This article was downloaded by: [b-on: Biblioteca do conhecimento online UP]On: 22 April 2014, At: 21:34Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Internet CommercePublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/wico20

An Examination of Attributes AffectingConsumers' Perceptions of E-tailerQualityRose Sebastianelli a & Nabil Tamimi aa Department of Operations and Information Management ,Kania School of Management, University of Scranton , Scranton ,Pennsylvania , USAPublished online: 04 Dec 2013.

To cite this article: Rose Sebastianelli & Nabil Tamimi (2013) An Examination of Attributes AffectingConsumers' Perceptions of E-tailer Quality, Journal of Internet Commerce, 12:3, 268-283, DOI:10.1080/15332861.2013.859039

To link to this article: http://dx.doi.org/10.1080/15332861.2013.859039

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

An Examination of Attributes AffectingConsumers’ Perceptions of E-tailer Quality

ROSE SEBASTIANELLI and NABIL TAMIMIDepartment of Operations and Information Management, Kania School of Management,

University of Scranton, Scranton, Pennsylvania, USA

An experimental task is carried out in which participants ranke-tailers described in terms of five attributes (reputation of retailer,site usability, security, delivery, and customer support). Conjointmodels estimated at the individual level reveal that the most impor-tant attribute to perceptions of e-tailer quality is security, followed bysite usability and reputation of retailer. Agglomerate hierarchicalclustering, based upon the relative importance of attributes to part-icipants, is used to group individuals into different segments. A pro-file for each segment is provided in terms of demographic andbehavioral customer characteristics.

KEYWORDS attribute importance, conjoint analysis, e-tailerquality, segmentation

INTRODUCTION

The growth in electronic commerce (e-commerce) has been extraordinary,with many of today’s business transactions being conducted online. Thelatest report published by the U.S. Census Bureau estimates retaile-commerce sales for the first quarter of 2013 at $61.2 billion, an increaseof 2.7% over the fourth quarter 2012 (total retail sales increased only 1.1%over this same period). Compared to the first quarter of 2012, this estimaterepresents an increase of 15.2%, while total retail sales increased by 3.7%.Moreover, e-commerce sales in the first quarter of 2013 accounted for5.5% of total sales (http://www.census.gov). Consequently, determiningwhat creates a quality experience for online shoppers is paramount for

Address correspondence to Dr. Rose Sebastianelli, Kania School of Management,University of Scranton, 423 Brennan Hall, Scranton, PA 18510, USA. E-mail: rose.sebastianelli@scranton.edu

Journal of Internet Commerce, 12:268–283, 2013Copyright # Taylor & Francis Group, LLCISSN: 1533-2861 print=1533-287X onlineDOI: 10.1080/15332861.2013.859039

268

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

electronic retailers (e-tailers) attempting to meet and=or exceed customerexpectations and stay competitive in an evermore crowded virtualmarketplace.

Examining quality in the e-commerce arena (e-quality) has been an areaof research interest for more than a decade. While early studies generallyemphasized conceptualizing e-quality using multi-dimensional frameworks,a number of researchers have approached this issue empirically by developingmulti-item scales to measure e-quality constructs. The common theme amongthis body of work has been the identification and evaluation of attributes (ordimensions) affecting online consumers’ perceptions of e-quality and=orsatisfaction with e-commerce. The current study, while within the context ofthis line of inquiry, focuses on determining the effects of selected e-tailerattributes on consumers’ judgments of e-tailer quality. Conjoint analysis, amethod long used in marketing research studies (e.g., Green and Rao 1971)but relatively new to e-commerce (e.g., Schaupp and Belanger 2005; Chen,Hsu, and Lin 2010) is used. Conjoint analysis provides the means to decomposeoverall judgments into ‘‘part-worths’’ that show the relative contributions ofeach attribute to the response variable (i.e., perceived quality of the e-tailerwebsite). A documented advantage of this approach is that it forces respon-dents to ‘‘trade-off ’’ among attributes in developing their overall evaluations.In this way, the relative importance of each attribute can be discerned in amorerealistic decision-making scenario (when asked to rate the importance ofattributes directly, respondents have a tendency to rate them all as important).

In this article, researchers present the results of an experimental study inwhich participants provide overall judgments (rankings) of e-tailer qualitydescribed in terms of five attributes (reputation of retailer, site usability, secur-ity, delivery, and customer support) shown to be salient in previous research.These rankings serve as the basis for estimating individual-level conjointmodels fromwhich the relative importance of attributes is derived. Researchersuse cluster analysis, based on the relative importance of these attributes, togroup individuals into different segments. They provide a profile of each seg-ment, not only in terms of its perceptions of how these attributes influencee-tailer quality, but also in terms of demographic and behavioral customer char-acteristics. Linking perceptions of e-quality to observable customer characteris-tics has implications for e-tailers wishing to develop effective, targetedstrategies for improving the quality of online shopping experiences.

