5 4 I E E E S O F T W A R E J a n u a r y / F e b r u a r y 2 0 0 1 0 7 4 0 - 7 4 5 9 / 0 0 / $ 1 0 . 0 0 © 2 0 0 1 I E E E

the demise of e-commerce sites when sitesare too late, too buggy, or too complex forease of use.

Many Internet analysts correctly pre-dicted that a significant number of busi-ness-to-consumer sites would fail during theyear 2000 due to a lack of customer reten-tion and repeat sales. Webmergers esti-mated that 150 dot-coms failed during2000 and more will follow this year.3 Thosesites that continue to succeed have and willexpend significant resources modifyingtheir sites to improve customer retention.

Many of the dot-com statistics do nottake into account the global aspect of on-line marketing. The potential for financial

gain in a global market is great, yet little isknown about global ventures’ success ratesin terms of meeting customer needs on a lo-cal level. On a global scale, we could arguethat cultural diversity and sensitivity mustbe considered to ensure that the onlineshopping experience is the same for eachcustomer regardless of locality. The fierceonline competition that has led to the de-mise of poorly designed online sites nation-ally may occur globally if nothing is done toaddress global usability.

What can be done strategically to reachout to a global market? We propose the useof a Web-based usability assessment modelthat promotes customer satisfaction as an

focusA Global Perspective onWeb Site Usability

Shirley A. Becker and Florence E. Mottay,

Florida Institute of Technology

online business failures are increasing as customers turn away from Online

Online businessfailures areincreasing ascustomers turnaway fromunusable orunfriendly sites.From a globalperspective,usabilityrequires culturalsensitivity inlanguagetranslation,along with theappropriate useof color, design,and animation.

“There is a widening customer experience gap online. Companieswho bridge this gap will win.”1

Although many companies have succeeded in developing onlinebusiness applications, numerous others have failed. Many of thefailures resulted from a lack of corporate vision by not takingWeb usability into account. A study by Deloitte and Touche

stated that approximately 70 percent of retailers lack a clearly articulated e-commerce strategy and considered their site as testing the waters for onlinedemand.2 This corporate “build it and they will come” mentality has led to

usability engineering

integral part of online business applicationdevelopment. This usability assessmentmodel is an outgrowth of our collaborationwith industry in the pursuit of more effec-tive online development efforts. From aglobal perspective, our work is in an ex-ploratory phase. However, with the currentexpansion of online business applications inthe global market, we believe our assess-ment findings can be useful.

Strategic usability factorsThomas Powell4 formally describes Web

usability as allowing the user to manipulatethe site’s features to accomplish a particulargoal. The targeted customer assesses usabil-ity for simplicity, understandability, andease of use. The perception of usability is in-fluenced by user characteristics, such as gen-der, age, educational level, and technologyskills. Usability perception is also affectedby cultural differences associated with, forexample, design layout, use of color and an-imation, and information content.

We developed the usability assessmentmodel, which Figure 1 shows, to identifyand measure usability factors that impacta customer’s online experience. We’ve ex-panded these factors into more than 100usability elements, not shown for spacereasons, that have been used during us-ability assessments of commercial sites.5

The following usability factors are brieflydefined.

Page layoutPage layout is the visual presentation of

the Web page by means of backgroundcolor, white space, horizontal and verticalscrolling, font size and color, and other de-sign elements. The layout affects ease of useand quick identification of page compo-nents. Layout can be influenced by culturaldifferences in usability, such as the signifi-cance of a particular color, use of graphics(for example, country flags or symbols), ortextual organization (left to right or topdown).

NavigationNavigation is the navigational schema in

terms of breadth and depth of search pathsand traversal mechanisms. Simplicity is pro-moted through the effective use of links,frames, buttons, and text. Navigational

considerations, from a global perspective,include ready access to other country sitesfrom a home page (understandable in anynative language) or via a navigationalschema on each page. Figure 2 illustratesglobal aspects of navigation on a Web site.

Design consistencyDesign consistency is the consistent loca-

tion of page components within and acrosspages. Various components requiring con-sistency include textual descriptions, labels,prompts, and messages. Consistency ofcolor is required for links, background, andtext, among others. Design consistency pro-motes ease of use by applying a commonlook and feel to each page in a particularsite or across global sites. Figure 3 shows ahigh level of design consistency in Yahoo’svarious country Web sites.

