Consumer feelings and behaviours towards well designed websites

Consumer feelings and behaviours towards well designed websites

Information & Management 48 (2011) 166–177 Contents lists available at ScienceDirect Information & Management journal homepage: www.elsevier.com/loc...

1MB Sizes 2 Downloads 47 Views

Information & Management 48 (2011) 166–177

Contents lists available at ScienceDirect

Information & Management journal homepage: www.elsevier.com/locate/im

Consumer feelings and behaviours towards well designed websites Natalia Vila a,b,*, Ine´s Kuster a,b,1 a b

Marketing Department & IDOCAL Research Institute, Univeristy of Valencia, Spain Dpto Comercializacio´n e Investigacio´n de Mercados., Edif. Departamental Facultad de Economia (1er piso)., Avda Tarongers, s/n., 46.022 - Valencia, Spain

A R T I C L E I N F O

A B S T R A C T

Article history: Received 5 February 2009 Received in revised form 20 September 2009 Accepted 26 April 2011 Available online 7 May 2011

We analysed the effect of a well designed website in terms of five indicators: purchase intention, positive attitudes, trust, satisfaction and perceived risk. These effects (measured by 21 items) were successfully combined into a single construct using Rasch’s Model. The construct was then tested by building a website designed by experts for a fictitious clothes company. It was compared with four less-welldesigned websites created by modifying the well designed website by removing one of the four major constructs [web security; customer service; amount and quality of information provided; and usability]. These websites were surfed by 350 consumers (in five subsamples); the experts were then asked to express their perceptions and attitudes of the sites a posteriori. The association between the five websites and the 21 items was displayed visually through a perceptual map built with DYANE software. This showed that a well designed website does not always have the best effect on all 21 items measured. ß 2011 Elsevier B.V. All rights reserved.

Keywords: Mass media communications Consumer attitudes & behaviour Marketing & advertising Rasch models Web effects Perceptual map

1. Introduction Website and Internet technologies are today well established and dependable, however it is important to know what factors impact website success [5]. Previous research focused on identifying website design factors controllable by the company, which could increase online sales and result in customer satisfaction, trust, and reduce perceived risk. These factors were classified in Fig. 1. Managers, particularly those of SMEs, should be able to develop transactional websites that Internet users will visit; however, not all visits lead to purchases. We performed a test to determine how simultaneous manipulation of several ‘‘key’’ variables affected purchase intention. As in prior tests, four design variables (web security, customer service, informative content, and usability) were modified to measure their effect on purchase intention and other measures of success. Prior researchers have compared secure and non-secure websites [29], navigable and non-navigable websites [8], etc., but the idea of simultaneously considering the effect of different effects on website results is new. According to statistics from the Spanish

* Corresponding author at: Marketing Department. Dpto Comercializacio´n e Investigacio´n de Mercados., Edif. Departamental Facultad de Economia (1er piso)., Avda Tarongers, s/n., 46.022 - Valencia, Spain. Tel.: +34 96 3828312; fax: +34 96 3828333. E-mail addresses: [email protected] (N. Vila), [email protected] (I. Kuster). 1 The work on this paper was financed by the Project CTIDA 2006/350, Oficina Ciencia y Tecnologı´a (Generalitat Valenciana). 0378-7206/$ – see front matter ß 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.im.2011.04.003

National Statistics Institute and the E-Commerce and Direct Marketing Association, the most important determinant of online shopping is perceived security: nearly half of consumers in a McAffe study terminated an order or abandoned their shopping cart due to security fears. Even to get a good deal, 63% would not purchase from a website that did not display a trust mark or security policy [4]. Together with web security, it seemed likely that the amount of product information and the presence of a wide range of services were really important to the customer when shopping online. As the E-Commerce and Direct Marketing Association concluded, 80% of online buyers purchased online because they could obtain information content and additional services had been offered. Finally, the relevance of usability of the site has been noted in several works. Other indicators are: (i) more than 83% of Internet users leave a website if they feel they have to interact too many times to find a product or service; (ii) 58% of visitors who experience usability problems do not return to the site; (iii) about 60% of the time, people do not find the information they are seeking; (iv) $25 billion is lost every year due to poor website usability; and (v) the average e-commerce site could increase its sales by 100% if it had improved usability. Our work was based on constructs from IS, marketing, and psychology in an integrated theoretical framework of online consumer behaviour [13]. Specifically we focused on two major objectives. Firstly, to build a tool capable of measuring the different desired effects of a well designed website in terms of satisfaction, online trust, perceived risk, and purchase intention. Our tool was created to provide a one-dimensional measure of a well designed

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

Key factors for building a well designed website and effects Desired effects

USABILITY SYSTEM RELATED Easu to use and quick

+ Confidence + Attitude

CONTENT RELATED Images, prices, sales, catalogues

Well Website Design

+ Satisfaction (-) Perceived Risk

SERVICE RELATED Delivery, payment, refunds.

+ On line purchase intention

SECURITY

SUCCESSFUL WEBPAGE FACTORS

167

2.1. System usability-related factors for a well designed website: speed and ease of use A ‘‘systematic website structure’’ [11] must provide good online engineering performance, which involves ease of access, speed, and navigability. Although usability has also been suggested as needing with other factors such as content, interactivity, and responsiveness, we decided to consider only two dimensions: speed and ease of use [25]. We split ‘‘speed’’ into two factors: waiting time, including attitude towards it, and website delay. We measured ‘‘ease of use/navigability’’, by considering the cognitive effort required to use the website. This concept has been widely studied. In easily navigable websites, users know where they are at all times and where they need to go, because the way to do so is indicated appropriately. See Table 1 for our set of variables. 2.2. Content-related factors for a well designed website: amount and quality of information Several authors have focused on measuring whether the information provided by the online service is precise, updated, comprehensive, relevant to the user, flexible, and periodically extended or renewed.

Fig. 1. Key factors for building a well designed website and effects.

website. This construct was created using the Rasch Model to obtain measurements that do not depend on the instrument used and that provided scales that were not linked to the objects being measured, that is, that did not vary from one interviewee to another. The advantage of this method was that a joint measure of different items could be created for the same dimension or construct. Secondly, our work sought an easy way of visually displaying the similarities between the desired effects of a website and the five websites: a perceptual map was employed, using the ANAFACO technique included in DYANE software. Perceptual maps were not used because there was no simultaneous comparison of several websites. To identify which website came closest to our measured effects, four subsamples of 65 individuals each were exposed to the four manipulated websites. The subsamples were compared with each other and with a control subsample of 110 Internet users who were showed the well designed website. Our intention was to provide managers with some general recommendations on how to distribute their resources when it came to designing a transactional website, so that they did not use unnecessary effort in providing concepts that were less useful in stimulating online sales. 2. Key variables in a well designed website During the past decade, researchers have applied TAM to examine IT usage and have verified that user perceptions of usefulness and ease-of-use are key determinants of technology adoption [7,17]. This framework was based o the premise that new technologies are complex and potential users are uncertain about their success in adopting innovations. Researchers have suggested several ways to improve commercial website design; these influence consumer perceptions, attitudes, and behaviours towards the website, the company and its products. The key factors may be classified in three major blocks: system, contents, and service. We followed the suggestions of one stream of the research [6] which postulated that the perceived quality of a website depended on the user’s evaluation of these three parts and, furthermore, that the three factors influence one another; thus that is important that companies commit to them all.