LITERATURE REVIEW

While there has been a proliferation in research on quality in e-commerceand the closely related issues of customer satisfaction and purchase intentionin e-tailing, the researchers focus on reviewing studies that deal primarilywith identifying salient attributes (or dimensions). These studies can beclassified broadly into those that are conceptual (providing multidimensional

Consumers’ Perceptions of E-tailer Quality 269

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

frameworks) and those that are empirical (deriving underlying dimensionsthrough scale development). Moreover, they include studies that investigatehow these dimensions (or attributes) affect customer perceptions, satisfac-tion, and willingness to transact online; particularly relevant are those fewthat have used conjoint analysis within this context.

Dimensional Frameworks

Service quality has often been the starting point for researchers attempting todefine e-quality. Many studies have included at least one of the five dimen-sions that comprise the well-known SERVQUAL scale developed by Berryand Parasuraman (1991). Van Iwaarden and colleagues (2003) found thatall five SERVQUAL dimensions (reliability, tangibles, empathy, responsive-ness, and assurance) can be applied to website design and website use. How-ever, most researchers discovered that not all determinants of e-quality fit intothe SERVQUAL framework as dimensions unique to the online environment,not previously defined, began to emerge. For example, Cox and Dale (2001)argued that the lack of human interaction during an online experience makessome service quality attributes irrelevant for virtual operations and based theirconceptual model on ease of use, customer confidence, on-line resources,and relationship services. Nonetheless, SERVQUAL dimensions remained per-tinent and appeared consistently in studies about e-commerce quality, inparticular the dimensions of reliability (e.g., Yang and Jun 2002; Long andMcMellon 2004; Lee and Lin 2005) and assurance referred to as ‘‘security’’(e.g., Wang, Tang, and Tang 2001; Janda, Trocchia, and Gwinner 2002; Jun,Yang, and Kim 2004). Alzola and Robaina (2005) provided an excellent dis-course on the applicability of SERVQUAL to e-commerce, exploiting synergiesbetween it and scales appearing in the e-commerce literature. In so doing,they defined the following five e-quality dimensions and related them tothe SERVQUAL scale: fulfilled promises (reliability), design (tangibles), perso-nalization (empathy), guarantee (responsiveness), and security (assurance).

Madu and Madu (2002) defined e-quality more broadly. In addition toborrowing dimensions from service quality, they borrowed dimensions fromproduct quality (Garvin 1984) and introduced dimensions that are unique tovirtual operations. Their framework includes the following 15 e-qualitydimensions: performance, features, structure, aesthetics, reliability, storagecapability, serviceability, security and system integrity, trust, responsiveness,product=service differentiation and customization, web store policies, repu-tation, assurance, and empathy. Product quality dimensions are redefinedfor e-commerce. For example, performance, a dimension of product quality,here refers to ‘‘ease of use and accuracy of content on the website.’’ Features,another dimension of product quality, refer to ‘‘extras such as links to othersites.’’ Dimensions unique to e-quality include trust, defined here as the will-ingness of users to disclose personal information over the Internet, and

270 R. Sebastianelli and N. Tamimi

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

storage, defined as capability and ease with which information can beretrieved when needed. While comprehensive, this framework is conceptualrather than empirically-based.

E-quality Scales

A number of researchers have taken an empirical rather than conceptualapproach to quality in e-commerce by developing scales to measure latent(underlying) e-quality constructs. SITEQUAL, a 9-item scale developed byYoo and Donthu (2001), identified four e-quality dimensions: (1) ease ofuse, (2) aesthetic design, (3) processing speed, and (4) security of personaland financial information. These four dimensions were based on data col-lected from convenience samples of students. Using information gatheredfrom website designers and consumers, as well as undergraduate students’ratings of e-commerce sites, Loiacono, Watson, and Goodhue (2002)developed WebQual. They identified the following 12 dimensions: informa-tional fit-to-task, interactivity, trust, response time, ease of understanding,intuitive operations, visual appeal, innovativeness, flow=emotional appeal,consistent image, online completeness, and better than alternative channels.Both SITEQUAL and WebQual focus exclusively on the website interface.

