E-Commerce Websites: a Qualitative Evaluation

Adriano Bessa Albuquerque Arnaldo Dias Belchior

Abstract E-commerce is considered an excellent alternative for companies to reach new
customers. However, many e-commerce websites have a short life because they dont meet the
minimal software quality requirements. To help the development of quality e-commerce websites,
this work identifies and ranks the main quality attributes to this application domain. The survey of
the attributes was based on the specialied literature and on the analysis of si!nificant websites.
Keywords" E-commerce, metrics, quality evaluation.
Todays world, becomin! more competitive every day, is demandin! from companies the
flexibility to ad#ust themselves to the permanent situations of market chan!e, readiness for
constant innovation and warranty of the quality of products and services.
$t the same time, %nternet, after bein! used initially as a !reat source of trade information,
!radually came to be used as an important environment for tradin! values&e-commerce'.$ccordin!
to (upiter )esearch, e-commerce transactions" business-to-business and business-to-consumer,
could exceed *+,- trillion per annum by .//0.However, the sites do not completely satisfy their
customers in several aspects and those deficiencies can eventually threaten the very existence of
many of those companies in the market.
This paper is intended to collaborate in the effort to produce e-commerce sites of quality,
throu!h the identification and rankin! of their main quality characteristics, as well as a survey of
the different developers and users points of view. To attain the desired quality of software
products, it is necessary to produce models that enable evaluation of those products quality.
$ccordin! to %+1, the main purpose of software quality evaluation is to supply referential
quantitative results to the software products that are reliable, understandable and acceptable to
anyone interested. *ser satisfaction and economic return are also important considerations.
The quality characteristics were or!anied usin! the 2uy 3odel to +oftware 4uality
Evaluation that was already used with satisfactory results in several application domains. That
model ranks a set of software quality attributes, or!aniin! them into three ob#ectives. Each
ob#ective is composed of quality factors. Those factors can be subdivided into sub-factors.
*sability is a quality ob#ective that refers to the characteristics that allow use of the e-
commerce site in the most diverse situations, not only durin! its development process, but also
durin! its operation and maintenance. 5onceptual )eliability concerns the e-commerce sites
capacity to implement, satisfactorily, what was specified and desi!ned. The )eliability of the
)epresentation refers to the e-commerce sites representation characteristics that affect its
understandin! and manipulation alon! its life cycle.
The fuy model sta!es used for the evaluation of e-commerce website quality are described
below
2irst +ta!e" establishment of the evaluation ob#ect and the set of items to be evaluated.
2or the three quality ob#ectives, hundred and sixteen sub-factors for e-commerce website
developers were found and appraised, contained in ei!hteen quality factors. 2or e-commerce
website users, seventy-ei!ht sub-factors were found, constitutin! a subset of the specialists6 sub-
factors, contained in nine quality factors.
The evaluators !ave !rades from / to 7 to each sub-factor, which underwent fuification,
that is, the transformation of those !rades into normal trian!ular fuy numbers accordin! to Table
8.%ts pertinence functions !raph is presented in 2i!ure 8 below.
$ normal trian!ular fuy number can be represented by 9 &a, m, b', where the values :a;
and :b; identify, respectively, the inferior and superior limits of the trian!le base. The value of
:m; corresponds to the trian!le hei!ht.
+econd +ta!e" obtainin! specialists and users profiles.
The specialists6 and users 6profiles were obtained throu!h the completion of the +pecialist
%dentification 4uestionnaire and the *ser %dentification 4uestionnaire respectively, !eneratin! a
wei!htin! measured by each evaluator, which will influence in the final result.
Third +ta!e" determinin! items identified in the first sta!e&de!ree of importance'.
2ield research provided opinions from thirty specialists and thirty users in four important
<railian +tates, accordin! to the questionnaire defined in the first sta!e. The evaluators were
!uided in the sense that they had to evaluate each of the quality sub-factors surveyed, accordin! to
its importance for the e-commerce website application domain.
2ourth +ta!e" fuy treatment of specialists6 collected data, in the evaluation of each collected
item.
%n that sta!e, the fuy treatment of data collected was performed in the evaluation of the
developers, users, and consolidated&developers plus users',usin! similar matrix, considerin! each
specialist6s and each users wei!htin!, obtained in the second +ta!e, as determined by the model.
2ifth +ta!e" fuy a!!re!ation of software quality attributes, in each quality model hierarchical
level.
%n this sta!e, the results obtained by the sub-factors were put to!ether to calculate the factor
results, accordin! to the model. The results obtained by the factors were #oined to calculate the
ob#ective results. The results obtained reveal the quality standard for e-commerce websites.
Table . displays the complete set of those quality attributes. The sub-factors are or!anied,
inside of the factors, in decreasin! order of importance, accordin! to the consolidated evaluation
&developers and users',that is, the #oint result of the specialists6 evaluation and e-commerce site
users.
The +ecurity factor was considered the most important. The result obtained indicates that, in
e-commerce websites, security is fundamental, especially when it comes to electronic payments,
which cannot be vulnerable to any kind of attack, and when it comes to the sub#ect of site
authentication itself. This factor obtained the defuification value of =.->,that is ..?very
important and ->? indispensable.
The %nte!rity factor was rated second in consolidated evaluation, reinforcin! that an e-
commerce website has to mana!e and control its stored data correctly and appropriately. The
consolidated value obtained was =.=@,indicatin! @7?very important and =@?indispensable.The
result shows that of the twenty consolidated quality sub-factors with the best rankin!, seven are
sub-factors related to +ecurity, thus showin! the importance of aspects such as authentication,
access control, privacy, vulnerability, electronic payment security, and others.
$ll factors obtained a !ood final evaluation, however, the factors +ecurity and %nte!rity
obtained the best score of all in the consolidated evaluation. This result portrays the Aeb context,
wherein electronic 5ommerce is inserted. <ecause these applications are public, accessed by a
vast population of users and it is not hard to find cases of security systems defrauded by hackers
who be!in to !ain access to unauthoried data.
9owadays, there is a preference for e-commerce websites that have resources a!ainst
unauthoried access and a!ainst interruption of processin!, even in abnormal situations, and sites
that !uarantee data preservation, that is, where modifications may only be made by appropriate
parties.