 Information on how to buy 36.429 30.714 Information on postage and packing costs 32.143 40.000 Technical product description 30.714 61.429 Well-know brands 20.000 32.143 Physical store distributor 17.142 32.143 The results show that the ability to make a reservation (corr=-0.237, sig=0.000) and payment options information (corr=-1.143, sig=0.017) are given the greatest value in the case of travel. For a laptop computer, the most frequently selected attributes are the provision of technical information (corr=0.308, sig=0.000), having a bricks and mortar presence (corr=0.174, sig=0.003), stocking wellknown brands (corr=0.138, sig=0.021) and offering a product guarantee and returns policy (corr=0.150, sig=0.012). The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 48 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 Furthermore, for each product we analysed whether there were any significant correlations between attributes. This is because a fractional factorial design (Hair et al., 2000) must be orthogonal, that is, there should be no correlation between attributes. The 2-tailed significance of the Fisher statistic indicated that there was a relationship between some of the attributes: for the travel product, the correlated pairs were payment information and the existence of several payment options (2-tailed sig. =0.011); and the firm’s reputation and price (2-tailed sig. =0.039). In the case of the laptop, the related attributes were information on how to pay and price (2-tailed sig. =0.015); and reputation and price (2tailed sig. =0.001). We therefore removed the information on how to pay and the firm’s reputation, because, in addition to their statistical significance, we recognise the conceptual relationship between each respective pair. 4. Application of conjoint analysis The first step was to select the most important attributes –those which were most frequently chosen in the pilot study [Bauer and Scharl, (2000); Bart et al., (2005)]. Then, to limit the amount of information given to the respondents (a high number of attributes generates a great number of profiles for the participants to evaluate), we chose a number of attributes that would give the right balance. Finally, because of the significant differences, we used a different list of attributes for each product. There were nine in total: with seven attributes in common and two which were different for each product. The next step was to decide on the levels for each attribute, making them realistic in order to increase the validity of the preferences (Table 3). Table 3: Attribute and attribute levels Attributes Attribute levels Price of product /service High products price Medium products price Low products price Product guarantee and returns policy Yes No Data security and privacy policies Both policies Either policy No policy Common to both products Product’s images Yes No The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 49 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 Existence of alternative payment Credit card Cash payment Pay by instalments Options to make phone or e-mail contact Yes No Information on postage and packing costs Yes No Option to reserve product Yes No Only for pleasure trip Posting customer reviews Yes No Technical product description Yes No Only for laptop computer Physical store distribution Yes No Our study consists of six attributes with two levels and three attributes with three levels, for each product analysed, giving a possible 1,728 combinationsiv. Given the difficulty of evaluating such a high number of combinations we used a fractional factorial design, which provides an appropriate fraction of all the possible combinations of the attribute levels. The orthogonal matrixv was designed using the SPSS 17.0 Orthoplan procedure, which captures the main effects of each attribute. This matrix consists of eighteen profiles, sixteen of which were used to estimate the model parameters and the remaining two were used to validate the results. We used the full profile method of data collection. Each participant was asked to rank the eighteen combinationsvi on a scale of 1 (least preferred) to 7 (most preferred). The preference model chosen is the part-worth function model, which is suitable for categorical data. To estimate the model’s parameters, the relative importance of the attributes and the partial utility of the levels, we used the SPSS 17 Conjoint Procedure. For the internal validation measures we used the Pearson correlation coefficient and the Kendall tau coefficient. The population sample used to obtain the data consisted of university students, who had not participated in the initial questionnaire. The convenience sample includes students from five university courses, which covers a broad spectrum of the student population. We contacted 290 participants, and received 274 responses in the case of choosing a pleasure trip and 270 for the laptop computer. The The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 50 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 respondents’ demographic data is shown in Table 4. The sample descriptives for both the pilot study and the conjoint analysis were very similar and we can therefore assume that it is acceptable to apply the most important attributes for the choice of an e-retailer that were identified in the pilot