2.3. Service-related factors for a well designed website: pre-purchase, during, and post-purchase Service quality and service value has received particular attention in assessing good designs, but some researchers have also started to examine on-line services and associated concepts [9,14]. We decided to measure online purchase intentions using methods used in prior work [18], considering that customer service-related success factors should include the effects of pre-, during-, and post-purchase service that may enhance the user’s perceived quality. 2.4. Web security factors for a well designed website Of all the ‘‘customer services’’, perceived security has received most attention in recent years, since a consumer will not be happy to provide personal details on websites that do not provide web security. Consumers may even be satisfied with and trust websites that do not offer any services but they are unlikely to trust and be satisfied with unsecured websites. 3. Determining factors for well website success Our work focused on websites whose aim was to invite purchase. Obviously, measuring website success in monetary terms is intuitively attractive; however, there are certain limitations and other measures could be more appropriate [10], e.g., subjective performance measures such as; (i) purchase intention [12,24,26]; (ii) attitude towards the website [1,2,22]; (iii) satisfaction; (iv) trust [3,23]; and (v) low risk [21]. Thus, a favourable website in terms of usability, amount and quality of informative content, customer service (before, during and after the transaction), and web security, will be perceived as a quality website. The site will therefore be able to provide superior levels of purchase, trust, satisfaction, and attitudes with a lower level of perceived risk. 3.1. Purchase intention Many studies have used purchase intention as an indicator of website success, considering that purchase intention is also a reflection of customer loyalty. Thus they have analysed the

168

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

Table 1 Well designed website variables. System usability Speed Easy of use/navigability

Informative content Full product information

Company information

Customer service Pre-purchase services

During purchase services

Post-purchase services

Security Security policy

Privacy policy

Waiting time, both for downloads and to reach the home page. Attitude towards waiting time and website delays. A well-structured website is perceived as intuitive, simple, consistent and attractive. In contrast, slow, complicated and disorganised websites do not encourage navigation and shopping even less so.

Product assortment, brand and model assortment. Enlarge, zoom and rotate functions for the product. New product launches, or changes in current products. Photographs of the products and from different angles so the product can really be appreciated. Price information, especially next to the products. Highlight special offers, promotions/discounts. Product information next to the photograph and not by the side, to avoid confusing the consumer. Product differentiation and comparisons. Company history. Company’s mission statement. Company notices, news, etc.

Information on availability and stock levels. Preliminary information on terms and conditions of sale, payment etc., with the chance to contact the company and obtain additional information. Preliminary information on security and privacy terms. Commitments on delivery dates and the promises made. Offer convenient times. (ii) Shorten the steps required to complete the transaction. Allow consumers to change their requirements/orders while the order is being processed. Use simple, self-explanatory forms when requesting personal data in the corresponding languages and with the appropriate cultural adaptations. Help customers during the process without claiming that the system is busy or out of service. A priority is to ensure that the correct price is calculated when an order is processed. Provide help to resolve any transaction errors. Allow consumers to review previous transactions. Accept several payment methods: credit cards, debit cards, cheques and cash (electronic or digital). Offer several delivery periods, at different product prices. Say thank you for the transaction. (iii) Introduce order tracking mechanisms. Loyalty programmes (frequent miles programmes, customer loyalty programmes and company credit cards for local countries, programmes for special members, etc.). Create customer discussion forums/chats. That is, clubs or chat rooms (members clubs, product-based clubs, chats with company employees, chats with interest groups, discussion groups, live talks etc.).

Encrypted technology. Authentication systems. Approval mechanism. Digital signatures included in the message itself. Third party insurance. Privacy policy on the home page.

Source: Own elaboration.

variables that most affect online purchase intention, with its main antecedents having been found to be the way that companies design their websites in terms of usability, information content, customer service, and security. Therefore: H1. A good website creates increased purchase intention in comparison to poor websites with less usability, information content, customer service, and perceived security. 3.2. Attitude towards the websites Several authors have found that some website elements have a positive influence on an individual’s interaction with the site. These include the amount and quality of the information content provided and convenient, interactive service that result in the user experiencing pleasure in the activity. These attributes increase perceived quality and arouse positive attitudes. Therefore, H2. A good website produces a positive attitude than one with less usability, information content, customer service and perceived security.

3.3. Satisfaction Antecedents to online satisfaction have been considered to fall into one of three groups: (i) information, (ii) system usability and (iii) customer service. Action on each of these is likely to increase customer satisfaction to the point of making the user return to the website. Therefore, H3. A good website provides greater satisfaction than a website with less usability, information content, customer service, and perceived security. 3.4. Online trust Companies need to use signals which can reduce uncertainty in situations where the consumer is unfamiliar with the products or services: then the website is the only indicator of the way the company will behave towards the consumer, so its design and contents must provoke online trust in the company. The factors affecting online trust can be grouped into the characteristics of the website and the Internet user [20]. The two variables that help to create a better website are: (1) privacy/web security (achieved by

