Interactivity Reexamined: A Baseline Analysis of Early Business Web Sites

Article excerpt

The World Wide Web is a networked hypertext system containing digitized texts, audio, and visual data (Snyder, 1996). Fueled by increasing promotion in mass media and the popularity of online services (Internet World, 1997; Maddox & Mehta, 1997), the Web, as part of the Internet, is growing at a pace faster than any previous new communication media (Berthon, Pitt & Watson, 1996). Although it has been available to the general public only since the early 1990s, the Web is now regularly accessed by millions of users. While different sources report different numbers of Web users, all reports agree that Web usage is growing tremendously. One study conducted in August, 1997 estimated that there were 55.4 million Web users over the age of 12 in the United States (Relevant Knowledge, 1998). This is up from 22 million users the previous year. Many traditional mass media (such as television, radio, newspapers, and magazines) have established their presence on the Web through online mirror versions.

In the latest edition of Diffusion of Innovations (1995), Rogers discussed the multiplier effect of Internet technology upon its adoption of its interactive nature. The rapid adoption of the Web can be explained by the attributes of successful innovations suggested by Rogers (1983). In terms of relative advantage, Web users can enjoy business profits and a reputation as being informed and knowledgeable. Its graphic and audio-vidual capabilities make it it an attractive medium to use. In terms of compatibility with previous ideas and social values, the Web is a combination of traditional print and electronic media. The name of the technology, the World Wide Web, easy to remember and vividly illustrating the worldwide interconnection function of the system, also facilitates its success. In terms of low complexity, the what-you-see-is-what-you-get format of web browsers makes them user friendly and reduces the perceived complexity of the medium. Finally, the Web has high trialibility. Trialibility is the degree to which an innovation can be sampled in small quantity or with low cost. Open access to the Web through computers at schools or public libraries minimize the risk for computers in trying the technology.

A major force behind the growth of the Web is its potential for generating revenue. A recent report by ActivMedia (1998) on 3,500 net markets estimated that the revenues generated from the Web would be around $24.4 billion US in 1997. Most of the revenues (85%) came from product and service sales or fees and online advertising. The rest came from equipment and web site development. Some 65% already claimed to be enjoying profits. Moreover, web sites with more experience on the Web were more likely to report making profits than those who were new to the Web. Although one may doubt the validity of the statistics because of the study sponsor's vested interests in the Internet, the proliferation of Internet-related publications and research indicates that the Web has become a viable business opportunity that many want to cash in on. The growing interest of business in establishing their presence on the Web is shown in a 1996 study of 367 marketing executives in the U.S. Seventeen percent of them reported that their companies already had web sites, and 31% of the companies were planning to set up a web site in the next six months (Paustian, 1996).

The present study first attempts to deconstruct the meaning of interactivity and then reports the results of a content analysis which examined the interactivity levels of business web sites. Business web sites are chosen for study over other web sites to assess interactivity because these sites are the most common. They are most likely to benefit from interactivity and possess financial resources that drive the technological development of the Web. Indeed, as of April 1997, 88% of all registered domain names on the Web were commercial domains ending with ".com" (Kosters, 1997).

The World Wide Web as a Medium

Although the Web is still a medium accessible only to computer users, it has reached the "critical mass" threshold suggested by Rogers (1995) to assure that its adoption rate will become self-sustaining. In addition, the Web can be accessed using a traditional TV with a special accessory through TV-based online services such as WebTV or NetChannel. The device, which resembles a remote-control, is available from major electric appliance manufacturers such as Sony and Magnavox. Although WebTV has only 150,000 subscribers after its first year in the market (Magill, 1997), the alternative of accessing the Web through traditional TV sets will facilitate Web penetration rate among the general public.