Given that a consumer’s experience with e-tailers goes beyond the web-site interface, others have developed scales that consider all aspects of pur-chasing via the Internet. For example, eTailQ (Wolfinbarger and Gilly 2003)consists of 40 items that were reduced to four underlying e-tailing dimensions:(1) fulfillment=reliability, (2) website design, (3) customer service, and (4)security=privacy. Long and McMellon (2004) developed a 53-item scale andbased on responses from a convenience sampling of students, derived sevendimensions. These seven dimensions include the five SERVQUAL dimensions(reliability, tangibles, empathy, responsiveness, and assurance) plus communi-cation (clarity=content=intent) and the purchase process (ordering=shipping=packaging). Also quite comprehensive are the E-S-QUAL and E-RecS-QUALscales developed by Parasuraman, Zeithaml, and Malhotra (2005). Based ondata collected from a random sample of Internet users, E-S-QUAL yieldedthe following four e-tailing quality dimensions: efficiency (ease of using thesite), fulfillment (extent to which the site’s promises are fulfilled), system avail-ability (correct technical functioning), and privacy (degree of protection).E-RecS-QUAL revealed three e-customer service dimensions: responsiveness(effective handling of problems), compensation (degree consumers arecompensated for problems), and contact (availability of assistance).

E-tailing Quality

Most studies that have examined e-tailing quality in particular have oftendone so within the context of understanding its impact on purchase intention

Consumers’ Perceptions of E-tailer Quality 271

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

and customer satisfaction. Trust, although conceptualized in different ways,has been studied extensively as an important attribute affecting a customer’swillingness to purchase online (e.g., Kimery and McCord 2002; Lee, Choi,and Lee 2004; Zhang 2005). Gefen, Karahanna, and Straub (2003) arguedpersuasively that online shopping involves not only interacting with ane-tailer’s website but with the e-tailer itself, consequently proposing a modelthat integrates TAM (Technology Acceptance Model) with trust. Among theantecedents to trust considered was familiarity with the e-vendor. Their find-ings suggest that online customers are influenced by both their trust in thee-vendor and technological aspects of the website interface, specificallyperceived usefulness and perceived ease of use. A number of other studiesabout online shopping attitudes and behavior also considered vendor char-acteristics such as reputation (see review by Li and Zhang 2002), and e-tailerreputation has been part of conceptual frameworks (e.g., Bramall, Schoefer,and McKechnie 2004) and empirical studies (e.g., McKnight, Choudhury, andKacmar 2004) on e-trust.

Several studies lend support to the notion that vendor characteristics, inaddition to website characteristics, can impact customer perceptions, satisfac-tion, and purchase intention. For instance, Shergill and Chen (2005) foundthat reliability in order fulfillment had a significant effect on online shoppers’perceptions; Lee and Lin (2005) found that overall service quality in onlinestores positively influences customer satisfaction. Other studies continuedto focus on website characteristics, with site interactivity and informationquantity and quality emerging as important determinants of customer satis-faction and online purchase behavior (Trocchia and Janda 2003; Park andKim 2006; Ballantine 2005). These themes persist in more recent work thatexamines the role of relationship quality (based on information, system,and service quality) on customer commitment and retention (Sun 2010).

Conjoint Studies

Few studies in e-commerce have used conjoint analysis, although it has beena popular marketing research tool for decades. Keen and colleagues (2004)used the approach to understand consumers’ purchase decisions when facedwith alternative retail formats, namely store, catalog, and the Internet. Theyfound that the two attributes most important in guiding the decision-makingprocess were retail format and product price.

A study that is more comparable to the one presented in this article is thatof Schaupp and Belanger (2005), who used conjoint methodology to evaluateonline consumer satisfaction. They considered attributes related to technology(security, usability and site design, privacy), shopping (convenience, trust-worthiness, delivery), and product (merchandising, product value, productcustomization). Each attribute was conceptualized at three levels; for example,delivery was defined as (1) provide a tracking number, (2) minimization of

272 R. Sebastianelli and N. Tamimi

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

delivery time, and (3) customer made aware of delays. They used the full pro-file approach with written paragraph descriptions, but they asked participantsto rate their level of satisfaction with each e-tailer resulting in metric (ratherthan ranked) responses. Consequently, they used ordinary least squares toestimate the conjoint models. Based on data collected from a sample of under-graduate students, they found that the most important attribute by far was priv-acy followed by merchandising.

Closely following the design of Schaupp and Belager, Chen and collea-gues (2010) used conjoint analysis to examine the impact of seven attributes(security, privacy, usability, convenience, trust, delivery, and product value)on online purchase intention. The data collection process involved twostages. In the first stage, participants rated the levels (features) for most ofthe attributes directly; the second stage involved evaluating four simulatedonline shopping websites that varied only two attributes: security and priv-acy. Based on responses from a sample of undergraduate students, conjointmodels were estimated and the resulting part-worths used to cluster indivi-duals into segments. This yielded three groups that were profiled in termsof their preference structures and labeled as usability=delivery-oriented,security=trust-oriented, and convenience=trust-oriented. In addition to seg-menting participants, the researchers concluded that online shopping web-sites should focus on usability, delivery, security, trust, and convenience inorder to increase online purchase intention.