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

statements and a seal) and (2) pleasurable usability (achieved through speed and ease of use). Thus: H4. A good website builds more online trust than a website with less usability, information content, customer service and perceived security. 3.5. Perceived risk Finally, credibility is related to online trust and can be an antecedent to it. Credibility can be defined as believability. Several authors have found that website credibility is a determining factor in reducing perceived risk of the customer [15]. Indeed, one of the main obstacles to online sales is customer perceived purchase risk. Therefore, companies should take particular care when designing their websites in order to reduce perceived risk and make online shopping pleasurable. Three terms have been used to describe credibility: information, functional credibility, delivery credibility. Therefore, H5. A good website causes the potential customer to experience less perceived risk than those - websites with less usability, information content, customer service, and perceived security). The five concepts in these five hypotheses (purchase intention, attitude, satisfaction, trust, and perceived risk) were combined into a single construct designed to measure the desirability of a good website. The 21-item construct is shown in Table 2. 4. Methodology A qualitative study was first carried out using seven focus groups; to determine the sector to be studied, establish the population for its analysis, and examine the key elements in website design. Then, a website was designed and implemented carefully for the chosen sector; this site was replicated four times and a site was then manipulated to (i) reduce its usability; (ii) reduce the quantity and quality of its information content; (iii) reduce its customer service, and (iv) reduce its perceived security. Table 2 Construct ‘‘desired effects in a well designed website’’ (desired properties in a web). Purchase intention PI.1. I would buy from this website. PI.2. I would create a personalised account with this website. PI.3. I would use my credit card to shop on this website. PI.4. I would recommend this website to other people. Satisfaction PS.1. I think I made the correct decision in using this website. PS.2. Experience of this website has been satisfactory. PS.3. I am satisfied with this website. PS.4. I am satisfied with the service provided by this website Attitudes to the website PA.1. I get on well with this website PA.2. I would like to visit this website again. PA.3: I feel comfortable navigating this website. PA.4: This site is a good place to spend my time PA.5. I consider this website to be a good site for fashion. Perceived risk PR.1. It might be risky to buy articles from this website. PR.2. It is highly likely to make a mistake in buying a product from this website. PR.3. There is a significant risk in buying a product from this website. PR.4. It is highly likely that the products I buy from this page will not meet my expectations. On line trust PC.1. This website deserves a lot of respect. PC.2. This website is honest and true. PC.3. This website seems to be sincere in its promises. PC.4. You can trust this website completely.

169

Each of the five resulting websites was then tested empirically to determine the effect of the deterioration of the four sites on the user’s satisfaction, etc. With the original site thus interviewees saw only one of the 5 websites. Finally, a quantitative study was carried out to (i) construct a one-dimensional measurement tool for the desired effects of a website, and identify which items were more closely associated with a good website. 4.1. Qualitative study Before starting our testing, we had assumed that: (1) an ecommerce website should be agile and well designed, and its content well-organised; (2), its perceived security should be high; (3) it should have plentiful, quality information content (such as sufficient information on prices) and product information content (such as varied images with rotation and zoom options); and (4) eservices (i.e. alternative methods of payment, delivery, etc.). Our first test was thus intended to learn more about the general public’s online shopping behaviour and determine what products they consumed. To achieve this, we formed seven focus groups, six consisted of six people while the seventh had eight participants, Each group was asked to discuss which variables they considered most important in the design of a website. They generally agreed that usability, web perceived security, and information content and services affect the transactions. Moreover, they pointed out other key aspects of website design (e.g., the relevance of product and price information in a catalogue), and noted that services would only be able to stimulate purchase if other variables were present. The focus groups also suggested that, after books, clothes were sold by most online sales and that online clothes availability was particularly interesting to young people, and that in Spain 15% of online shoppers bought an item of clothing. For these reasons our study focused on developing and experimenting with a clothes website directed at consumers between 18 and 35 years of age. The investigation was therefore conducted in the Spanish textile industry, which has been under pressure for decades, primarily from Asia, which has been forcing small and mediumsized textile companies to seek new ideas to help them recover their competitiveness. 4.2. Website design 4.2.1. The good website Initially, two graphic designers created a well designed fictitious website for a non existent clothing company: Resaka. The target customer was assumed to be from the segment of middle class young people. This website was structured in 6 sections shown horizontally on a menu bar that appeared on all the web pages. These sections were prepared ad hoc for the experiment: company, services, catalogue, work with us, online shopping and contact. (See the Appendix A). There was also a logo in red and black designed by the graphic designer; it also appeared on several of the website pages. Chill out style music was included on the website as being appropriate to the target public. Verisign and Visa logos appeared top right on all the pages to maintain web security conditions. The shopping form the user was requested to complete also included these logos and other measures to guarantee secure shopping. 4.2.2. Low usability website: slow and untidy Based on the well website described above and to alter usability in relation to this web, two types of changes were done, changes in

170

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

website speed and changes in ease of use. These changes were pretested with a sample of students to identify any statistical differences between both webs. Firstly, to alter the speed of the shopping process, the number of clicks was increased from ‘‘4’’ on the well website to ‘‘9’’ on the low usability website. Thus, Internet users were unable to purchase rapidly as recommended. In the website manipulated for low usability and in order to lengthen the shopping process, the click to ‘‘his fashion’’ and ‘‘her fashion’’ was removed from the main page, as was the ‘‘online shopping’’ click represented by a star. The stages required to put an item of clothing in the basket were extended to 9. This decision to increase the number of links to reduce usability is based on previous works that also have studied usability pointing out that Internet users mainly use links to navigate. Secondly, to alter ease of use, two other measures were adopted. On one side, the horizontal navigation bar on the well designed website was moved from the top to the lefthand side and the bottom. The designers were asked to make the menu bar L-shaped (in two parts) and only on the main page. On another side, the surfer was forced to move the mouse to visualise the links at the bottom of the page and it was decided to eliminate the menu from all the other pages, leaving only the Resaka logo and a start button to return to the home page. 4.2.3. Low security web To manipulate the second variable, it was decided to remove the logos and references to Verisign and Visa from all the Resaka sections. Both references were also eliminated from the shopping process (when asked to provide personal details and choose the payment method). Paragraphs were also eliminated from another three sections which made some reference to web security. Additionally, in the customer service section, the references to web security were omitted. Likewise, in the section on terms and conditions of sale, the paragraphs which referred to web security were eliminated. The online shopping section itself was also modified. Thus, when it came to choosing the payment method, all references to web security were eliminated. Finally, and given that we did not want to modify the privacy variable, it was decided to leave the link to the privacy policy only on the home page and reduce it to the minimum expression (with no allusion to web security). These changes were pre-tested with a students sample to identify any statistically significant differences between both websites. This group of changes to reduce web security was made based on a previous work [16]. These authors state that web security is perceived under the following circumstances: when there are web privacy policies (logos such as VeriSign and Visa), when not too much information is required, privacy is ensured through procedures, rules and/or legal protection on the use of information, secure technologies are used (more than firewalls) and the company is linked to credible, responsible associations (such as VeriSign and Visa). In contrast, the absence of these indicators is perceived as a lack of security and so all these indicators were eliminated as mentioned above. Yoon explains that transaction security can be achieved by VeriSign and Visa logos [28]. Thus, although other variables could be used to measure web security, these two indicators provide an intuitive, clear indication of the level of security being offered; so, by eliminating clear, decisive paragraphs explaining the company’s commitment to the customer, this study was able to proceed in line with previous works referred. 4.2.4. Low informative content website In order to manipulate the website’s informative content, it was decided to remove the section on the product catalogue from the home page. This meant that the user could not get an idea of how the garments look in real scenarios, possible combinations or

