automatic website evaluations: the case of hotels in hong kong

Information Technology amp Tourism Vol 8 pp 255ndash269 1098-305806 $6000 + 00Printed in the USA All rights reserved Copyright 2006 Cognizant Comm Corp

wwwcognizantcommunicationcom

AUTOMATIC WEBSITE EVALUATIONS

THE CASE OF HOTELS IN HONG KONG

STEVEN CHAN and ROB LAWdagger

Department of Computing and daggerSchool of Hotel amp Tourism ManagementThe Hong Kong Polytechnic University Hung Hom Kowloon Hong Kong

Websites are now the basic requirement to an increasing number of communication and businessstrategies The usability of a website effectiveness of its interface as well as its amount of infor-mation ease of navigation and user friendliness of its functions are central to the success of thesestrategies Yet these important design qualities remain to date undefined and difficult to evaluatein a convenient and objective way In this article we describe the development of the AutomaticWebsite Evaluation System (AWES) an automated website evaluation tool AWES is applied to61 Hong Kong hotel websites HTML tags representing categories of the information pertinent toeffective website design are automatically collected and classified The results can provide objec-tive and quantitative guidance to website design

Key words Web page design Hong Kong Language interpreter Evaluations

Introduction align text into tables or place images (separatecomputer files) into specific locations in order topresent the layout of a website Often this creativeWith the emergence of low-cost broadband

transmission and improvements in personal com- process is in the hands of web page designerswhose background and training may be more tech-puter systems online shopping is increasingly be-

coming just another way for consumers to shop nical than design oriented The result hence maybe a website design that is technology heavy butand for businesses to market and sell their prod-

ucts and services As websites function in many that fails to satisfy the needs of customers andconsequently businesses An area where this isaspects in the business environment their appear-

ance organization and contents can have a con- clearly the case is in the tourism industry In HongKong tourism is a major source of income Insiderable influence on the behavior of online con-

sumers Seeing this businesses are now giving 2004 Hong Kong received 218 million visitors(Tourism Commission 2005) and it was commonmore attention to website design Usually web

pages are created by using a programming lan- for visitors and potential visitors prior to their vis-its to use the Internet to make their travel plansguage that employs tags to indicate boldface font

Address correspondence to Steven Chan Department of Computing The Hong Kong Polytechnic University Hung Hom KowloonHong Kong Tel (852) 2766 7254 Fax (852) 2774 0852 E-mail csstchancomppolyueduhk

255

256 CHAN AND LAW

Importantly then travel websites are often the Usabilityprospective visitorsrsquo first contact with Hong Kong

Users generally determine the usefulness of aTo date however there has been little attentionwebsite based on its usability or ease of use Thegiven to the important topic of the design and theusability of web pages has been measured in a va-impact of the design of tourism-related websitesriety of ways Ivory (2000) presented a web pageon usersmatrix composed of 42 features of a web pageThe Internet is transforming transaction pro-which can affect usability These 42 features werecesses in the tourism and travel industries makingthen grouped under three headings of page compo-them more competitive Buhalis (2002) indicatessition page formatting and overall page Most ofthat e-business can cope with the transformationthese features under page composition and format-from the traditional business operation Ultimatelyting (eg the number of words and font style) arethe efficiency and development of this medium arerepresented quantitatively Some of the more ab-directly related to the technologies and the inter-stract features clustered under an overall pageface through which the websites are developed Ithowever cannot be measured quantitatively some-is therefore important to establish and if possi-what compromising the objectivity of the metricble codify what would constitute good web page

design practice in the area of e-commerce in travelWebsite Structureand tourism It would certainly be useful if web

designers could refer to some commonly agreed Ivory and Hearst (2002) suggested measuresupon standards that allow the convenient objec- for assessing design quality and usability of web-tive and quantitative evaluation of important website structure that represent website design ele-site features (Law Ho amp Cheung 2004) Exam- ments within a triangular schema At the base ofples of these features are the appearance of text the triangle are text links and graphics Thesefonts color contrast and multimedia These stan- provide the foundation for elements at the upperdards can then be applied automatically levels but are most obviously operational in the

This article describes a study that makes an ini- interface Website elements that are assigned totial attempt to establish a standard for evaluating upper levels of the triangle schema such as for-the design of websites and an automatic process matting serve to assist users in forming a repre-for evaluating the design of web pages in the con- sentation of how to go about accessing a sitersquostext of hotels Although some prior studies such contentas the work by Ivory (2000) have evaluated webpages in the context of travel and tourism these Color Contraststudies were primarily related to or based on for-matting Modifying the approach that was initiated Color contrast has an effect on the readability

of a website and the combination of text color andby Ivory and Hearst (2002) a page format struc-ture that classifies the formatting of website into background color can persuade viewers to use a

website Color contrast can also drive users awayfour main areasmdashcontrast font multimedia andothersmdashis proposed in this research Figure 1 from using a website Lindsay (1999) suggested

that the colors used in computer graphics arepresents the proposed web page format structurebased on psychological andor biological consider-

Literature Review ations and that colors can attract users to browsethe website and to stay with it Researchers haveTo date work on website evaluation has mainly

focused on the areas of usability and of website proposed various guidelines for choosing the rightcolors For example Schraff and Ahumada (2002)structure More narrowly it has taken account of

basic design elements such as the use of fonts and recommended the use of plain backgrounds withmoderate-to-high-contrast text on patterned back-colors This section briefly reviews current web-

site evaluation metrics and related references in grounds Similarly Ivory (2002) discussed the im-portance of using a small number of browser-safethese areas