Information contentInformation content includes timely and

correct error messages, prompts, button la-bels, textual descriptions, help, and cus-tomer service information. From a globalperspective, information translated fromone language to another should be gram-matically correct, not archaic, and appro-priate for cultural differences. Local termi-nology for a shopping cart, for example,includes shopping trolley and shopping bag.Figure 4 shows an example of effective in-formation content with buttons appropri-ately labeled for local use.

J a n u a r y / F e b r u a r y 2 0 0 1 I E E E S O F T W A R E 55J a n u a r y / F e b r u a r y 2 0 0 1 I E E E S O F T W A R E 55J a n u a r y / F e b r u a r y 2 0 0 1 I E E E S O F T W A R E 55J a n u a r y / F e b r u a r y 2 0 0 1 I E E E S O F T W A R E 55

Strategic goals• Customer satisfaction• Financial• Business process effectiveness• Learning and innovation Design

layoutNavigation Design

consistency

Informationcontent

Clearlylabeledfields

. . .Facilitation

of dataentry

Performance

Customerservice

Reliability

Security

Environment• Browser• Monitor• Modem . . .

User profile• Age• Gender• Computing skills• Native language . . .Localization factors• Reading• Language• Custom . . .

Usabilityassessment

Figure 1. The usabilityassessment modelincorporates usabilityfactors as well as the user profileand computingenvironment. All of these affect acustomer’s perceptionof Web site usability.

PerformancePerformance is measured according to

consumer wait and system response times.Currently, there is significant global dispar-ity in terms of modem speed and personalaccess to the Internet. Cultural sensitivitytranslates into sensitivity concerning down-load time. Performance-related cultural in-sensitivity is demonstrated by the high useof animation in many Asian and SouthAmerican Web sites affiliated with US com-panies (we found animation disparity forEuropean and Japanese-based companies aswell). Yet their North American and Euro-pean sister sites, where Internet access with

higher modem speeds is more readily avail-able, minimize the use of animation.

Customer serviceCustomer service is additional informa-

tion and support mechanisms that are read-ily available from the organization to en-hance the shopping experience. Thisincludes, for example, email and mail ad-dresses, phone numbers, and interactivechat rooms. It can also mean that help isavailable in a native language.

ReliabilityReliability is defined in terms of site

5 6 I E E E S O F T W A R E J a n u a r y / F e b r u a r y 2 0 0 1

Figure 2. Illustrationof navigationalaspects of globalusability. The worldmap supports globalnavigation by show-ing available countryWeb sites for aselected area on themap. The secondWeb page illustratesinconsistent globalnavigation. In termsof global usability,not all country Websites navigateconsistently to othercountry Web sites.(It’s possible that theWeb sites cited inthis article havesince changed.)

Figure 3. Illustrationof design consis-tency. Note that theGerman and Englishsite designs lookvery similar.

National Schema:This page allows a user tohighlight a region to display alist of countries for that region.

Navigation inconsistency:There is a link from the Swedish to the German and US pages butno link from the German to theSwedish page.

crashes, downtime, error messages, and con-sistent response times. A common usabilityproblem related to reliability results whenSQL, JavaScript, and other cryptic errorboxes are displayed to the end user. Anothercommon problem results from a miscalcula-tion in the number of hits during peak peri-ods of Web use. In terms of global-related re-liability, these problems will have a majoreffect on customer usability.

SecuritySecurity is concerned with privacy and

limited access to personal information. Thesecurity issues facing American consumersextend to customers worldwide regardingthe misuse and unauthorized distribution ofcredit card numbers, addresses, phone num-bers, income, and other personal data.

Other usability componentsOur usability assessment model includes

a user profile of the targeted customer baseand the customer’s computing environment,which is important in ensuring that modemspeed, browser type, and screen size aretaken into account during the assessmentprocess. A usability assessment also consid-ers other environmental factors. Moreover,the user profile and environment data mightneed to be localized based on a particularcountry’s or region’s characteristics.

The usability assessment model also in-cludes the organization’s strategic goals to

ensure that these are weighed during usabil-ity decision making. Typically, strategicgoals require a balance of financial, cus-tomer, business process, and internal learn-ing perspectives.6 Strategic goals will dictatewhether cultural sensitivity (driven chieflyby customer satisfaction goals) or culturalinsensitivity (driven chiefly by financial,time-to-market goals) take priority in the de-velopment of online business applications.