recommended contexts. These changes were pre-tested with a sample of students to identify any statistically significant differences between both websites. The garment option in the online shopping section was also eliminated as was the back view of the garments so that the potential shopper could only see the garment in one size, one colour and from the front. The informative paragraph on the materials, colour etc., for each garment was also omitted. Price information was removed from the home page so that the user had to click on each garment to find out the price. Also, in the case of products at reduced prices, references to the previous price were eliminated so that the user could not estimate the saving. All this meant that the user who went on to the website manipulated for low informative content would find a site where only one image and one colour were shown for each garment and it was not possible to enlarge, rotate or see how it would look in other colours. Also prices were not indicated clearly or immediately, nor any possible savings. 4.2.5. Low service level website To manipulate customer service, the ‘‘customer service’’ link was completely removed from the home page. This removed the suggestions box, Resaka members club, the explanation on exchanges and refunds and the section, your purchases. These changes were pre-tested on a sample of students to identify any statistically significant differences between both websites. A second link, ‘‘contact’’, was also removed from the main menu and the contact telephones and e-mails were removed from the other pages on the site. The information on the company headquarters included in the well web under the link ‘‘contact’’ was moved to the link ‘‘company’’ also on the home page. All references to customer service were removed from the ‘‘terms and conditions of sale’’ section. Some images of the alternative manipulations of the well website are available in Appendix A. 4.3. Quantitative study Our experiments are supported by the opinions of 350 interviewees who agreed to participate in exchange for a pen drive (USB) worth 15 euros. The interviewees were contacted in two underground stations and invited to take part in the experiment. After accepting the invitation, they were taken to a computer room to surf the website of the fictitious company, Resaka. Total sample composition was 57% women and 43% men. 40% had 3 year University studies and 48% had a secondary education. More than 50% of the interviewees said they earned more than 2000 euros a month. The age range varied between 18 and 25 years. Of the total sample, 110 were exposed to the well website, while groups of 65 interviewees were exposed to each of the four website manipulations described in the above section. They were all told that they had a fictitious cheque for 200 euros which they could spend shopping on the website. After 30 min they were given a questionnaire to reflect their opinions after surfing. 4.3.1. Output measurement scales To measure the different concepts in the model, several 7-point Likert scales were used based on proposals in the literature (Table 2 shows the items considered in detail and the authors that support each scale). 4.3.2. Statistical methodology Firstly, in pursuit of this work’s first objective, we applied the Rasch Model using Linacre’s Winsetp programme version 3.60.01. This programme is increasingly enabling objective measurements in the Business Management sphere given that the Rasch Model

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

does not work with a statistical paradigm (eliminate items from a scale which presents problems) but with a measurement paradigm (consider the possible elimination of the subjects interviewed whose opinion is distorting scale validity). With this second paradigm, the data fit the model rather than the other way round. The great advantage is that it does not compromise scale content validity by beginning to eliminate items which, with a different interviewee sample, would be perfectly valid. By applying this model, the ideal is to transform the scores obtained into objective measurements which are presented in logit units for comparative analysis under the one-dimensionality hypothesis. However, a criticism of the Rasch model is that it is overly restrictive or prescriptive because it does not permit each item to have a different discrimination [27]. In spite of this, Rasch model is one of the psychometric models that can help researchers create measures. The most famous application of Rasch model can be found in large education data sets, such as TIMSS and NELS, in the forms of student academic achievement scores. But Rasch model can also create attitudinal scales, such as student engagement level or liking of a particular topic. In this sense, Winsteps software is a computer program that has been developed to calibrate item and person parameters in the family of Rasch models. WINSTEPS begins with a central estimate for each person measure, item calibration, and rating scale category structure calibration. An iterative version of the PROX (normal approximation) algorithm is used to reach a rough convergence to the observed data pattern. The unconditional maximum likelihood method, also called the joint maximum likelihood (JML) method, is then iterated to obtain more exact estimates, standard errors, and fit statistics. Secondly, and to achieve the second objective, perceptual maps were obtained which could represent two groups of stimuli by points in the same geometric space i.e. the subjects (which in the study are the five segments of interviewees exposed to 5 different websites), and the 21 items (which make up the successful website construct). Proximity of two points on the perceptual map signifies a close association between them. Thus, it is to be expected that the point which represents the ideal subject (well website) will appear close to the 21 items which represent the construct and points which are distant will indicate a remote association. This tool is very useful because it represents the items and stimuli (websites in this case) visually in the form of points. A small distance on the map between a website and an item reflects a close association between both. This perceptual map was obtained using DYANE version 3. This software is similar to other software such as SPSS, that also perform correspondence factor analysis. This analysis detects associations and oppositions existing between subjects (webs in our study) and variables (success indicators in our study), measuring their contribution to the total inertia for each factor. The projection of the subjects and the variables onto the same set of factorial axes enables a two-dimensional graph to be drawn which assists in the interpretation of the results. 5. Discussion 5.1. Developing a tool to determine website success The main objective was to provide a measure of the desired effects of a website, or in other words, an instrument for measuring website success. As explained above, we have argued that this success will comprise the concepts of favourable attitude towards the website, satisfaction, trust, low perceived risk and purchase intention. The information obtained from 350 interviewees who took part in this research was analysed using the Rasch Model. In this model, the latent variable ‘‘website success’’ is conceived of as a 21-item

171

Fig. 2. Map of items and individuals interviewed.

concept measured on a 7-point Likert scale, as justified above. The Rasch model was applied to the entire sample rather than on each experiment to obtain valid results with a better fit, lower estimation error and across all the scenarios analysed. However, for experimental purposes, the model was applied to each subsample independently, and practically equivalent results were obtained. The likelihood ratio test between the estimations for the whole sample and each of the 4 manipulated subsamples shows much symmetry. The Rasch model can be used to establish a hierarchy of website success concepts, defining their presence among the participants in the study (Fig. 2). Additionally, as Fig. 2 shows the most relevant items for measuring success are at the bottom of the map. In fact, this model was applied in the search for a measure which would enable comparison of a set of heterogeneous factors associated to website success. Table 3 shows the variables in relation to the value obtained in the measure column with the items ordered from least to most important. Thus, the

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

172 Table 3 Measure of website success (21 variables). Entry number

Raw score

Count

Measure

PA.4 PI.3 PI.2 PR.4 PA.1 PR.1 PI.1 PC.1 PR.2 PR.3 PC.4 PA.2 PS.3 PS.1 PA.5 PS.4 PC.2 PI.4 PC.3 PS.2 PA.3 Mean S.D.