RESEARCH METHOD

Design

In an effort to balance the need to represent salient e-tailing quality dimen-sions cited in the literature while keeping the number of overall judgmentsrequired by study participants in the conjoint task manageable, five attributeswere used to describe e-tailers. The first attribute is reputation of the retailer.Not only has this dimension been included in a number of studies related totrust, but it is also associated with SERVQUAL dimensions that apply toe-tailing quality (i.e., reliability, responsiveness). The second and thirdattributes used are site usability and security, both of which have beenrepeatedly included in prior studies and e-quality scales (i.e., SITEQUAL,WebQual, eTailQ, E-S-QUAL). In order to capture aspects of online shoppingthat go beyond the website interface, delivery (to represent the fulfillmentdimension in eTailQ and E-S-QUAL) and customer support (to represente-customer service dimensions from E-RecS-QUAL) were included as thefourth and fifth attributes. Each of these is conceptualized at two levels, asshown in table 1.

A complete factorial design of the five attributes results in a total of 32(2� 2� 2� 2� 2) different e-tailer descriptions. The full profile approach

Consumers’ Perceptions of E-tailer Quality 273

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

was used to collect the data. Because this method uses the complete set ofattributes, it yields 32 e-tailer profiles for participants to judge. In order toalleviate respondent fatigue during the conjoint task, a one-half fractionalfactorial design was used to implement the full profile approach. Thisreduced the number of evaluative judgments required by each participantfrom 32 to 16.

Data Collection

Research participants were recruited from a university setting to take part inan experimental study about ‘‘online shopping’’ during spring 2012. Flyerswere used to publicize the study and provide details regarding compensationand session times. The conjoint task was one of several tasks to be completedduring the session. The study was carried out over multiple sessions.

For the conjoint task, participants were presented with 16 multiple cuestimulus cards; each had a verbal description of an e-tailer in terms of the fiveattributes. All groups of cards were randomized prior to presentation. Parti-cipants were instructed as follows: You have 16 cards (designated A throughP) and each describes a particular online retailer. Arrange the cards from bestonline retailer to worst online retailer based on the description provided. Placethe rank (1¼ best to 16¼worst) on the line next to the retailer designation.Participants were asked to rank, rather than rate, the e-tailer descriptions forc-ing them to ‘‘trade off’’ among the attributes. A subset of participants com-pleted a second conjoint task. The set of cards used in the second conjointtask were created according to a different one-half fractional factorial designthan those used in the original conjoint task. These participants, therefore,evaluated two sets of e-tailers, each set being described by the same attributesand their levels, but neither set containing duplications of the other. This pro-cedure is referred to as the alternate forms method with spaced testing and isused to assess the reliability and predictive validity of the estimated conjointmodels. Participants also completed a brief survey that gathered data ondemographics and online browsing and shopping behaviors.

TABLE 1 Attributes and Levels for E-tailer Descriptions

Attribute Levels

Reputation 1. Lesser known small specialty retailer.2. Widely recognized, established retailer.

Site usability 1. Web interface is not user-friendly.2. User-friendly web interface.

Security 1. No visible assurance of secure transactions.2. Encryption technologies to assure secure transactions.

Delivery 1. Order tracking number provided.2. Excellent record for on-time delivery.

Customer support 1. Toll-free number is available for customer service.2. Live-person chat support available for customer service.

274 R. Sebastianelli and N. Tamimi

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

Data Analysis

Conjoint models are estimated at the individual level for all participants.The rankings of e-tailer descriptions are analyzed using monotone analysisof variance (MONANOVA), a method that is especially suited for factorialdesigns and appropriate for estimating part-worths for attribute levelsfrom ordinal judgments. MONANOVA selects the best monotone trans-formation of the data, over all possible monotone transformations, suchthat the greatest percentage of variance can be accounted for by the maineffects.

Reliability and predictive validity are assessed using the rankings for thealternate set of e-tailer profiles provided by a subset of the participants.These rankings are analyzed using MONANOVA to obtain a second set ofestimated individual conjoint models. Pearson correlations between thepart-worths of the first and second models are computed to measurereliability. To assess predictive validity, the first set of estimated individualconjoint models are used to predict each participant’s rankings for thealternate set of 16 e-tailer profiles. The predicted rankings for the secondset of e-tailer profiles are then correlated (by computing Spearman rankcorrelations) to the actual rankings.