Country-centricity and usabilityAs a result of our study of usability asso-

ciated with US companies, we discoveredthat organizations tend to develop country-centric sites to support their global market.(We limited our study, and so our discus-sion, to US-centric usability, although theusability concept could apply to any coun-try.) US-centricity is imposing a Web usabil-ity look and feel from an American perspec-tive onto localized Web sites. The resultmight be an emphasis of English as a pri-mary language on all international Websites with little regard for native-languagesupport. The result might also include alack of concern (or awareness) for gram-matical inconsistencies or incorrect transla-tions to a native language.

US-centricity can come about unknow-ingly, for example, when an English-language Web site is directly translated intonative-language Web sites. Other possiblereasons for US-centricity are when a com-

J a n u a r y / F e b r u a r y 2 0 0 1 I E E E S O F T W A R E 57

Figure 4. Buttonlabels are appropriatefor local use of agiven Web site.

The user has the option of an English or Chinese version of thisWeb site. In either case, the buttonlabel for the other selection is written in the appropriate languagefor ease of use.

pany deems it economically feasible tomaintain only English-supported countrysites, translates one US-based Web designinto many international sites, or uses im-plied design standards regardless of culturaldifferences. Figure 5 illustrates this conceptof cultural insensitivity whereby site pagesfor global use are written in English.

Usability problems that we encounteredrange from simple grammatical mistakes tothe overuse of animation, which severelyslows download time. A number of US-centric usability issues can negatively affecta local customer’s online experience:

■ The use of culture-specific icons may beinappropriate, confusing, or unknownat a local level. A common example isthe shopping cart icon. Other countriesuse different terminology to representthe shopping container, such as a trolleyor a bag.

■ The use of a particular color for back-grounds, error messages, or textual in-formation may be inappropriate, con-fusing, or misleading. A color mighthave different meanings in differentcountries. The color red means error orwarning in the US although this isn’t thecase in Asian countries. One or morecolors might represent nationalism for acountry. Yellow, for example, is foundon many German sites, as this is a na-tional color.

■ Commonly used English words andphrases, as well as trademarks, are oftennot translated into the native language. Lo-cally, these words might be misunderstood,difficult to pronounce, or their meaningmight be unknown (see Figure 6).

■ Direct translation of English to a nativelanguage can result in unintuitive orconfusing labels and instructions. Onone particular site, the English word“map” was translated directly into theFrench word “plan,” which is not self-explanatory in French. Plan du site—plan of the site—would have been a bet-ter phrase for improved readability.Figures 7 and 8 show examples of Websites in which the direct translationmight affect local usability.

■ A main or home page for accessingcountry or regional sites is in English.The user must select a country or optionfrom a list of English words with notranslation support for the native lan-guage. (Some sites have remedied this inpart by providing a visual map of theworld, as Figure 2 shows).

■ The use of animation varies by countrysite. For several US companies, theirAsian and Central and South Americasites have significant animation whencompared to North American and Euro-pean sites. For several European andJapanese sites, the US site containedmore animation. Figure 9 shows an ex-ample of a European company withvarying degrees of animation associated

5 8 I E E E S O F T W A R E J a n u a r y / F e b r u a r y 2 0 0 1

Figure 5. Severalinternational sitesthat are in English.These sites illustratethe reliance on theEnglish language forinternational sites.The user would have to understandEnglish, for example,to ask for directionsin a native language.

Figure 6. An example of a site with potential for confusion:Selected English words are not translated into the nativelanguage. In this case, the English words are difficult topronounce and may not be understood in French.

The France site is inEnglish.

English wordshard to pronouncein French.

The site is in English.Directions provided for Euro-pean travel may be selectedin many different languages.

with its country sites.■ Navigational schema varies by country

site. Inconsistencies in navigation makeit difficult to traverse consistently acrosssites. Some country sites allow access toa home page; others allow access to aparticular region of the country, whileothers access all countries (Figure 2shows this limitation).

Usability strategiesIn pursuing a global market, organizations

should be sensitive to cultural differencesthat might impact usability. Several strategiesare available that can help with usability, de-pending on the organization’s goals.