AUTOMATIC WEBSITE EVALUATIONS 257

Figure 1 Web page format structure

colors default link colors and color combinations and easy access to other web pages (HaffnerRoth Engel amp Meinel 2000) Effective hyper-with adequate contrastlinks are central to the overall web browsing expe-rienceUse of Text

In Ivoryrsquos (2000) proposed web page matrixTables

there are several criteria related to text font Likecontrast the style and size of font can affect the There are many ways to group and present text

and graphics but the most popular are tables andreadability of a web page Similarly Schraff andAhumada (2002) noted that the careful selection frames Wan (2000) noted that the use of frames

not only slows load times it makes it more diffi-of a font style and size is important in websitedesign and that different styles of applications cult for users to make a page becomes a ldquofavor-

iterdquo This may account for the greater popularityneed different styles of font The font size alsoaffects the text on a page Specifically Ivory of tables as a formatting tool Tables also allow

the addition to web pages of borders and graphics(2002) suggested that the average font size shouldbe between 8 and 12 when the text is describing and allow convenient spacing and text alignment

Kim and Albers (2003) suggested that using tablesthe contenton a web page can increase the readability of thepageMultimedia

Broadly speaking multimedia includes any fa-Page Length and Layout

cility that can display text graphics images andsound Multimedia has been found to have the The basic web page elements must be laid out

on a page of a certain length and for the optimalability to motivate customers to use a website (Ni-col Gutfreund Paschetto ampB Rush 1999) Be- effect Marcus (1997) stated that the optimally or-

ganized screen layout requires the manipulation ofcause multimedia presentations can increase cus-tomersrsquo satisfaction most users prefer websites three primary resources These include i) use an

underlying grid layout ii) standardizing the screenwith multimedia and the technology is widely usedin business websites including large-scale in- layout and iii) grouping related items Page for-

mat has been posited as a major issue in websitehouse and data-heavy business applications (Barryamp Lang 2001) Hyperlinks are a type of multime- structure (Ivory amp Hearst 2002) having a poten-

tially major impact on user satisfaction Bhattidia and have the crucial ability to allow users fast

258 CHAN AND LAW

Bouch and Kuchinsky (2000) recommended that Text in black on a white background is certainlymore visible and easy to a human eye than blackwebsite pages should avoid the use of scroll bars

as visitors do not like to spend time on scrolling text on a blue background The need for contrastwould be evident in an example of white text onpagesa gray background where the white text couldonly be read with difficultyMethodology and Findings

The Automatic Website Evaluation SystemBackground (AWES) uses two formulae as proposed in the

W3C Working Draft (2000) to evaluate contrastTo access a website an Internet user will browseOne formula is for calculating color differenceswith software such as Microsoft Explorer or Net-and the other one is for calculating differences inscape Navigator These browsers read speciallycolor brightnesscoded web pages and process them so that they

The color difference formula isare viewable on a computer screen To receive thisinformation the userrsquos computer makes a request

value = 0ndash765 = [max(Red value 1 Red value 2)over the Internet to an agent computer called aminus min(Red value 1 Red value 2)] + [max(Greenweb server The web server then responds to the

request by sending the encoded web page back to value 1 Green value 2) minus min(Green value 1the user and the userrsquos browser makes the web Green value 2)] + [max(Blue value 1 Blue valuepage viewable on a computer monitor or other de-

2) minus min(Blue value 1 Blue value 2)] (1)viceWhen a web page is accessed the browser soft-

The color brightness formula isware interprets specially coded web pages and dis-plays the contents on the computer screen The

value = 0ndash255 = ((Red value times 299) + (Greenpage itself is generally coded in html (Hyper Textvalue times 587) + (Blue value times 114))1000 (2)Markup Language) which is simply a standard

way of representing the format and contents of allDetail of Textinformation on a web page The html tag contents

are classified into different property groups The Even in a multimedia environment text re-grouping criteria are the following features mains a central way for web pages to deliver im-

portant information about products or servicesUse of Background Color Decisions about design features such as combina-

tions of font color font size and font family canGood web page design often makes good usehave a significant impact on how effectively textof color but it is important that the color contrastcan contribute to the clarity and immediacy of therelation between the informative text and the back-web page message The issue of variety is perti-ground be fully understandable At each extremenent here On the one hand many web pages willa background may be specified as either a purebenefit from a style that applies a limited selectioncolor or an image (picture) If a background colorof text design features On the other hand carefulor image is specified the browser begins by paint-and informed design may also make use of a moreing the specific portion or the entire page withvaried selection of features Certainly font proper-color or image and then proceeds to add the con-ties should be taken into account in any web pagetent Values of 500 and 125 for color differencedesign evaluation system In our system if a de-and brightness are assigned to the web pages thatfault setting is used the specific attribute is as-contain only pictures as background and that carrysigned a value of zerono informative text

A background image has the potential to reduceMultimedia

the readability of a display so it is important notto use images with dense detail and vivid colors Non-text elements such as images multimedia

objects logos photographs artwork Java appletsSimilarly low-contrast pages are difficult to read


and other types of web page contents also contrib- interpreter in that it targets the web page designlanguage and breaks the analytical task into aute to making an attractive web page In addition

to images and text elements web developers have number of different partsAWES utilizes commonly available browsersembraced multimedia technologies that allow web

pages to include presentations short movies or as the graphical user interface to allow users toinput a specific web page address (URL) to beother content with synchronized audio and video

tracks Some developers have also started using evaluated and provides several options for usersto choose from based on the expected output dis-such technologies as Real Video a form of

ldquostreamingrdquo multimedia that allows users to watch play Figure 3 shows the input interface of AWESIn this research we analyzed the home pagesonline videos with relatively little delay Multime-

dia may make a stronger and more immediate im- of the websites of 66 Hong Kong hotels (Table 1)These hotels comprised all the members of thepression on visitors than text and is also taken

into account in the evaluation system Hong Kong Hotels Association that have theirwebsites listed on Expediacom Travelocitycomand Hotelscom which represent the worldrsquos lead-Number of Tables and Length of Front Pageing intermediary websites for making hotel room