study to the conjoint analysis. Table 4. Samples characteristics (in percentages) Characteristics Inicial study (n=140) Pleasure trip (n=274) Laptop computer (n=270) 18 to 27 years 92 92 91,5 Age 27 + years 8 8 8.5 Female 65 63.6 63.5 Sex Male 35 36.4 36.5 Yes 65 65.7 67.8 Previous online purchase No 35 34.7 32.2 Yes 78 78.1 80 Future purchase intention No 22 219 20 ANALYSIS OF THE RESULTS OBTAINED FROM THE CONJOINT ANALYSIS In the example of a pleasure trip, the attributes that are most valued by the participants are the virtual store’s security and privacy policy, having a product price list and availability of product images (Table 5). The participants’ satisfaction with each level of the attributes enables us to identify the sample’s preferred online retailer. This will be the store that combines the levels with the greatest partial utility: providing low product prices, a product guarantee and returns policy, product images, data security and privacy policies, the option for the customer to pay by instalments, the option to make telephone or e-mail contact, clearly visible postage and packing costs, the option to reserve products and posting customer reviews. The total of the partial utilities of these levels indicates the total utility attributed to the preferred online store, plus the constant, which gives a total value of 5.836vii. Table 5. Relative attribute importance and part-worth utilities of attribute levels. Pleasure trip. Attributes Relative importance Levels Part-worth utility estimates Std.error 1 -0.201 0.099 2 -0.123 0.130 Price of product/service 13.565 3 0.324 0.130 The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 51 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 1 0.368 0.074 Product guarantee and returns policy 11.183 2 -0.368 0.074 1 0.394 0.074 Product’s images 11.364 2 -0.394 0.074 1 0.383 0.117 2 0.182 0.140 Data security and privacy policies 17.013 3 -0.565 0.117 1 -0.039 0.099 2 0.001 0.116 Existence of alternative payment 10.361 3 0.038 0.116 1 0.370 0.080 Make phone or e-mail contact 10.621 2 -0.370 0.080 1 0.335 0.074 Information on postage and packing costs 10.268 2 -0.335 0.074 1 0.259 0.074 Option to reserve product 8.390 2 -0.259 0.074 1 0.218 0.080 Posting customer reviews 7.235 2 -0.218 0.080 Constant= 3.147. Pearson´s R: Value = 0.990, sig. =0.000. Kendall´s tau: Value = 0.883, sig. =0.000. Kendall´s tau for holdouts: Value = 1, sig. =0.000 As for the reliability of the results, the Pearson correlation coefficient is 0.990, and the Kendall tau is 0.883, which indicates that the results obtained are reliable. Kendall tau coefficient for the two holdout profiles and their value of 1 confirms the validity of the results. In the case of the laptop computer, the attributes given the highest value by the participants are the data security and privacy policies, the product price and the product guarantee and returns policy (Table 6). The levels that comprise the ideal profile are low product prices, a product guarantee and returns policy, product images, the implementation of a data security and privacy policies, the ability to pay by instalments, the ability to contact the store by telephone or e-mail, clearly visible postage and packaging costs and a physical store. The sum of the partial utilities of these levels is 5.68viii, which is the highest global utility that a profile can attain. Table 6: Relative attribute importance and part-worth utilities of attribute levels. Laptop computer. The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 52 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 Attributes Relative importance Levels Part-worth utility estimates Std. error 1 -0.102 0.118 2 -0.075 0.138 Price of product/service 12.762 3 0.177 0.138 1 0.365 0.088 Product guarantee and returns policy 11.356 2 -0.365 0.088 1 0.341 0.088 Product’s images 10.034 2 -0.341 0.088 1 0.315 0.118 2 0.243 0.138 Data security and privacy policies 16.003 3 -0.558 0.138 1 -0.064 0.118 2 0.158 0.138 Existence of alternative payment 10.565 3 -0.093 0.138 1 0.255 0.088 Make phone or e-mail contact 8.062 2 -0.255 0.088 1 0.297 0.088 Information on postage and packing costs 9.315 2 -0.297 0.088 1 0.384 0.088 Technical product description 11.270 2 -0.384 0.088 1 0.350 0.088 Physical store distribution 10.633 2 -0.350 0.088 Constant= 3.038. Pearson´s R: Value= 0.986, sig. =0.000 Kendall´s tau: Value= 0.900, sig. =0.000. Kendall´s tau for holdouts: Value= 1, sig. =0.000 With regard to the reliability of the results, the Pearson correlation coefficient values, Kendall tau (0.986; 0.9) and Kendall tau for the holdout profiles show that the results are reliable. Finally, we compared the results from conjoint analysis with the replies from the 257 students who responded for both products [Keen, et al. (2004); Chiam et al. (2009)]. Four of the attributes attained similar and have high evaluations in both cases (Table 7). Table 7: Relative attribute importance and part-worth utilities of attribute levels for both products. (Sample size = 257 respondents). The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 53 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 Pleasure trip Laptop computer Attributes Relative importance Pleasure trip Relative importante Laptop computer Levels Part-worth utility estimates Std. Error Part-worth utility estimates