1290 1351 1364 1478 1496 1500 1515 1540 1556 1564 1615 1616 1631 1652 1657 1689 1692 1697 1759 1771 1774 1581.3 133.5

350 350 350 350 350 350 350 350 350 350 350 350 350 350 350 350 350 350 350 350 350 350.0 0.0

0.59 0.47 0.45 0.22 0.19 0.18 0.15 0.10 0.06 0.05 0.06 0.06 0.10 0.14 0.15 0.23 0.23 0.24 0.39 0.42 0.43 0.00 0.28

Model S.E.

Infit

0.04 0.04 0.04 0.04 0.04 0.04 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.00

1.32 1.10 1.10 1.70 0.75 1.62 0.96 0.79 1.52 1.47 0.68 0.86 0.83 0.68 0.96 0.85 0.71 0.94 0.76 0.72 0.69 1.00 0.32

MNSQ

Outfit ZSTD 4.2 1.3 1.4 8.0 3.7 7.2 0.6 3.0 6.1 5.5 4.9 1.9 2.3 4.8 0.4 2.0 4.3 0.7 3.4 4.0 4.5 0.3 4.1

MNSQ 1.48 1.10 1.10 1.75 0.77 1.69 0.96 0.82 1.56 1.53 0.66 0.84 0.83 0.67 0.96 0.97 0.71 0.94 0.73 0.69 0.66 1.02 0.35

PTMEA corr. ZSTD 5.8 1.4 1.3 8.4 3.4 7.8 0.5 2.6 6.4 6.2 5.1 2.2 2.4 4.8 0.5 0.4 4.2 0.8 3.8 4.4 4.9 0.1 4.4

0.60 0.69 0.66 0.40 0.75 0.36 0.69 0.69 0.41 0.40 0.70 0.75 0.61 0.68 0.68 0.60 0.67 0.70 0.66 0.68 0.75

Exact match OBS%

EXP%

27.7 32.0 34.0 24.0 40.6 24.9 38.6 38.0 29.7 27.1 41.7 32.6 39.1 45.1 32.6 42.9 44.6 35.1 40.6 44.6 44.0 36.2 6.7

30.4 31.2 31.4 32.0 32.2 32.0 32.1 32.5 32.7 32.7 33.4 33.5 33.8 34.0 34.0 34.5 34.5 34.5 35.5 35.6 35.8 33.2 1.5

Item

PA.4 PI.3 PI.2 PR.4 PA.1 PR.1 PI.1 PC.1 PR.2 PR.3 PC.4 PA.2 PS.3 PS.1 PA.5 PS.4 PC.2 PI.4 PC.3 PS.2 PA.3

Entry number. 21 proprieties/items of a successful website. Count. Total sample of interviewers. MNSQ. Mean-square fit statistics show the size of the randomness, i.e. the amount of distortion of the measurement system. Values greater than 1.0 indicate unpredictability (unmodeled noise, data underfit the model). Statistically, mean-squares are chi-square statistics divided by their degrees of freedom. Mean-squares are always positive. ZSTD. Standardized fit statistics are t-tests of the hypothesis ‘‘Do the data fit the model (perfectly)?’’ These are reported as z-scores, i.e. unit normal deviates. They show the improbability of the data, i.e. its significance, if the data actually did fit the model. More than 2.9 indicate lack of predictability. Standardized values are positive and negative. PTMEA. Correlation of each point with the total scale.

most relevant variables for measuring website success are: PA.3: ease of navigating the website, PS.2: satisfactory experience of navigating the website, PC.3: sensation of sincere promises to customers, and PI.4: recommendation of this website to others. The least relevant items for measuring website success proved to be: PA.4: good place to spend time, PI.3: use of credit card to shop on this website, PI.2: creation of a personalised account on this website, and PR.4: likelihood that the products purchased do not satisfy expectations. However, none of the items drop in the analysis, so all of them have to be considered to measure success. The items related to perceived risk may be comparatively less important in this scale because they have a negative sense and they are measured on the opposite side (less perceived risk, higher success). Table 3 shows the validity of the variables in the proposed measurement instrument. As can be seen, MNSQ statistical values are between recommended values 0.5–1.5 and PTMEA correlations reach high positive values. PTMEA values must also move in a similar interval (above 0.5), indicating in this case that the measurement of the item contributes to the whole measure. However, not all Zstd statistic values exceed 2 in absolute terms, advising a more detailed study on measurement reliability and validity because values below 2 indicate over fit and lower variability of data. In any case, Rasch’s probabilistic model allows us to determine measurement reliability and validity. Results also confirm satisfactory values related to item and interviewee measurement reliability. Therefore, both interviewee responses and the measures used to measure website success are consistent and stable. Additionally, information on measurement validity support that a good level of global fit has been obtained as the infit and outfit values for both statistics (MNSQ and ZSTD) are adequate. Therefore, the proposed measurement instrument enables website success to be evaluated (i.e. the construct measures what it was intended to measure). For the first statistic (MNSQ), values should vary between 0.5 and 1.5. For the second one (ZSTD), values should

be 2. Both statistics are used to examine the validity of each item. Given that the values obtained for both move around recommended values, this means the item does not significantly deviate from the Guttman assumption of the Rasch model. Thus, the INFIT AND OUTFIT statistics for the 21 items indicate that all of them meet this criterion. So, all of them have to be considered to measure website success with a valid instrument.’’ Finally, it is interesting to check instrument one-dimensionality and invariance, as these are the underlying hypotheses in the Rasch Model. In this regard, explained variance is 54.1% which, together with the high reliability levels and high PTMEA values, shows that the one-dimensionality hypothesis required by the model has been fulfilled. The variance value, however, also indicates that other concepts not contemplated a priori in the measurement model could be included. In that regard and in relation to the invariance test, the five websites have been compared with each other for each of the measurement instrument items. The results show that only 17 out of the 120 relations analysed obtain significant differences (t values >2 and probability less than 0.05). This implies that this significant difference is only present in certain pairings and for certain variables. In view of the above, the proposed instrument is appropriate, reliable and valid for measuring website success from the consumer’s point of view. 5.2. Factor analysis: 5 websites and 21 items (effects) Having built the tool to measure the construct ‘‘desired effects of a well website’’, Correspondence Factor Analysis was applied to achieve the second objective. This analysis shows to what extent the different items are closest to each of the 5 website pages used in the study. Thus, it is also possible to verify the hypotheses by examining the relative distance between the well website and the different items. Correspondence Factor Analysis shows that two dimensions explain a high variability of the data (0.843). So, the perceptual map that visually represents both dimensions can be

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

173

Table 4 Dimension contributions to website inertia and effect inertia (OTR). Websites analysed (stimulus)