The number of tables on a web page and thereservations The hotel websites were divided into

length of a websitersquos front page are also includedthree categories of luxury midpriced and econ-

as parameters in our evaluation system Tables areomy in line with their star ratings Generally

used to present informationmdashwhether that infor-speaking fuve-star hotels fall into the luxury cate-

mation be textual or graphicalmdashin rows columnsgory four-star hotels are in the midpriced cate-

and cells This allows flexible page layout designgory and two- and three-star hotels are in the

as web developers can freely insert space and ex-economy category As Hong Kong does not have

tra lines between words and tags without disturb-an official star rating system our samples are cate-

ing the ultimate appearance of the page Nowa-gorized with reference to the three most prominent

days however web developers want greaterwebsites (Expediacom Travelocitycom and Hotel

control over spacing in a document and they maycom) The evaluation system is set to the most

create tables by simply adding extra spaces be-commonly used screen resolution 1024 times 768 pix-

tween column entries This process can be facili-els In this research the study evaluates the front

tated by table tags that format everything betweenpage the page that users most often first retrieve

these tags as preformatted textPage length has been cited as an important pa-

Procedurerameter of readability Black (1997) suggested thata well-designed web page should not require any

AWES consists of three modules a scanningscrolling The automatic evaluation software uses

module a parsing module and an evaluation mod-a pixel count of the computer display to calculate

ule The scanning module retrieves web pagethe length of the front page of each website

source file as input to the browser It filters thecontents of the page and breaks the file into mean-

Implementationingful groups of characters called tokens Thisphase also separates the contents of a web pageAWES has been developed to evaluate the

readability of websites of Hong Kong-based ho- into two groups (group 1 and group 2) Group 1is retained as meaningful content that serves astels The use of this system is intended to produce

evaluations that are not just automatic but also input into the evaluation module It will be furtheranalyzed later Group 2 is discarded as not beingconsistent transparent and objective Figure 2

shows the logic flow for AWES The system be- relevant to page layout Figure 4 shows the htmlsource file of a sample web page used by AWESgins by downloading a specific web page URL It

then parses the source file and reports characteris- Figure 4 contains four major classes of instruc-tions html tags style sheet JavaScript and com-tics of the corresponding web page via a browser

The system is similar to a programming language mentcontent texts Each of these four classes of

260 CHAN AND LAW

Figure 2 Overall framework of AWES

instructions is represented differently html tag is ldquohrefrdquo is the property name The property nameis followed by a path Here the path is ldquotopin normal font style sheets are in bold JavaScript

is underlined and context texts are italicized A htmlrdquo Words and values are separated withspecial characters such as spaces indicated by ho-tag starts with a special character the triangular

opening bracket lt which is followed by a reserve rizontal parallels ldquo=rdquo and double code Some tagsalso require an ending tag to end the function pro-word and ends with a triangular closing bracket

gt Some tags will have properties and these prop- vided by a tag Take line 22 as an example failureto include the tag ltcentergt in the sourceerties can be specified inside the tag An example

of such a specification can be seen in line 25 of would cause all hyperlinks in the defined table andits contents to be centered All reserved words ofFigure 4 In the phrase lta href=ldquotophtmlrdquogt

ldquoardquo is the reserved word for hyperlink and layouts that are related to html tag are prestored

Figure 3 Interface of AWES


Table 1

Selected Hotels for Analysis

Hotel Website

Economy (2ndash3 stars n = 35)B P International wwwbpihcomhkCentral Park Hotel wwwcentralparkhotelcomhkintrohtmCharterhouse wwwcharterhousecomCity Garden Hotel wwwcitygardencomhkDorsett Olympic Hotel wwwdorsettolympiccomhkEmperor Happy Valley Hotel wwwemperorhotelcomhkEmpire Hotel Hong Kong wwwasiastandardcomhotelindexphpHarbour Plaza North Point wwwharbour-plazacomhpnpindexhtmlHarbour Plaza Resort City HK wwwharbour-plazacomHPRCSourceoutindayhtmHarbour View International House wwwharbourymcaorghkHoliday Inn Golden Mile wwwgoldenmile-hkholiday-inncomHotel Concourse wwwhotelconcoursecomhkImperial Hotel wwwimperialhotelcomhkIsland Pacific Hotel wwwislandpacifichotelcomhkKimberley Hotel wwwkimberleycomhkLuk Kwok Hotel wwwlukkwokhotelcomMajestic Hotel wwwmajestichotelcomhkMarco Polo Gateway wwwmarcopolohotelscomgatewayMetropark Hotel wwwmetroparkhotelcomThe Metropole Hotel wwwmetropolecomhkthomehtmNew World Renaissance Hotel KLN wwwrenaissancehotelscomhkgnwNewton Hotel Hong Kong wwwnewtonhkcomNewton Inn North Point wwwnewtoninncomPanda Hotel wwwpandahotelcomhkPark Hotel wwwparkhotelcomhkRamada Kowloon wwwramadahongkongcomRegal Riverside Hotel wwwregalriversidecommmmregalriversideRosedale on the Park wwwrosedalecomhkRoyal Park Hotel wwwroyalparkcomhkRoyal Plaza Hotel wwwroyalplazacomhkSalisbury-YMCA of Hong Kong wwwymcahkorghksaleshtmlhst001ehtmSouth Pacific Hotel wwwsouthpacifichotelcomhkStanford Hillview Hotel wwwstanfordhillviewcomenindexaspStanford Hotel wwwstanfordhongkongcomenindexaspYMCA International House wwwymcaintlhousehkorg