Std. Error 1 -0.212 0.098 -0.103 0.119 2 -0.125 0.129 -0.074 0.138 Price of product/service 13.522 12.765 3 0.337 0.129 0.177 0.138 1 0.381 0.074 0.365 0.089 Product guarantee and returns policy 11.351 11.357 2 -0.381 0.074 -0.365 0.089 1 0.402 0.075 0.342 0.088 Product’s images 11.570 10.038 2 -0.402 0.075 -0.342 0.088 1 0.375 0.116 0.314 0.120 2 0.177 0.140 0.245 0.138 Data security and privacy policies 16.826 16.001 3 -0.552 0.116 -0.559 0.138 1 -0.051 0.099 -0.065 0.118 2 0.002 0.115 0.159 0.139 Existence of alternative payment 10.412 10.564 3 0.049 0.115 -0.094 0.139 1 0.364 0.081 0.255 0.087 Make phone or email contact 10.389 8.063 2 -0.364 0.081 -0.255 0.087 1 0.339 0.074 0.299 0.088 Information on postage and packing costs 10,408 9.312 2 -0.339 0.074 -0.299 0.088 1 0.257 0.073 - - Option to reserve product 8.35 - 2 -0.257 0.073 - - 1 0.220 0.079 - - Posting customer reviews 7.182 - 2 -0.220 0.079 - - 1 - - 0.386 0.087 Technical product description - 11.271 2 - - -0.386 0.087 1 - - 0.350 0.088 Physical store distribution - 10.628 2 - - -0.350 0.088 The most important attribute is the concern for privacy and security, which was accorded the highest value in the travel category. According to Bart et al. (2005, p.135), privacy is given a greater The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 54 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 value for products which involve sensitive data, such as a trip, since this requires information such as a customer’s whereabouts and activities. In second place is pricing information, which is also more important in the pleasure trip example (Chiam et al., 2009). Two other attributes, the product guarantee and return policy and information on payment methods, are given similar relative importance. There are differences in the relative importance given to the other attributes. Whereas for the pleasure trip, product images is the third most important, for the laptop, the important attributes are the technical description of the product and the fact that the supplier also has a physical store. These results may be explained by the product type. In the case of the computer, the individual is considered to be in a situation of high rational involvement, which means that the search for information is focused more on the technical aspects of the product. For the trip, on the other hand, this can be viewed as a situation requiring high emotional involvement for the individual, in which there is a search for information, but with a greater focus on the hedonic-pleasurable elements. The different attribute levels present similar partial utilities, except in the case of the payment methods, since for the pleasure trip there is a preference for paying by instalments, whereas for the computer the option of paying up front has a greater utility. In order to test whether there are significant differences in the participants’ choices for the two products, and given that the samples are related, we applied the Wilcoxon signed-rank test for the seven attributes that are common to both product types and their levels (Table 8). Table 8: Wilcoxon signed-rank test. Relative attribute importance Trip/laptop Part-worth utility estimates Trip/laptop Z -2.028(a) -0.071(b) Asymp. Sig. (2-tailed) 0.043 0.943 a Based on positive ranks. b Based on negative ranks. The results show that there are no significant differences in the partial utilities of the levels of the two product types, but there are differences in the relative importance of the attributes. The sum of the relative importance of the seven attributes analysed is 84.48% in the case of the trip and 78.1% for the laptop computer. This means that, in the case of the laptop, the participants consider that almost 22% of an e-retailer’s total importance is derived from the provision of the technical details of the product and the existence of a physical store. Similarly, there are differences in the importance given to the attributes. In the case of the pleasure trip, five out of the seven attributes analysed obtain higher values in this respect than for the laptop. Only the existence of alternative payment options is considered to be slightly more important than in the case of the trip. Finally, turning to the characteristics of the respondents, we tested The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 55 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 for differences between the consumers’ preferences. The variables used to segment the sample were gender and previous experience in Internet purchasing. Using these variables, no statistically-significant differences were found in the relative importance of the attributes or in the partial utility of the levels, for either product. DISCUSSION, PRACTICAL IMPLICATIONS AND LIMITATIONS Firstly, from the results of the initial study to determine the attributes of the profiles for the conjoint analysis, it is clear that the least important attributes are aspects of a web page’s design or the variables which have little relevance to the task (Eroglu et al., 2001). However, the variables which assist the shopping task, such as price, product guarantee and returns policy, the data security and privacy policies, information on how to buy, product images or technical description, all achieved high selection percentages from the participants. These