Cluster analysis, based on the estimated part-worths of the individualconjoint models, is used to group participants into segments that are similarwith respect to perceptions of e-tailer attributes. The Ward’s minimum vari-ance algorithm, a hierarchical agglomerative procedure that minimizeswithin cluster sum of squares, is used to perform the analysis.

RESULTS

Respondent Profile

A total of 122 participants took part in the study; 4 participants did not rankorder the e-tailers correctly and therefore were eliminated from the analysis.Therefore, the sample size is n¼ 118. Participants include undergraduate stu-dents, graduate students, and professional staff. Respondents range in agefrom 19 to 60, with an average of 26.86 years. About 29% are older than22, suggesting that the majority consists of typical undergraduate students.They are predominately female (65.3%), and most are employed at leastpart-time (26% full-time and 37% part-time). Additional information regard-ing online shopping behaviors, such as browsing and purchasing frome-tailers, is provided in tables 2, 3, and 4. Over 62% have made between1 and 6 online purchases in the last 3 months; almost 80% make at least 6purchases from online retailers, on average, per year. The vast majority ofrespondents (almost 74%) report frequent browsing of e-tailers, on a dailyor weekly basis.

Consumers’ Perceptions of E-tailer Quality 275

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

Conjoint Results

Conjoint models were estimated at the individual level for all 118 parti-cipants. Summary results for the estimated part-worths of attribute levelsacross all participants are presented in table 5. The range in the estimatedpart-worth values between the levels of a given attribute indicates the attri-bute’s importance; the greater the range the more important the attribute’scontribution to the overall judgment of e-tailer quality. Table 6 presents asummary of the relative importance of the five attributes.

The relative importance is expressed as the mean percentage contri-bution (across all participants) to the overall judgment (the standard devi-ation in percentages indicates the level of consistency among participants).On average, the attribute most important to perceptions of e-tailer qualityis security, followed by site usability and reputation of retailer. Deliveryand customer support are found to be substantially less important.

As noted earlier, the alternate forms method with spaced testing wasused to assess the reliability and predictive validity of the estimated conjoint

TABLE 3 Online Purchasing Behaviors of Study Participants

Frequency of purchasingfrom online retailers (per year) Number Percent

0 to 5 25 21.2%6 to 12 35 29.7%13 to 20 33 28.0%21 to 36 16 13.6%36 or more 9 7.6%

TABLE 2 Online Purchasing Behaviors of Study Participants

Number of online purchasesmade in the last 3 months Number Percent

None 8 6.8%1 to 3 39 33.1%4 to 6 35 29.7%7 to 9 14 11.9%10 or more 22 18.6%

TABLE 4 E-tailer Browsing Behavior of Study Participants

Frequency of browsingonline retailers Number Percent

Daily 37 31.4%Weekly 50 42.4%Monthly 26 22.0%Only a few times per year 5 4.2%

276 R. Sebastianelli and N. Tamimi

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

models. A subset of 34 participants completed a second conjoint task, whichinvolved ranking an alternate set of 16 e-tailer profiles. Correlations betweenthe part-worths estimated for the first and second set of conjoint modelsrange from 0.871 to 1.000, with a mean of 0.972 and a median of 0.981, indi-cating a high level of reliability. To assess predictive validity, the first set ofestimated individual conjoint models was used to predict each participant’sranking for the alternate set of e-tailers. Spearman rank correlations betweenthe predicted and actual rankings range from 0.626 to 0.997, with a mean of0.875 and a median of 0.928. These results indicate a relatively high level ofpredictive validity as well.

Cluster Analysis Results

Agglomerative hierarchical clustering (AHC), based upon the relative impor-tance of attributes to individual participants, was used to form segments thatmaximize similarity within groups (and dissimilarity between groups). Afour-cluster solution is obtained, and each segment can be described basedon its perceptions of the relative influence each attribute has on e-tailer qual-ity (see figure 1).

Segment 1 is the largest with n1¼ 61. While this segment places moreimportance on security compared to the other four attributes, its relative

TABLE 5 Summary Results: Estimated Part-Worth Values for Attribute Levels

Model Minimum Maximum MeanStandarddeviation

Lesser known, small specialty retailer. �4.610 2.214 �1.403 1.385Widely recognized, established retailer. �2.214 4.610 1.403 1.385Web interface is not user-friendly. �4.610 0.443 �1.476 1.192User-friendly web interface. �0.443 4.610 1.476 1.192No visible assurance of secure transactions. �4.610 �0.010 �3.319 1.222Encryption technologies to assure securetransactions.

0.010 4.610 3.319 1.222

Order tracking number provided. �2.061 2.062 �0.011 0.628Excellent record for on-time delivery. �2.062 2.061 0.011 0.628Toll-free number is available for customer service. �1.464 2.045 �0.040 0.597Live-person chat support available for customerservice.