Midpriced (4-stars n = 20)Eaton Hotel Hong Kong wwwlanghamhotelscomeatonhongkongCosmopolitan Hotel Hong Kong wwwcosmopolitanhotelcomhkenframehtmlExcelsior Hong Kong wwwmandarinorientalcomhotel538000001aspHarbour Plaza Hong Kong wwwharbour-plazacomhphkindexshtmHarbour Plaza Metropolis wwwharbour-plazacomhpmeHotel Miramar wwwmiramarhkcomHotel Nikko Hong Kong wwwhotelnikkocomhkHyatt Regency Hong Kong hongkongregencyhyattcomKowloon Hotel wwwthekowloonhotelcomGreat Eagle Hotel Hong Kong wwwlanghamhotelscomlanghamhongkongLe Meridien Cyberport wwwlemeridien-cyberportcomMarco Polo Hongkong Hotel wwwmarcopolohotelscomthehongkonghotelindexhtmlThe Park Lanemdasha Summit Hotel wwwsummithotelscomhotel_conciergecfmid=80394ampRegal Airport Hotel wwwregalairportcommmmregalairporten_US_mainhtmlRegal Hong Kong Hotel wwwregalhongkongcommmmregalhken_US_mainhtmlRegal Kowloon Hotel wwwregalkowlooncommmmregalklen_US_mainhtmlRegal Oriental Hotel-Hong Kong wwwregalorientalcommmmregalkaitaken_US_mainhtmlRoyal Garden Hotel wwwrghkcomhkThe Royal Pacific Hotel amp Towers wwwroyalpacificcomhkSilvermine Beach Hotel wwwresortcomhk

Luxury (5 stars n = 11)Grand Hyatt Hong Kong wwwhongkonghyattcomInterContinental Grand Stanford HK wwwhongkonginterconticomIsland Shangri La Hong Kong wwwshangri-lacomenLangham Place Hotel Mongkok HK wwwlanghamhotelscomlanghamplacehongkongMandarin Oriental Hong Kong wwwmandarinorientalcomMarriott JW Hong Kong wwwmarriottcompropertypropertypagemimarshaCode=HKGDTParkview Suites wwwparkviewlifestylecomPeninsula Hong Kong wwwpeninsulacomindexhtmlRenaissance Harbour View Hotel wwwmarriottcompropertypropertypagehkghvRitz-Carlton Hong Kong wwwritzcarltoncomhotelshong_kongSheraton Hong Kong Hotel Tower wwwstarwoodhotelscomsheratonsearchhotel_detailhtml

262 CHAN AND LAW

Figure 4 HTML source code

in a table inside the scanning module After the also be represented with style sheets These arealso grouped into group 1 Style sheets can be rep-scanning processing the tags that match the re-

served words in the list will be collected as one resented in three different ways i) they can bestated inside a tag as at line 18 of the example ingroup (group 1) to be parsed later and the others

will be classified into another group (group 2) Figure 4 ii) specified inside the source file (line4ndashline 7) or iii) stored in an external file TheFigures 5 and 6 show the group 1 and group 2

codes for our example rules for writing style sheets are slightly differentfrom those for tags In line 5 of Figure 4 bodyThe html tag properties and their values can


Figure 5 Source that will be used for evaluation (group 1)

background-colorFFFFAA the characters in- tains reserved words a property name a propertyvalue and a delimiter The system will then storeside the curly brackets are the properties In

front of the brackets we find the tag namenames the property name and corresponding value Forexample ltspan style=ldquofont-size30rdquogt the re-referencing this property Most of the style sheets

will affect the web page layout and therefore all served word ldquofont-sizerdquo is the Fonts group and ithas a value of 30 Other characters in the tag willstyle sheets will also be grouped into group 1

JavaScript is a programming language that is be omitted After parsing all the tagsstyle sheetsthe data will be passed to the evaluation modulecommonly used on web pages It provides interac-

tive functions and enhances usability but its use is Finally the page length is evaluated by a spe-cific function using JavaScript The function codenot within the scope of this study All JavaScript

codes are thus passed over and put into group 2 is added at the end of the html code to calculatethe actual size The result is calculated and passedFinally as content text will not affect the layout

design of a web page it is also classified into into the system for later use The computed webpage size is next compared with a manually mea-group 2 After the data are scanned and classified

group 2 data are discarded Group 1 data are sured size The comparison result shows that theaccuracy of the calculated size is more than 95stored and then passed into the analysis (parsing)

module and the remaining 5 is the error in loading anexternal style sheet file that leads to an incorrectInside the parsing module AWES uses a table

with presaved reserve words Tags in group 1 html layoutThe evaluation module is for characteristicsfrom the scanning module will be classified into

corresponding property groups by searching re- analysis and printing output Results of all tags ingroup 1 from the parsing module are summed upserved words that are matched in the table After

a property group is identified its tag will be bro- in this module The evaluation module presentsthe output results either in a table or in Excel for-ken down into different parts A common tag con-