Horizontal axis of the map factor 1 (OTR%)

Vertical axis of the map factor 2 (OTR%)

Well designed website Non-secure website No customer service website Non-informative website Non-usable website

0.6095 0.0150 0.0440 0.3676 0.0170

0.0089 0.6413 0.3168 0.0090 0.0321

Website effects (desired properties) Purchase intention PI.1. I would buy from this website. PI.2. I would create a personalised account with this website. PI.3. I would use my credit card to shop on this website. PI.4. I would recommend this website to other people. Attitudes to the website PA.1. I get on well with this website PA.2. I would like to visit this website again. PA.3: I feel comfortable navigating this website. PA.4: This site is a good place to spend my time PA.5. I consider this website to be a good site for fashion. Perceived risk PR.1. It might be risky to buy articles from this website. PR.2. It is highly likely to make a mistake in buying a product from this website. PR.3. There is a significant risk in buying a product from this website. PR.4. It is highly likely that the products I buy from this page will not meet my expectations. Satisfaction PS.1. I think I made the correct decision in using this website. PS.2. Experience of this website has been satisfactory. PS.3. I am satisfied with this website. PS.4. I am satisfied with the service provided by this website On line trust PC.1. This website deserves a lot of respect. PC.2. This website is honest and true. PC.3. This website seems to be sincere in its promises. PC.4. You can trust this website completely.

Horizontal axis (OTR%)

Vertical axis (OTR%)

0.0262 0.0048 0.0040 0.0000

0.1798 0.0000 0.0602 0.0808

0.0929 0.0591 0.0020 0.3933 0.0430

0.0590 0.0150 0.0364 0.0013 0.0540

0.1102 0.0749 0.0788 0.0288

0.0713 0.0019 0.0006 0.0002

0.0129 0.0030 0.0110 0.0213

0.0010 0.0510 0.1374 0.0040

0.0081 0.0176 0.0272 0.0249

0.1702 0.0575 0.1263 0.0092

The numbers in the cell represent the contribution of each point in the map (website or property) to the inertia of the factor/dimension. If a point has a higher value, it explains better the variance of the factor/dimension. So, it gives a positive or negative direction to the factor. The higher values have been bolded. This means that the horizontal axis better explains Well designed website and Non-informative website, while the vertical axis is powerful to explain the other three websites.

used to accept/reject our hypotheses as the variance explained (0.8043) is higher than the recommend value (0.75). Secondly, it is possible to analyse to what extent both dimensions explain the behaviour of the 5 websites. As Table 4 shows, the positive part of the horizontal axis explains the behaviour of the well website (OTR = 0.6095) against the noninformative website (OTR = 0.3676) and to a lesser extent the non usable website (OTR = 0.170). Both are on the negative side. Seeing what variables group with each of these two blocks, shows that the well website is associated with most of the variables concerning less perceived risk (PR.1, PR2 and PR.4). This website inspires also short term satisfaction (PS.4, PS.3, PS.2, PS.1). In contrast, the non-informative website and to a lesser extent the non-usable website are associated with long term pleasure attitudes to the website (PA.4, PA.2 and PA.1) and with educed perceived risk variables (PR.3). Therefore, the horizontal axis could be called ‘‘long term pleasure versus short term satisfaction’’. The positive part of the second dimension, or vertical axis explains the behaviour of the no service website (OTR = 0.33) and, to a lesser extent, the non-usable website (OTR = 0.321). They are both in contrast to the non-secure website (OTR = 0.6413). Specifically, it can be seen that no service website obtained a high average evaluation in purchase intention (PI.1, PI.2, PI.3 and PI.4). The non-secure website is linked to trust (PC.1, PC.2, PC.3 and PC.4) and non-usable website to positive attitudes (PA.3, PA.5). Therefore, the vertical axis could be called ‘‘action versus feeling’’, which is the same as ‘‘purchase intention versus trust’’. To summarise then, and as shown visually in Fig. 3 and numerically in the above Table 4, the considered websites seem to

emphasise different variables. It should be remembered that the names of the websites correspond to the variable that was manipulated to get less customer service, product/firm informative content, web security and usability. The well website was named ‘‘WELL’’ because it kept the highest levels of customer service, product/firm informative content, web security and usability according to previous literature. The variables that best described each web are the following: - The no service website is the one most associated with a high purchase intention (actions) (PI.1, PI.2, PI.3 and PI.4). So, a website could increase purchase intention even when low customer services are offered. Therefore, we must reject H1, since the well website does not have the greatest purchase intention in relative terms. - The non-informative website is the one most associated with positive attitudes (long term pleasure) (PA.4, PA.2 and PA.1). Also, the non-usable website develops positive attitudes (PA.5, PA.3 and PA.1). We must therefore reject H2 since in relative terms the well website does not produce the most positive attitudes to the website. - The non-secure website is the one most associated with online trust (feelings) (PC.1, PC.2, PC.3, PC.4). We must therefore reject H4 since in relative terms the well website does not have the greatest online trust in relative terms. - The non-usable website is associated with a mix of items. It offers positive attitudes (PA.5, PA.2 and PA.1), although it is also perceived low risk (PR.3); with purchase intention (PI.1) and satisfaction (PS.2). In other words the non-usable website does not emphasise any property, although it is the one which comes

174

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

Fig. 3. Representation of the five websites and the desired effects.

closest to the majority of them. Its assessment is the most varied, as shown by its proximity to the intersection between the coordinate axes. - The well website is the one most associated with low perceived risk (PR.1, PR2.2, PR,4) and with satisfaction (PS.1, PS.2, PS.3, PS.4). Therefore, we can accept H3 as the well website does stand out in satisfaction and also H5, as a lower level of perceived risk is associated with the well website. In view of the results and verification of the hypotheses, it is not possible to state that the well website design guarantees greater success than websites which omit certain characteristics. Thus, the well website does not always have a greater effect than another nonwell website in all the desired effects. These results do not mean that the interviewees do not perceive differences between website, quite the opposite in fact, they perceive very big differences, but they do not link the well website to all the desired effects. This may be due to the fact that the ‘‘well website’’ concept varies from public to public so that not all consumers expect a website to be simultaneously easy and convenient to navigate, very secure, with varied, plentiful information and a high level of customer service. These results must be understood in the context of a particular industry (clothing), for a particular segment (young people) and in a specific country (Spain). So, some website design factors could be more valuable in relative terms than others: (i) in other industries (i.e. website services for leisure industries; or website security for financial services), (ii) other publics (i.e. security for mature

markets) and (iv) other countries (i.e. informative websites for Germany). Furthermore, the spatial representation technique used determines that the stimuli (the websites analysed) and the items are spread over the two-dimensional space in relation to the association coefficients (frequencies) between each stimulus and each item. So, a relative map is created considering different variables (properties) and different websites at the same time. This means that results should be understood in relative terms as websites and items are grouped together based on their higher proximities (expressed by interviewees during the interviews). Thus, the fact that a property is not associated to a website, does not mean that the website does not have that property, but that in comparative terms it is less close to that property than other websites which are associated to the property. Taking these explanations into account, our results show that a well website design is mainly associated to satisfaction and low perceived risk. However, when surfers feel satisfied and perceive low risk, other consequences may start to emerge. Satisfaction and low risk can therefore be viewed as two main requirements for obtaining further results. 6. Conclusions, implications, limitations and future lines for research This study was intended to achieve two major objectives. Firstly, to develop a measurement instrument which would