�2.045 1.464 0.040 0.597

TABLE 6 Average Relative Importance of Attributes

Attribute Relative importance (Mean %) Standard deviation

Security 50.195% 26.433%Site usability 20.014% 17.833%Reputation 19.998% 19.188%Delivery 5.162% 5.968%Customer support 4.630% 5.641%

Consumers’ Perceptions of E-tailer Quality 277

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

importance is not as great when viewed in comparison to the other groupprofiles. Moreover, this group places higher importance on e-vendor charac-teristics (delivery and customer support) relative to other groups. Conse-quently, even though the two most important attributes to Segment 1 arethe website characteristics of security and site usability, it can be dis-tinguished based upon the relatively higher importance it places on e-vendorcharacteristics (reputation of retailer, delivery, and customer support). Thissegment is more equally oriented between site and vendor attributes thanother segments. And as the largest segment, its profile with respect to demo-graphics and online shopping behaviors mimics that of the entire sample(n¼ 118).

Segment 2 consists of 25 individuals and compared to Segment 1, placeshigher relative importance on the two website attributes of security andusability. Consequently, this group can be labeled as security=site usabilityoriented. Relative to other segments, this group tends to be younger (averageage is 25), with a higher percentage browsing e-tailers frequently (daily orweekly) and making more online purchases per year (60% make at least 13).

Segment 3 is the smallest with n3¼ 13. This group clearly places thehighest relative importance on the reputation of the retailer, and to a lesserdegree on security. This reputation=security-oriented segment tends tobrowse e-tailers less frequently and make fewer online purchases per year,on average, compared to other groups. Moreover, a larger percentage of

FIGURE 1 Relative importance of attributes by segment (color figure available online).

278 R. Sebastianelli and N. Tamimi

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

individuals in this segment report being unemployed (perhaps the reason forless frequent online shopping).

Finally, Segment 4, comprised of 19 individuals, is security oriented tothe exclusion of all other attributes. The group tends to be older (averageage is 29.36) with a smaller percentage browsing e-tailers on a daily orweekly basis, although it resembles the sample at large with respect to onlinepurchasing. This group has a higher percentage of females (74%) than anyother segment.

DISCUSSION AND LIMITATIONS

The results from this conjoint study indicate that security is the most impor-tant attribute, on average, affecting perceptions of e-tailing quality, followedby site usability and reputation of the retailer. The major advantage of usingconjoint methodology, as already noted, is the ability to force individuals to‘‘trade-off’’ among attributes when judging e-tailer quality, thereby providinga more realistic way to determine the relative importance of various attri-butes. Furthermore, the individual level conjoint models estimated exhibithigh levels of reliability and predictive validity. It appears that conjointmethodology has potential for e-commerce research that focuses on under-standing how various attributes affect users’ preferences, opinions, andperceptions about e-quality.

Cluster analysis on the derived relative importance of attributes ident-ified four distinct segments that offer some insight into how differences indemographics and=or browsing and online shopping behaviors might influ-ence perceptions of e-tailer quality. For instance, the segment that valuedsecurity to the exclusion of all other attributes tends to be older and lesslikely to browse e-tailers on a frequent basis. The reputation=security-oriented segment not only tends to browse e-tailers less frequently, but alsomakes fewer online purchases. This segment has the highest percentage ofunemployed individuals compared to other segments. The other twosegments reveal that site usability and the vendor characteristics of deliveryand customer support gain relative importance for those who are younger,tend to browse online retailers on a more frequent basis, and make moreonline purchases.

As in most research, a number of decisions are required in a conjointstudy. These include selecting the type of model, data collection method,the attributes and levels, measurement scale for the response, and estimationmethod. This study differs from previous applications of conjoint analysis ine-commerce research (Schaupp and Belanger 2005; Chen et al. 2010) in sev-eral important ways. First, this study is the only study to use quality of e-taileras the response variable (rather than customer satisfaction or online purchaseintention). Second, participants ranked e-tailer profiles that were constructed

Consumers’ Perceptions of E-tailer Quality 279

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

using a fractional factorial design of all attributes being considered. This taskresults in a more complete ‘‘trade-off’’ among the attributes based on a true‘‘overall’’ evaluative judgment. Previous studies have used various combina-tions of rating or ranking attribute levels directly with evaluating paragraphdescriptions or judging websites created by manipulating only a subset ofthe attributes being considered. Third, this is the only study to assess thereliability and predictive validity of the estimated conjoint models. Finally,while the majority of participants in this study are typical undergraduates, theyare not exclusively undergraduate students. This offers some variation interms of demographics and online browsing and shopping behaviors notpossible in studies based solely on undergraduate students.