264 CHAN AND LAW

Figure 6 Source that will not be used for evaluation (group 2)

mat and the results are displayed on a computer ever is quite diverse It is very difficult to detectthe contrast of the background color in a specificscreen Tables 2 3 and 4 show the results of the

analysis of the luxury midpriced and economy section that has superimposed text However thevalues obtained are still above the range as pre-hotelssented by the W3C Working Draft (2000) Inter-estingly although it has been noted that the use ofAnalysis and Discussiona ldquowelcomerdquo page may try the patience of some

Table 5 shows the comparison results of theweb surfers (Burns 2002) as Table 2 shows over

three hotel categories50 of the luxury hotels do still have a welcomepage

Color

Although the range of values for color is quite Use of Textwide there is no significant difference among thethree categories This result is likely due to the Like color the use of text varies in a wide

range of values The difference among the threefact that most homepages use a picture as back-ground The color distribution of the picture how- categories is however not insignificant In gen-


Table 2

Result of Luxury Hotels

Group 1

Group 2

Group 3 Group 4Contrast

Font Details

Multimedia OthersNo of No of No of

Color Font Font Font No of No of No of No of No of PageHotel Name Difference Brightness Color Sizes Family Graphs Flashes Music Links Tables Size

Grand Hyatt Hong Kong 451 160 0 1 1 7 0 0 5 1 092InterContinental Grand Staford 765 255 4 3 1 19 0 0 23 5 202Island Shangri La Hong Kong 765 125 1 2 1 114 0 0 41 28 115Langham Place Hotel Mongkok 500 125 3 4 2 43 0 0 49 27 116Mandarin Oriental Hong Kong 765 255 1 1 1 11 0 0 70 29 116Marriott JW Hong Kong 500 125 1 2 1 19 1 0 61 1 276Parkview Suits 581 198 1 1 1 N1 0 0 2 2 096Peninsula Hong Kong 500 125 1 1 1 61 0 0 48 17 228Renaissance Harbour View Hotel 500 125 1 2 1 21 0 0 58 1 347Riz-Carlton Hong Kong 500 125 1 1 1 141 0 0 40 15 110Sheraton Hong Kong Hotel Tower 765 255 1 1 1 59 0 0 47 23 292

Table 3

Result of Midpriced Hotels

Group 1

Group 2


Font Details



Eaton Hotel Hong Kong 306 102 5 2 3 6 2 0 2 7 113Cosmopolitan Hotel Hong Kong 425 140 1 1 1 37 1 1 21 6 094Excelsior Hong Kong 765 255 1 0 0 29 0 0 40 19 125Harbour Plaza Hong Kong 713 245 2 3 1 35 0 0 13 10 134Harbour Plaza Metropolis 645 220 2 3 1 50 0 0 15 10 134Hotel Miramar 765 255 4 2 2 9 5 0 16 12 098Hotel Nikko Hong Kong 765 245 1 1 1 23 0 0 16 6 124Hyatt Regency Kong Kong 483 175 5 2 1 15 0 0 23 5 225Kowloon Hotel 677 227 2 3 1 50 0 - 14 10 134Langham Hotel Hong Kong 500 125 3 4 2 43 0 0 42 27 100Le Meridien Cyberport 500 125 1 1 1 27 0 0 33 5 202Marco Polo Hongkong Hotel 512 186 1 2 1 41 0 0 11 2 086The Park Langmdasha Summit Hotel 765 255 8 4 2 92 0 0 37 14 244Regal Airport Hotel 663 244 5 6 3 68 0 0 43 21 206Regal Hong Kong Hotel 753 250 5 6 2 69 0 0 75 21 163Regal Kowloon Hotel 703 242 6 5 2 69 0 0 79 21 207Regal Oriental Hotel-Hong Kong 623 223 3 6 4 70 0 0 76 21 178Royal Garden Hotel 258 88 3 2 2 11 0 0 2 2 087The Royal Pacific Hotel amp Towers 765 255 1 1 1 25 1 0 10 1 081Silvermine Beach Hotel 754 251 1 1 1 9 1 0 24 10 076

266 CHAN AND LAW

Table 4

Result of Economy Hotels

Group 1

Group 2


Font Details



B P International 144 57 2 2 1 5 1 0 4 4 132Central Park Hotel 201 59 1 1 1 0 1 1 2 1 102Charterhouse 500 125 0 0 0 15 0 0 5 5 096City Garden Hotel 765 255 0 0 0 4 1 0 0 1 100Dorsett Olympic Hotel 731 248 1 2 2 7 0 1 14 4 115Emperor Happy Valley Hotel 500 125 0 0 0 30 0 0 3 1 088Empire Hotel Hong Kong 500 125 0 0 0 36 1 0 9 2 098Harbour Plaza North Point 500 125 2 3 1 35 0 0 13 10 136Harbour Plaza Resort City HK 765 255 2 3 1 35 0 0 15 10 135Harbour View International House 687 231 2 1 1 21 2 0 10 11 137Holiday Inn Golden Mile 765 255 4 4 1 20 0 0 20 6 205Hotel Concourse 679 233 0 0 0 4 1 0 1 2 096Imperial Hotel 561 189 1 5 1 14 4 0 6 3 204Island Pacific Hotel 500 125 0 0 0 4 1 0 0 1 081Kimberley Hotel 765 255 6 5 3 27 0 0 15 7 109Luk Kwok Hotel 500 125 1 1 1 15 1 0 5 5 096Majestic Hotel 765 255 1 1 1 26 0 0 14 0 078Marco Polo Gateway 512 186 1 2 1 41 0 0 11 4 075Metropark Hotel 765 255 1 2 2 19 2 0 22 3 107The Metropole Hotel 765 255 1 4 1 14 0 0 24 4 181New World Renaissance KLN 500 125 1 2 1 20 0 0 56 1 209Newton Hotel Hong Kong 500 125 1 1 1 10 0 0 11 2 181Newton Inn North Point 500 125 1 1 1 8 0 0 9 2 158Panda Hotel 680 220 3 5 1 21 0 0 14 2 155Park Hotel 598 203 4 4 3 4 0 0 14 3 134Ramada Kowloon 500 125 0 0 0 1 1 0 0 0 075Regal Riverside Hotel 765 255 4 5 2 67 0 0 74 20 201Rosedale on the Park 525 171 3 2 2 9 0 0 12 3 126Royal Park Hotel 441 147 6 3 3 55 1 0 38 9 108Royal Plaza Hotel 441 147 8 4 3 55 1 0 21 9 098SalisburymdashYMCA of Hong Kong 765 255 2 1 1 16 0 0 14 1 100South Pacific Hotel 500 125 1 1 1 1 1 0 1 2 100Stanford Hillview Hotel 351 125 4 5 1 34 1 0 17 6 111Stanford Hotel 263 80 4 5 1 34 1 0 17 6 111YMCA International House 765 255 2 1 1 6 2 0 2 6 087