Common designA general design layout, with little or no

customization for particular country sites,might reduce the cost of upgrades and main-tenance associated with multiple sites. Forcustomers accessing more than one countrysite, it provides design consistency for easeof use. It is also easier to enforce global de-sign standards in terms of the site’s look andfeel. Figure 3 illustrates this concept for Ya-hoo sites, which have a high level of designconsistency.

The risk associated with this strategy isthat usability can be degraded when gram-matical mistakes, missing translations, andinappropriate colors, for example, are intro-duced during site construction and mainte-nance. Usability assessments uncover theseproblems before they reach the customer.

CustomizationA lot could be learned about cultural

sensitivity, concerning global site deploy-ment, from the international marketingstrategies of McDonald’s and Coca-Cola.When visiting a McDonald’s in Aruba, forexample, there is a localized food item—barbeque chicken—not found on the NorthAmerican menu. Similarly, Coca-Cola lo-calizes the flavor of its products to maxi-mize global sales. This localization conceptcould be applied in the development ofglobal online business applications to en-hance global usability. The downside to de-veloping customized Web sites for eachcountry, however, includes higher develop-ment and maintenance costs when each siteis built and maintained separately.

Combined common and customized designThis middle-of-the-road strategy supports

design consistency across all Web sites whilecustomizing a particular Web site to meet thelocality’s cultural needs. By standardizingcorporate logos, nav bars, graphics, andother standard look-and-feel componentsacross all sites, companies support the us-ability goals of simplicity and ease of use. Bycustomizing colors, icons, graphics, andother Web components to meet a given coun-try’s needs, companies promote understand-ability and ease of use. Perhaps most impor-tant, however, is the appropriate use of thenative language for each respective Web site.

Applying the customization or the com-

J a n u a r y / F e b r u a r y 2 0 0 1 I E E E S O F T W A R E 59

Figure 7. Directtranslation withpotential negativeconnotation. In thiscase, the Frenchtranslation of theEnglish word has anegative connotation.Though its meaningis explained to theuser, there may stillbe a negativeimpression.

Figure 8. English words may cause confusion when inter-preted in a native language.

The US name “Escrow” may be interpreted as “fraud” in French. Though the French version of escrow is escroc, it is pronounced the same. The company explains the meaning of the word, but onestill has to question whether this will overcome thenegative connotation of the word.

Academic initiative is English but is also composed of two French words in reverse order. For a non-English-speaking French person, academic initiative can be understoodto have meaning but is grammatically incorrect.

bined strategy instead of a common design re-sults in higher development and maintenancecosts. The higher costs are justified, however,by customer satisfaction achieved with cultur-ally sensitive sites. Although more research isneeded, the national fallout of business-to-consumer Web sites to date tells us that fiercecompetition and customer satisfaction bothplay a critical role in online success.

Usability assessments: A studyMuch of our work on usability assess-

ments of US sites has focused on user profiledata that included age, gender, computingskills, and other commonly used marketingdata. When profiling the consumer for aparticular country, however, there is addi-tional information that would assist in de-veloping an effective online business appli-cation. From a global standpoint, a userprofile for a country should include the levelof understanding (or popularization) ofcommonly used icons (such as a shoppingcart), words (such as the GO button label),

6 0 I E E E S O F T W A R E J a n u a r y / F e b r u a r y 2 0 0 1

Figure 9. Animationand performanceissues. These twoexamples illustratecountry centricity interms of animationand the impact onperformance. In eachcase, the country-of-origin Web site hasless animation thanthe other countrysite.

Table 1Comparison of country Web sites for a software company. The study was conducted using a 56K modem, 15-inch monitor on a notebook computer,

and Microsoft’s Internet Explorer browser.Country Use of animation Horizontal scrolling Oversized graphics English content

(Scale 1 – 5)1 (Yes or no) (Yes or no) 2 (Scale 1 – 5)1

US No animation Yes No Not applicableAustralia No animation Yes No Not applicableSweden 3 No No 1France 3 No No 1Japan No animation No No 2China 3 No Yes 1 (button label GO)Brazil 1 Yes Yes 4

1 Likert scale where 1 is the lowest point of allocation and 5 is the highest. A 1 indicates low significance; 5, high significance.2 Oversized graphics waste valuable information space and require more vertical or horizontal scrolling to find information.

The download time for the US site versus the Argentina site is significant. The Argentinasite has significant animation, which the USsite does not have.