variables have a utilitarian motive and include all the web page descriptors (verbal or pictorial) which appear on the screen, making it easier for consumers to achieve their purchasing aims. E-retailers therefore need to make it easy to identify and access the variables on their web pages which are most relevant to the task, given that online consumers need this information in order to make the decision to buy. Nevertheless, the variables relating to the attractiveness of the web page should not be overlooked. Secondly, although the security and privacy policies are not the most important attribute in the initial study, given the response frequency obtained for the two products analysed, the results of the conjoint analysis show that this is the attribute of the web page which is given the highest relative importance. In fact, according to ONTSI (2008), the risk attached to data security and confidentiality is one of the specific arguments that non-purchasers maintain against online purchasing. Therefore public institutions with policies to support electronic commerce and the firms which sell their products and services on the Internet should continue striving to minimise consumers’ rejection of online purchasing and their concerns regarding the problem of online security. They should publicise the advances made in guaranteeing the privacy of personal data. Thirdly, we can verify that there are differences in the relationship between product type and the most valuable attributes of a web page. Thus, in the case of the laptop, the technical description of the product and the fact that the e-retailer also operates from a physical store are given high importance. In fact, one of the fundamental reasons for Internet users not shopping on the web is their preference for physical stores, where they can see what they are buying and can gather all the technical and commercial information that they believe to be important. Therefore, companies with physical stores and an Internet outlet have significant competitive advantage over the pure-players, for certain products at least, since consumers prefer these websites. The International Journal of Management Science and Information Technology (IJMSIT) Issue 3 - (Jan-Mar 2012) (38 - 62) 56 ISSN 1923-0265 (Print) - ISSN 1923-0273 (Online) - ISSN 1923-0281 (CD-ROM), Copyright NAISIT Publishers 2014 Thus, in April 2010, the EU approved the new Vertical Restraints Block Exemption Regulation, which allows brand owners to prevent the sale of their products on Internet sites which do not also have a physical presence. They could, for example, choose suppliers with a bricks and mortar store, in order to present a uniform sales environment. Fourthly, although –in common with Bhatnagar and Ghose (2004) – we have been unable to prove any differences in the respondents’ preferences according to their characteristics. Conjoint analysis is a methodology which provides managers with useful information which they can apply to their web page design strategy for different consumer segments. Equally, web pages could be tailored according to aspects such as the consumer’s previous experience as an online shopper (Zhu and Zhang, 2010) or age (Kim and Forsythe, 2010). The Internet is a channel for marketing, information and assistance which allows a high degree of adaptation to each client’s profile, to help satisfy all of their consumer needs. Another alternative is proposed by Kamakura et al. (1994), Bhatnagar and Ghose (2004) and Ramaswamy and Cohen (2007), who suggest that the information regarding partial utilities obtained through conjoint analysis could be usefully applied to latent class models. These models could be used in future investigations to identify segments which describe the Internet shopper according to characteristics that are not known a priori to the researcher. Once these segments have been identified, the sociodemographic characteristics of the individuals that comprise each segment could be analysed. This would make it possible to identify whether significant differences exist between the characteristics of the individuals in each segment, thereby facilitating web page design improvements and providing appropriate information relating to the requirements of each segment. Finally, the limitations of this research arise from the methodology used. The number of attributes and levels must be decided by the investigator: a conjoint analysis cannot be carried out using a high number of them, since the factorial design would produce too many profiles for the individual to evaluate. Another limitation refers to the fact that we cannot generalize the results as we have used a convenience sample of university students. CONCLUSION In this study we have analysed consumers’ preferences regarding the attributes built into eretailers’ web pages. Unlike previous research, whose dimensions encompass several attributes, making it difficult to understand the importance ascribed to each one, our study used simple attributes which can be easily evaluated by individuals. This affords a better understanding of how they actually make their choices, maximising the global utility of each attribute. We also proved that the type of product analysed influences the importance of the e-retailer’s attributes. In both products, the attributes that are most valued