eral different hotel categories make use of a simi- changes in font styles (size color and family)would not match this needlar set of values of font size and font family This

likeness does not extend to font color but the find-ings for attributes in this group are not significant MultimediaThis is not unexpected as the main function of ahotel website is presumably to provide informa- There are significant differences in number of

graphics and number of links used on web pagestion rather than to entertain An entertainment-oriented web page would use a wider variety of Specifically midpriced and luxury hotels have

significantly more graphics than their economycolors sizes and font types The primary need fora visitor to a hotel website is to browse the web counterparts Websites in the luxury category also

use 15 more graphics than those in the mid-page comfortably A wide variety and constant


Table 5

An ANOVA Table for Different Hotel Categories

Economy Midpriced Luxury(2ndash3 Stars (4 Stars (5 Stars Overall

n = 35) n = 20) n = 11) (n = 66)

Mean SD Mean SD Mean SD Mean SD F-Value Sig Post Hoc

Group 1 ColorColor difference 5704 17125 617 160387 599273 134706 58933 16154 0547 0582 mdashBrightness 17774 65682 2054 57968 170273 58914 18488 6298 1612 0208 mdash

Group 2 Font detailsFont color 200 1955 300 2077 136 1120 220 1947 3071 0053 mdashFont sizes 217 1790 275 1860 173 1009 227 1724 1395 0255 mdashFont family 114 0879 160 0940 109 0302 127 0851 2217 0117 mdash

Group 3 MultimediaNo of graphs 2037 16821 3910 25020 4509 45713 3017 27684 5499 0006 M L gt ENo of flashes 066 0873 050 1192 09 0302 052 0932 1575 0215 mdashNo of links 1409 15300 2960 23694 4036 21974 2317 21629 9362 0000 M L gt E

Group 4 OthersNo of tables 446 4061 1150 7709 1355 11877 811 7964 10382 0000 M L gt EPage size 124 0398 141 9527 181 0919 138 0580 4558 0014 L gt E

priced category Hotels in the midpriced category of hotel websites are displayed on 138 screensHowever the home pages of luxury hotels are sig-in turn use 92 more graphics than the economy

hotels Such a difference is probably due to the nificantly longer than the home pages of economyhotels More specifically the home pages of lux-fact that room rates in the luxury and midpriced

hotels are higher than the economy hotels Thus ury hotels are 28 longer than midpriced hotelswhich in turn are 14 longer than the economythese higher end hotels need more graphics to in-

troduce their facilities and services As with the hotels Again the reason is the same as for the useof multimedia and hyperlinks These hotels pro-number of graphics there is a significant differ-

ence among the three hotel categories in terms of vide a more information-rich websitethe number of hyperlinks In particular luxury ho-

Conclusionstels have 36 more hyperlinks than midpricedhotels and the midpriced hotels use 110 more Our analysis of hotel web pages leads to spe-hyperlinks than their economy counterparts The cific conclusions in the three areas of Backgroundreason for such a difference is similar to that of Contrast Detail of Text and Multimediathe multimedia However there is no significantdifference in number of flashesanimations Background Contrast

In this measurement the analyzed sites per-Tables and Page Length

formed similarly The largest contrast and bright-ness value between text and background are 765In the attribute of number of tables the luxury

and midpriced hotels have significantly more ta- and 255 but the W3C Working Draft (2000) pro-poses that values at least of 500 and 125 are ac-bles than the economy hotels Hotels in the luxury

category use 17 more tables than hotels in the ceptable In this study the overall mean values ofcolor contrast and font brightness are 58933 andmidpriced category and the midpriced hotels use

158 more tables than the economy hotels This 18488 respectively for hotel websites in HongKong This implies that the visibilities betweenshows that websites of hotels at the high end have

more information presented in tables to show their text and background colors of the categories ofhotels are good as a wholecustomers In relation to page size home pages

268 CHAN AND LAW

Detail of Text registers in which the websites are operating (ege-commerce informative government service)

The results however reveal that the hotel web- and in the features and attributes that might be in-sites use a very small number (ranging from 127 cluded in the evaluations Another possible area ofto 227) of font color size and family This makes future research would be to compare and contrastthe content information more presentable and styl- these automated findings with the intuitions ofish Such a pattern also matches the general con- users We should also investigate to determinecept of using less text in order to visually pleasing whether we would get similar outcomes if an en-viewers Although we can quantify the family tire website is analyzed rather than only the homesize and color the results show no clear indica- page whether websites in different languages ex-tion or direct relationship among the three catego- hibit similar behavior what the most desirableries of hotels This could be attributed to the fact level is for contrast between text and backgroundthat this group relates to the components of the and what exactly users look for on a hotel websitedetail of text and design of web pages