Even though this study does not exactly replicate previous applications ofconjoint analysis to e-commerce, it is still important to consider the findings inthe context of prior research. Parallels between the results of this study andthose obtained by Chen and colleagues (2010) are notable. First, Chen andcolleagues also found the attributes of security and usability to be of highrelative importance, in their case, to determining online purchase intention.Second, they identified three segments based on cluster analysis of estimatedconjoint part-worths that bear some similarity to those obtained in thisstudy. For example, their group labeled security=trust-oriented resembles thisstudy’s segment named reputation=security-oriented, especially since the linkbetween reputation of retailer and trust has been established in previousresearch. The segment they label usability=delivery-oriented can be con-sidered somewhat comparable to this study’s first and second segments inwhich site usability and vendor characteristics share high relative importancewith security. Fewer similarities are apparent between this study and thatcarried out by Schaupp and Belanger (2005), however. Interestingly, in theirstudy, security ranked last in terms of relative importance to customer satisfac-tion. Moreover, usability was deemed to be of only moderate importance, afinding they admit contradicts previous research. As noted above, a numberof decisions must be made in designing a conjoint study. Perhaps theseseemingly inconsistent findings point to the need for understanding how thesedecisions might affect conjoint results. For instance, consider the possibility ofan attribute’s relative importance being influenced by how one conceptualizesits levels. In this study, two levels that are quite extreme are used toconceptualize security: no visible assurance of secure transactions versusencryption technologies to assure secure transactions. Schaupp and Belanger(2005), on the other hand, used three levels, none of which indicate a com-plete lack of security (confirmation screen, encryption, and password=ID-protected accounts). Can this be a plausible explanation for why securityis the most important attribute in this study and the least important attributein theirs? It may be, although Chen and colleagues (2010) conceptualizedsecurity using four levels (information encryption, accounts with ID and pass-word, confirmation screen, and digital certificate) and still found it to be one of

280 R. Sebastianelli and N. Tamimi

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

the most important attributes in determining online purchase intention. None-theless, this raises an important consideration (and perhaps an interesting areaof future study) when applying conjoint analysis to e-commerce research.

This study is, of course, not without its limitations. First, attention isrestricted to only five attributes in order to limit the number of judgmentsrequired in the experimental task. Obviously, there are other attributes thatinfluence customer perceptions of e-tailer quality, as supported by previousresearch, which could have been included. Second, the findings depend, atleast in part, on the design of the conjoint study (i.e., how each attribute isconceptualized, how the dependent variable is measured). Third, the sam-ple was not randomly selected, but participants were recruited to participatein a study about ‘‘online shopping.’’ This may have biased the sampletoward individuals with more experience shopping online and perhaps amore favorable attitude toward shopping online, compared to the generalpopulation.

CONCLUDING REMARKS

As with any research methodology, conjoint analysis is not without its limita-tions. Nonetheless, its advantage of providing a more realistic task makes it apromising methodology for better understanding the attributes affecting con-sumers’ perceptions of online shopping and other e-commerce transactionsand online activities. The ability to derive the relative importance of variousattributes to customers’ perception of quality, satisfaction, and purchaseintention from overall judgments, and then using these derived values to seg-ment customers, is valuable from both a research and practical perspective.

REFERENCES

Alzola, L. M., and V. P. Robaina. 2005. SERVQUAL: Its applicability in electroniccommerce B2C. Quality Management Journal 12 (4): 46–57.

Ballantine, P. W. 2005. Effects of interactivity and product information on consumersatisfaction in an online retail setting. International Journal of Retail and Distri-bution Management 33 (6): 461–471.

Berry, L. L., and A. Parasuraman. 1991. Marketing services—Competing throughquality. New York: The Free Press.

Bramall, C., K. Schoefer, and S. McKechnie. 2004. The determinants and conse-quences of consumer trust in e-retailing. Irish Marketing Review 17 (1=2): 13–22.

Chen, Y., I. Hsu, and C. Lin. 2010. Website attributes that increase consumerpurchase intention: A conjoint analysis. Journal of Business Research 63(9=10): 1007–1014.

Cox, J., and B. G. Dale. 2001. Service quality and e-commerce: An exploratoryanalysis. Managing Service Quality 11 (2): 121–131.

Consumers’ Perceptions of E-tailer Quality 281

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

Garvin, D. A. 1984. What does product quality really mean? Sloan ManagementReview 26 (1): 35–43.

Gefen, D., E. Karahanna, and D. W. Straub. 2003. Trust and TAM in online shopping:An integrated model. MIS Quarterly 27 (1): 51–90.