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

determine website success and unite different concepts associated to this success. Secondly, and following guidelines in the literature, to check if certain website characteristics are essentially relevant and determining factors for success. To achieve both objectives an well website was designed for a hypothetical textile company called Resaka which contemplated the characteristics indicated in the literature as being necessary for successful online sales. Four other versions of this well website were also created, eliminating a different characteristic from each. Information was collected from 350 Internet users between 18 and 35 years old. The first objective has provided two major conclusions. Firstly, the Rasch Model has proved to be an appropriate procedure for obtaining reliable, valid measurement scales. Secondly, an instrument able to evaluate website success has been designed which reflects different success-related aspects: low perceived risk, attitude, trust, satisfaction and purchase intention. Achieving this objective provides companies with useful information for a better understanding of the key dimensions to successful online sales. However, when this instrument is broken down and represented on a perceptual map (considering the 21 items) the idiosyncrasy of the spatial representation technique disperses the construct components in the space. Thus, the different websites are associated to the items which are closest to them on the map. This means that the map yields relative results (the websites are compared with the 21 items), but this does not invalidate the construct as an absolute tool for measuring website success. In relation to the second objective, it has been found that the different websites analysed produce very different effects. The results show that a well website in terms of web security, customer service, informative quality and usability does not have the best effects in absolute terms. Our study shows that the well website is outstanding in items concerning short term satisfaction and low perceived risk. Thus, a technically perfect website, with good web security indicators and plenty of informative elements satisfies the user. In contrast, this well website does not carry higher levels of purchase intention (actions), trust (feelings) or long term pleasure attitudes. The results show that the construct ‘‘desired effects’’ is not uniformly present in an well website. In other words, all the desired effects do not associate with a website which in principle is well designed in terms of web security (with logos and certifications), customer service (with links on periods, refunds, guarantees, contact, etc.), informative quality (with catalogues, colour options, zoom, offers, prices etc.), and usability (with horizontal navigation bars and fast screens where shopping is fast with few clicks). The following managerial implications may be inferred. Firstly, it is not possible to speak of a well website design which on its own is able to bring about all the desired effects. Consumer preferences vary widely, so that what some Internet users value does not coincide with what others value. This study only shows that a website which is as perfect as possible according to the literature will bring more satisfaction and less perceived risk than others. Therefore, if the company objectives include creating an image, achieving satisfied customers and credibility in the market, they should follow the guidelines for building well websites in terms of web security, customer service, informative content and usability.

175

This would make sense for textile companies which develop websites which are not strictly oriented at sales, although they may well supplement their physical channel with the opportunities offered by the new technologies. Secondly, if the manager’s objective is short term clothing online sales among young people, then account must be taken of the fact that the offer of online services is given a comparatively worse evaluation than other web design variables. In effect, our results show that to improve sales in the textile sector among young people, comparatively more care must be taken over usability (quick and easy to use), security (no risk to buy on line) and information (sizes, colours, and catalogue) than in detailing all the shopping services being offered. Therefore, in business situations where resources are scarce, young online clothes shoppers tolerate service deficiencies better than other missing variables (security, information and usability). This does not mean that services should be forgotten, just that they are less helpful in short term online sales in this particular sector, for this kind of public and in this specific country. Therefore, this implication must be understood in the context of the sector being analysed (textile) where the object of exchange is a tangible product. In other areas (such as leisure, tourism, finance, etc.), the implications may be different to those obtained in this study. Finally, and following previous works [19], it has been shown that appropriate layout can increase usability; product information and variety of product presentations can increase website informative content; an optimal range of services can increase website online services, and the presence of some security indicators can enhance website security. So, an appropriate balance of these elements must be studied for each industry, each country and each kind of public in order to improve website success (satisfaction, buying intention, pleasure attitudes, low perceived risk and trust). In sum, and as a result of the above, companies cannot pretend to be everything for everybody. The results show that young clothing shoppers who value web security are less interested in the service offered or in usability. That is why, when designing an well website, although there should be an optimum balance between the different design variables, the company must know a priori the characteristics of its target public in order to develop an informative sales channel appropriate to their requirements. For that reason, a market study would be useful to provide informative content on the values and needs of the company’s majority public, before the company’s website is left to the designers and experts. The limitations of this study include the fact that it concentrates on the Spanish textile sector which is a mature sector, and so the conclusions cannot be generalised to other sectors with other characteristics. A cross-cultural comparison to analyse the different manipulations in different cultural contexts would be interesting. Likewise, future studies could contemplate another type of statistical methodology, for example, multi-sample structural equation models capable of analysing cause-effect relationships between several concepts considered simultaneously for different segments/websites. Additionally, our model could be incorporated into larger models describing on-line intentions and behaviours. For example, a comprehensive model of on-line purchasing might include factors such as personal characteristics which also can affect perceptions and behaviours.

176

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177

Appendix A

References [1] R. Agarwal, V. Venkatesh, Assessing a firm’s web presence: a heuristic evaluation procedure for the measurement of usability, Information Systems Research 13 (2), 2002, pp. 168–186. [2] A.M. Aladwani, An empirical test of the link between web site quality and forward enterprise integration with web consumers, Business Process Management Journal 12 (2), 2006, pp. 178–190. [3] F. Belanger, J.S. Hiller, W.J. Smith, Trustworthiness in electronic commerce: the role of privacy, security, and site attributes, Journal of Strategic Information Systems 11, 2002, pp. 245–270. [4] H. Chang, W. Chen, Consumer perception of interface quality, security, and loyalty in electronic commerce’’, Information & Management 46 (7), 2009, pp. 411–417. [5] C.L. Corritore, B. Kracher, S. Wiedenbeck, On-line trust: concepts, evolving themes, a model, International Journal of Human-Computer Studies 58, 2003, pp. 737– 758. [6] J. Cox, B.G. Dale, Key quality factors in website design and use: an examination, International Journal of Quality & Reliability Management 19 (7), 2002, pp. 862– 888.