AcknowledgmentsMultimedia

The authors acknowledge the constructive com-There are significant differences in the way that ments provided by the three anonymous reviewers

different hotel categories used multimedia Luxury on the earlier versions of this paper This researchand midpriced hotels were found to have signifi- was partly funded by The Hong Kong Polytechniccantly more links graphs and tables than econ- University (Under contract No A-PA1B and 4-omy hotels The home pages of luxury hotels were Z09D) The authors would like to thank Carmenalso significantly longer than those of the econ- So and Bisney Lau for their assistance in softwareomy hotels These values could be because luxury development and data analysishotels offer a wider range of products and servicesto a larger and more demographically geographi- Biographical Notescally and economically dispersed customer base

Steven Chan is a Senior Lecturer in the Department ofFurthermore they also have the resources to pro- Computing at the Hong Kong Polytechnic University Hisduce more elaborate multimedia websites Inter- research interests are in E-business Web usability Net-

working Computer-aided learning and other IT applica-estingly midpriced hotels also have more tablestions in the computing industrylinks and graphs than economy hotels

In this research it was found that web pages ofhotels in Hong Kong usually emphasize functions Rob Law is an Associate Professor of Information Technol-

ogy in the School of Hotel amp Tourism Management at theover design Such an observation is probably dueHong Kong Polytechnic University His research intereststo the difficulty to quantify all components in de-are modeling and forecasting InternetE-commerce andsign and obtain enough meaningful data Still thetourismhospitality IT applications

findings from the use of multimedia number oftables and page length are useful for hoteliers as

Referencesthey look for ways to improve their websites

Barry C amp Lang M (2001) A survey of multimedia andweb development techniques and methodology usage

Future Work IEEE Multimedia 8(2) 52ndash60Bhatti N Bouch A amp Kuchinsky A (2000 June) Inte-

This initial study of automated website analysis grating user-perceived quality into web server designwas conducted in a focus area of Hong Kong hotel Computer Networks 1ndash16

Black R (1997) Web sites that work San Jose CAwebsites Still it has produced some useful find-Adobeings that are of interest both to researchers and

Buhalis D (2002) eTourism Information technologies forpractitioners and that could be generalized tostrategic tourism management New York Financial

many other sectors in tourism With this in mind Times Prentice Hallfuture work may increase the scope of the research Burns P (2002) Top 100 Internet mistakes you can not

afford to make Berkshire McGraw-Hill Professionalin the number of websites to be examined in the


Haffner E G Roth U Engel A amp Meinel C (2000) Lindsay W (1999 JulyAugust) Using color effectivelyin computer graphics IEEE Computer Graphics andWhat do hyperlink-proposals and request-prediction

have in common Proceeding of the International Con- Application 20ndash35Nicol J R Gutfreund Y S Paschetto J amp Rush Kference on Advances in Information System (pp 275ndash

282) S (1999) How the Internet helps build collaborativemultimedia application Communications of the ACMIvory M Y (2000) Preliminary findings on quantitative

measures for distinguishing highly rated information- 42 79ndash85Marcus A (1997) Graphical user interfaces In M Hel-centric web pages Retrieved August 20 2005 from

httpwwwcaviuniv- paris3frilpga ilpgasfleurysen- ander T K Landauer amp P Prabhu (Eds) Handbookof humanndashcomputer interaction (pp 423ndash440) Oxfordsnetbiblio-sensnetivoryetaihfw00pdf

Ivory M Y (2002) Characteristics of web site designs England ElsevierScharff F V amp Ahumada Jr A (2002) Predicting theReality vs recommendations Retrieved July 30 2005

from httpscholargooglecomscholarhl=enamplr=ampq= readability of transparent text Journal of Vision 2653ndash666cacheaVYOqiHm_UcJubit ischool washingtonedu

pubshcii03hcii03pdf Tourism Commission (2005) Tourism statistics Tourismperformance Retrieved July 25 2005 from httpwwwIvory M Y amp Hearst M A (2002) Improving web page

design IEEE Internet Computing 5(2) 56ndash63 tourismgovhkenglishstatisticsstatistics_performhtmlW3C Working Draft (2000) Test the color attributes visi-Kim L amp Albers M J (2003) Presentation information

on the small-screen interface Effects of table format- bility Retrieved August 28 2005 from httpwwww3orgTRAERThtmlcolorting IEEE Transactions on Professional Communica-

tion 46(2) 94ndash104 Wan H A (2000) Opportunities to enhance a commercialwebsite Information amp Management 38 15ndash21Law R Ho D amp Cheung C (2004) A study of the func-

tionality of hotel websites in mainland China and theUnited States Journal of Academy of Business and Eco-nomics 3(1) 202ndash209

256 CHAN AND LAW

Importantly then travel websites are often the Usabilityprospective visitorsrsquo first contact with Hong Kong