Green, P., and V. Rao. 1971. Conjoint measurement for quantifying judgmental data.Journal of Marketing Research 8:355–363.

Janda, S., P. J. Trocchia, and K. P. Gwinner. 2002. Consumer perceptions of Internetretail service quality. International Journal of Service Industry Management 13(5): 412–431.

Jun, M., Z. Yang, and D. Kim. 2004. Customers’ perceptions of online retailingservice quality and their satisfaction. International Journal of Quality andReliability Management 21 (8): 817–840.

Keen, C., M. Wetzels, K. Ruyter, and R. Feinberg. 2004. E-tailers versus retailers:Which factors determine consumer preferences. Journal of Business Research57 (7): 685–695.

Kimery, K. M., and M. McCord. 2002. Third-party assurances: Mapping the road totrust in e-retailing. Journal of Information Technology Theory and Application4 (2): 63–82.

Lee, G. G., and H. F. Lin. 2005. Customer perceptions of e-service quality in onlineshopping. International Journal of Retail and Distribution Management 33(2=3): 161–176.

Lee, S. M., J. Choi, and S. Lee. 2004. The impact of a third-party assurance seal incustomer purchasing intention. Journal of Internet Commerce 3 (2): 33–51.

Li, N., and P. Zhang. 2002. Consumer online shopping attitudes and behavior: Anassessment of research. Proceedings of Eighth Americas Conference onInformation Systems, August 9–11, Dallas, TX, 508–517.

Loiacono, E., R. Watson, and D. Goodhue. 2002. WEBQUAL: A measure of web sitequality. Marketing Educators’ Conference: Marketing Theory and Applications13:432–437.

Long, M., and C. McMellon. 2004. Exploring the determinants of retail service qualityon the Internet. Journal of Services Marketing 18 (1): 78–90.

Madu, C. N., and A. A. Madu. 2002. Dimensions of e-quality. International Journal ofQuality and Reliability Management 19 (3): 246–258.

McKnight, D. H., V. Choudhury, and C. Kacmar. 2004. Shifting factors and theineffectiveness of third-party assurance seals: A two-stage model of initial trustin web business. Electronic Markets 14 (3): 252–266.

Parasuraman, A., V. A. Zeithaml, and A. Malhotra. 2005. E-S-QUAL: A multi-itemscale for assessing electronic service quality. Journal of Service Research 7 (3):213–233.

Park, C., and Y. Kim. 2006. The effect of information satisfaction and relationalbenefit on consumers’ online shopping site commitments. Journal of ElectronicCommerce in Organizations 4 (1): 70–90.

Schaupp, L., and F. Belanger. 2005. A conjoint analysis of online consumersatisfaction. Journal of Electronic Commerce Research 6 (2): 95–111.

Shergill, G. S., and Z. Chen. 2005. Web-based shopping: Consumers’ attitudestowards online shopping in New Zealand. Journal of Electronic CommerceResearch 6 (2): 79–94.

282 R. Sebastianelli and N. Tamimi

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014

Sun, H. 2010. Transferring attributes of e-commerce systems into business benefits: Arelationship quality perspective. Journal of Electronic Commerce Research 11(2): 92–190.

Trocchia, P. J., and S. Janda. 2003. How do consumers evaluate Internet retail servicequality? Journal of Services Marketing 17 (2=3): 243–253.

Van Iwaarden, J., T. van der Wiele, L. Ball, and R. Millen. 2003. Perceptions about thequality of websites: A survey amongst students at Northeastern University andErasmus University. Information and Management 41 (8): 947–959.

Wang, Y., T. Tang, and J. Tang. 2001. An instrument for measuring customer satis-faction toward web sites that market digital products and services. Journal ofElectronic Commerce Research 2 (3): 89–102.

Wolfinbarger, M., and M. C. Gilly. 2003. eTailQ: Dimensionalizing, measuring andpredicting e-tail quality. Journal of Retailing 79 (3): 183–98.

Yang, Z., and M. Jun. 2002. Consumer perception of e-service quality: From Internetpurchaser and non-purchaser perspectives. Journal of Business Strategies 19 (1):19–41.

Yoo, B., and N. Donthu. 2001. Developing a scale to measure the perceived qualityof an Internet shopping site (SITEQUAL). Quarterly Journal of ElectronicCommerce 2 (1): 31–46.

Zhang, H. 2005. Trust-promoting seals in electronic markets: Impact on onlineshopping decisions. Journal of Information Technology Theory and Application6 (4): 29–39.

Consumers’ Perceptions of E-tailer Quality 283

Dow

nloa

ded

by [

b-on

: Bib

liote

ca d

o co

nhec

imen

to o

nlin

e U

P] a

t 21:

34 2

2 A

pril

2014