[7] T. Feng-Cheng, L. Ming Shing, Ch. Chih-Chung, Yao-Sheng Hsu, An extension of financial cost and TAM Model with IDT for exploring users’ behavioural intentions to use the crm information system, Social Behavior and Personality: An International Journal 37 (5), 2009, pp. 621–626. [8] C. Flavia´n, M. Guinalı´u, R. Gurrea, The role played by perceived usability, satisfaction and consumer trust on website loyalty, Information & Management 43 (1), 2006, pp. 1–14. [9] D. Hackman, S.P. Gundergan, P. Wang, K. Daniel, A service perspective on modelling intentions of on-line purchasing, Journal of Services Marketing 20 (7), 2006, pp. 459–470. [10] E.K.R.E. Huizingh, The antecedents of website performance, European Journal of Marketing 36 (11/12), 2002, pp. 1225–1247. [11] Y.J. Kim, T.-I. Eom, J.H. Ahn, Measuring IS service quality in the context of the service quality-user satisfaction relationship, Journal of Information Technology Theory and Application (JITTA) 7 (2), 2005, pp. 53–70. [12] H. Ko, C. Cho, M.S. Roberts, Internet uses and gratifications: a structural equation model of interactive advertising, Journal of Advertising 34 (2), 2005, pp. 57–70. [13] M. Koufaris, Applying the technology acceptance model and flow theory to online consumer behavior, Information Systems Research 13 (2), 2002, pp. 205–223.

N. Vila, I. Kuster / Information & Management 48 (2011) 166–177 [14] G.G. Lee, H.F. Lin, Customer perceptions of e-service quality in online shopping, International Journal of Retail & Distribution Management 33 (2/3), 2005, pp. 161–176. [15] D. Maditinos, L. Sarigiannidis, E. Dimitriadis, The role of perceived risk on Greek internet users’ purchasing intention: an extended TAM approach, International Journal of Trade and Global Markets 3 (1), 2010, pp. 99– 114. [16] C.N. Madu, A.A. Madu, Dimensions of e-quality, International Journal of Quality & Reliability Management 19 (3), 2002, pp. 246–258. [17] J. Moon, Y. Kim, Extending the TAM for a World-Wide-Web context, Information and Management 38, 2001, pp. 217–230. [18] S. Otim, V. Grover, An empirical study on web-based services and customer loyalty, European Journal of Information Systems 15, 2006, pp. 527–541. [19] J.W. Palmer, Web site usability, design, and performance metrics’’, Information Systems Research 13 (2), 2003, pp. 151–167. [20] P. Palvia, The role of trust in e-commerce relational exchange: a unified model, Information & Management 46 (4), 2009, pp. 213–220. [21] J. Park, S.J. Lennon, L. Stoel, On-line product presentation: effects on mood, perceived risk, and purchase intention, Psychology & Marketing 22 (9), 2005, pp. 695–719. [22] G.M. Rose, J. Lees, M.L. Meuter, A refined view of download time impacts on econsumer attitudes and patronage intentions toward e-retailers, International Journal of Media Management 3, 2001, pp. 105–111. [23] V. Shankar, G. Urban, F. Sultan, Online trust: a stakeholder perspective, concepts, implications, and future directions, The Journal of Strategic Information Systems 11 (34), 2002, pp. 325–344. [24] A. Schlosser, T. White, S.M. Lloyd, Converting website visitors into buyers: how website investment increases consumer trusting beliefs and on line purchase intentions, Journal of Marketing 70 (2), 2006, pp. 133– 148. [25] P. Van Schaik, J. Ling, The effects of graphical display and screen ratio on information retrieval in web pages, Computers in Human Behaviour 22, 2006, pp. 870–884. [26] T. Verhagen, W. van Dolen, Online purchase intentions: a multi-channel store image perspective, Information & Management 46 (2), 2009, pp. 69– 150. [27] W.-C. Wang, C-T.C. Chen, Item parameter recovery, standard error estimates, and fit statistics of the Winsteps program for the family of Rasch models, Educational and Psychological Measurement 65 (3), 2005, pp. 376– 404. [28] S. Yoon, The antecedents and consequences of trust in online-purchase decisions, Journal of Interactive Marketing 16 (2), 2002, pp. 47–63. [29] Ziane, S., Bacha, H. (2006). Availability and Security for Complex Enterprise web Services. The Business Review, Cambridge, 5(1), 25–329

177

Natalia Vila is a Professor in Marketing in the Department of Marketing in the Economics School, University of Valencia. Her research interest include competitive positioning, identification of strategic groups/competitive groups and the application of Multidimensional Scaling in marketing. Her research work has been published in European Journal of Marketing; International Marketing Review; Marketing Intelligence and Planning; Journal of Consumer Marketing; Journal of Food Products Marketing; Journal of Relationship Marketing; Innovative Marketing; Journal of Marketing Management; Journal of Strategic Marketing; European Journal of Innovation Management; Qualitative Market Research: An International Journal; Journal of Euromarketing; The Marketing Review; Journal of Global Marketing; Journal of Travel and Transport Marketing; Sex Roles, Equal Opportunites International, Journal of Teaching in International Business; Neural Computing & Applications, Journal of Marketing Trends, Micro & Macro Marketing, Multicultural Education & Technology Journal, Quality & Quantity, Tourism Economics, Innovar, International Journal of Quality & Reliability Management, Innovation Management: Policy and Practice, among others. She has presented papers at several Conferences such as EMAC, AM or AMS.

Ines Kuster is Professor in Marketing in the Department of Marketing in the in the Economics School at the University of Valencia. Her research attention has focused on the areas of strategic marketing and sales. She has published articles in several refereed journals (i.e. JQR; European Journal of Innovation Management; Journal of Business and Industrial Marketing; Innovative Marketing; Qualitative Market Research: An International Journal; European Journal of Marketing; The Marketing Review; Marketing Intelligence and Planning; Journal of Global Marketing;, Journal of Relationship Marketing; Annals of Tourism Research; Sex Roles; Equal Opportunites International, Journal of Teaching in International Business; Neural Computing & Applications, Journal of Marketing Trends, Micro & Macro Marketing, Multicultural Education & Technology Journal, Quality & Quantity, International Journal of Quality & Reliability Management, Sex Roles, Innovar, Innovation Management: Policy and Practice among others). She is author of several books in the sales field. She has presented papers at several Conferences such as EMAC, AM or AMS. She collaborates with several companies, helping them in marketing areas (recruiting salespeople, training sales managers, analysing commercial efforts, etc.).