Users generally determine the usefulness of aTo date however there has been little attentionwebsite based on its usability or ease of use Thegiven to the important topic of the design and theusability of web pages has been measured in a va-impact of the design of tourism-related websitesriety of ways Ivory (2000) presented a web pageon usersmatrix composed of 42 features of a web pageThe Internet is transforming transaction pro-which can affect usability These 42 features werecesses in the tourism and travel industries makingthen grouped under three headings of page compo-them more competitive Buhalis (2002) indicatessition page formatting and overall page Most ofthat e-business can cope with the transformationthese features under page composition and format-from the traditional business operation Ultimatelyting (eg the number of words and font style) arethe efficiency and development of this medium arerepresented quantitatively Some of the more ab-directly related to the technologies and the inter-stract features clustered under an overall pageface through which the websites are developed Ithowever cannot be measured quantitatively some-is therefore important to establish and if possi-what compromising the objectivity of the metricble codify what would constitute good web page

design practice in the area of e-commerce in travelWebsite Structureand tourism It would certainly be useful if web

designers could refer to some commonly agreed Ivory and Hearst (2002) suggested measuresupon standards that allow the convenient objec- for assessing design quality and usability of web-tive and quantitative evaluation of important website structure that represent website design ele-site features (Law Ho amp Cheung 2004) Exam- ments within a triangular schema At the base ofples of these features are the appearance of text the triangle are text links and graphics Thesefonts color contrast and multimedia These stan- provide the foundation for elements at the upperdards can then be applied automatically levels but are most obviously operational in the

This article describes a study that makes an ini- interface Website elements that are assigned totial attempt to establish a standard for evaluating upper levels of the triangle schema such as for-the design of websites and an automatic process matting serve to assist users in forming a repre-for evaluating the design of web pages in the con- sentation of how to go about accessing a sitersquostext of hotels Although some prior studies such contentas the work by Ivory (2000) have evaluated webpages in the context of travel and tourism these Color Contraststudies were primarily related to or based on for-matting Modifying the approach that was initiated Color contrast has an effect on the readability

of a website and the combination of text color andby Ivory and Hearst (2002) a page format struc-ture that classifies the formatting of website into background color can persuade viewers to use a

website Color contrast can also drive users awayfour main areasmdashcontrast font multimedia andothersmdashis proposed in this research Figure 1 from using a website Lindsay (1999) suggested

that the colors used in computer graphics arepresents the proposed web page format structurebased on psychological andor biological consider-

Literature Review ations and that colors can attract users to browsethe website and to stay with it Researchers haveTo date work on website evaluation has mainly

focused on the areas of usability and of website proposed various guidelines for choosing the rightcolors For example Schraff and Ahumada (2002)structure More narrowly it has taken account of

basic design elements such as the use of fonts and recommended the use of plain backgrounds withmoderate-to-high-contrast text on patterned back-colors This section briefly reviews current web-

site evaluation metrics and related references in grounds Similarly Ivory (2002) discussed the im-portance of using a small number of browser-safethese areas


















258 CHAN AND LAW




















































260 CHAN AND LAW










Table 1


Hotel Website




262 CHAN AND LAW




















264 CHAN AND LAW






Color




Table 2


Group 1

Group 2


Font Details




Table 3


Group 1

Group 2


Font Details




266 CHAN AND LAW

Table 4


Group 1

Group 2


Font Details










Table 5



n = 35) n = 20) n = 11) (n = 66)


















268 CHAN AND LAW

















































258 CHAN AND LAW




















































260 CHAN AND LAW










Table 1


Hotel Website




262 CHAN AND LAW




















264 CHAN AND LAW






Color




Table 2


Group 1

Group 2


Font Details




Table 3


Group 1

Group 2


Font Details




266 CHAN AND LAW

Table 4


Group 1

Group 2


Font Details










Table 5



n = 35) n = 20) n = 11) (n = 66)


















268 CHAN AND LAW


































































260 CHAN AND LAW










Table 1


Hotel Website




262 CHAN AND LAW




















264 CHAN AND LAW






Color




Table 2


Group 1

Group 2


Font Details




Table 3


Group 1

Group 2


Font Details




266 CHAN AND LAW

Table 4


Group 1

Group 2


Font Details










Table 5



n = 35) n = 20) n = 11) (n = 66)


















268 CHAN AND LAW

































Table 1


Hotel Website




262 CHAN AND LAW




















264 CHAN AND LAW






Color




Table 2


Group 1

Group 2


Font Details




Table 3


Group 1

Group 2


Font Details




266 CHAN AND LAW

Table 4


Group 1

Group 2


Font Details










Table 5



n = 35) n = 20) n = 11) (n = 66)


















268 CHAN AND LAW
































262 CHAN AND LAW




















264 CHAN AND LAW






Color




Table 2


Group 1

Group 2


Font Details




Table 3


Group 1

Group 2


Font Details




266 CHAN AND LAW

Table 4


Group 1

Group 2


Font Details










Table 5



n = 35) n = 20) n = 11) (n = 66)


















268 CHAN AND LAW














































264 CHAN AND LAW






Color




Table 2


Group 1

Group 2


Font Details




Table 3


Group 1

Group 2


Font Details




266 CHAN AND LAW

Table 4


Group 1

Group 2


Font Details










Table 5



n = 35) n = 20) n = 11) (n = 66)


















268 CHAN AND LAW

































Table 2


Group 1

Group 2


Font Details




Table 3


Group 1

Group 2


Font Details




266 CHAN AND LAW

Table 4


Group 1

Group 2


Font Details










Table 5



n = 35) n = 20) n = 11) (n = 66)


















268 CHAN AND LAW
































266 CHAN AND LAW

Table 4


Group 1

Group 2


Font Details










Table 5



n = 35) n = 20) n = 11) (n = 66)


















268 CHAN AND LAW

































Table 5



n = 35) n = 20) n = 11) (n = 66)


















268 CHAN AND LAW
































268 CHAN AND LAW
































automatic website evaluations: the case of hotels in hong kong

Documents

Transcript of automatic website evaluations: the case of hotels in hong kong