US site navigation is complex because of the extensive use ofanimation and frames. The Dutchsite has little animation and is verysimple in design.

and colors (such as red). A usability assess-ment, based on the model in Figure 1, canuncover this information.

To illustrate the importance of usabilityassessments in uncovering design flaws, wecompared seven country sites for a US-based, global software company. Table 1summarizes the results. The usability ele-ments included animation, horizontal scroll-ing, graphics, and English content.

It’s interesting that although these siteswere customized, each had usability prob-lems. The US and Australian sites did nothave animated components, thus minimiz-ing download time. However, both sitesmade use of horizontal scrolling, which neg-atively impacted readability. The China andBrazil sites had oversized graphics, whichwasted valuable information content space.All non-English sites had various amountsof English embedded in the text.

The company that we studied and summa-rized in the table is a large, well-establishedsoftware company selling multiple productsin an international market. Common aspectsof all the company’s sites included consistentuse of background colors, fairly consistentpage design, mixed English with native lan-guage, good use of vertical white space, andthe use of the folder design standard (popu-larized by Amazon.com design).

T he number of non-US customers us-ing online business applicationscontinues to increase very rapidly.To take advantage of this opportunity, com-panies must understand the target market interms of localized and common onlineneeds. In this respect, we have only just be-gun to understand the usability issues thatinfluence short- and long-term use of onlinebusiness applications.

We are developing a tool, an automatedenvironment, that will let users enter theirassessment of a particular Web page or site.The tool implements the usability modelshown in Figure 1. It supports data entry forone or more selected usability elements inorder to analyze the user’s perspective on

Web site usability. The tool’s report genera-tor allows for data analysis based on userprofile or environmental selection criteria.Our future endeavors will expand our toolto incorporate our findings on global us-ability for more effective assessments.

AcknowledgmentWe would like to thank Anthony Berkemeyer and

Natalie Roberts for their usability expertise and theirinvaluable assistance in uncovering global usabilityissues.

References1. M. Hurst and E. Gellady, “Building a Great Customer

Experience to Develop Brand, Increase Loyalty andGrow Revenues,” www.creativegood.com/creativegood-whitepaper.pdf (current 16 Jan 2001).

2. R. Spiegel, “Report: 70 Percent of Retailers Lack E-Commerce Strategy,” E-Commerce Times; www.ecommercetimes.com/news/articles2000/000126-1.shtml(current 16 Jan 2001).

3. J. Weisman, “E-Commerce 2000: The Year of LivingDangerously,” E-Commerce Times, 29 Dec. 2000;www.ecommercetimes.com/perl/story/6380.html (current 16 Jan 2001).

4. T. Powell, Web Design: The Complete Reference, Os-borne McGraw-Hill, Berkeley, Calif., 2000.

5. S. Becker, A. Berkemeyer, and B. Zou, “A Goal-DrivenApproach to Assessing the Usability of an E-commerceSystem,” Cutter Information Technology J., Apr. 2000,pp. 25–34.

6. R.S. Kaplan and D.P. Norton, The Balanced Score-card—Translating Strategy into Action, Harvard Busi-ness School, Boston, 1996.

J a n u a r y / F e b r u a r y 2 0 0 1 I E E E S O F T W A R E 61

About the Authors

Shirley A. Becker is a professor of computer science at the Florida Institute of Technol-ogy, Melbourne, and codirector of its Software Engineering Research Center. Her funded re-search includes Web usability and testing, Web-enabling tools and technologies, e-commercesystems development, and database systems. She recently served as editor of the Journal ofDatabase Management and serves on several editorial review boards. Becker received her MSand PhD in information systems from the University of Maryland, College Park. She is amember of the IEEE, the ACM, and the Association for Women in Computing.

Florence E. Mottay is a graduate student in software engineering and a research as-sistant at the Center for Software Engineering Research, Florida Institute of Technology, Mel-bourne. Her research interests are in software testing, formal languages, mathematical mod-els, and e-commerce. She was awarded for excellence in mathematics by the United StatesAchievement Academy (1997) and for academic excellence by the American Association ofUniversity Women (1998). Mottay received a BS in applied mathematics from Florida Instituteof Technology.

Contact the authors at the Florida Inst. of Technology, 150 West University Blvd., Melbourne,FL 32901; abecker@cs.fit.edu; fmottay@fit.edu.