Unlike traditional mass media which represent a one-to-many communication model, the Web represents both many-to-one and many-to-many models (Hoffman, Novak, & Chatterjee, 1995; Morris & Ogan, 1996). Many individual consumers can initiate communication to the same web site at the same time. This many-to-many scenario is unique to the Web, because many points of origination and destination coexist in cyberspace. With such features as cyberchat and listserv mailing list that simultaneously connects people with common interests, the Web becomes a means for many-to-many communications. There is no single source of message origination or single destination on the Web. Negroponte (1996) describes such use of computer networks as a "digital age" in which all information is digitized. Information is customized to the demand of the consumer. Gilder (1990) predicts the new computer technology will increase the power of the people by "blow[ing] apart all the monopolies, hierarchies, pyramids, and power grids of established society (p.31)." The medium will change from a mass-produced and mass-consumed commodity to an endless feast of niches and specialties.

As a marketing communication medium, the Web has been described as "a cross between an electronic trade show and a community flea market (Berthon, Pitt & Watson, 1996, p.44)." It allows visitors to browse a company's products or services at the user's own pace and facilitates informal communication between the company and the consumer. Consumers who do not like face-to-face communication in real trade shows and flea markets can avoid it by browsing the Web while enjoying the same experience. A web site can perform many different functions for a business, its consumers, and other stakeholders such as investors and employees.

Among the most widely used marketing functions of the Web are: 1) to provide consumers with continuously updated product information without limitation of space; 2) to generate qualified leads for salespeople by identifying customer queries; 3) to support customer service so that customers can contact the company anytime they wish with complaints nd suggestions; 4) to serve as a customer research tool, collecting consumer information by conducting surveys and monitoring visitor behavior on the Web; 5) to conduct sales promotion activities such as as giving away samples of computer software through free download services, distributing electronic coupons and inviting customer participation in sweepstakes; 6) to distribute and accept orders from visitors during the web site visit; 7) to hold the attention of consumers through an interactive input-output process and customize communications precisely to individual consumers (Balthazard & Koh, 1997; Berthon, Pitt, & Watson, 1996; Hoffman & Novak, 1996; Stern, 1995); and 8) to project a favorable corporate image as a high-tech and consumer-oriented company (Maddox & Mehta, 1997).

The Concept of Interactivity

Interactivity is a critical concept in computer-mediated communications, because it is seen as the key advantage of the medium (Morris & Ogan, 1996; Pavlik, 1996; Rafeli & Sudweeks, 1997). Technologists such as Ted Nelson (1990) suggest that human-computer activities exemplify the human impulse to create interactive representation. The outcomes of interactivity are engagement in communication and relationship building between a company and its target consumers. Researchers have noted that the quest for improving interactivity guides future technological development for the Web (Robb et al., 1997).

From an interpersonal communication perspective, interactivity has been defined as "the extent to which messages in a sequence relate to each other, and especially the extent to which later messages recount the relatedness of earlier messages (Rafaeli & Sudweeks, 1997 p.3)." By the same token, many communication researchers use face-to-face communication as the standard of interactivity and evaluate the interactivity of mediated communication (such as the Internet) by how closely it simulates face-to-face communication (Walther & Burgoon, 1992). This conception ignores the characteristics of computer-mediated communication which allow asynchronous communication. Participants may choose the time and the duration of interaction. From a mechanical perspective, interactivity has been defined as "the extent to which users can participate in modifying the form and content of a mediated environment in real time (Steuer, 1992 p. 84)."

Using an artistic approach, Laurel (1996) contends that interactivity is an experience like acting in a theater. Many agents participate within representation contexts in computer-human interactions.

In a business setting, interactivity tends to be seen as the "combination of rich content, active intelligence, collaborative communications to create a compelling consumer experience (Robb et al., 1997 p.5)" or "a person-to-person or person-to-technology exchange designed to effect change in the knowledge or behavior of at least one person (Haeckel, 1998 p.64)." These approaches to interactivity can easily lead to subjective interpretations of the nature of interactivity, because individuals have different perceptions of richness, engagement, fantasy, or relationship with previous messages. Moreover, Steuer's mechanical perspective is not precisely applicable to the Web. For instance, if an individual visitor writes a message to the Webmaster, he will not be considered interacting with the web site, because he does not modify the form or content of a web site.

Despite its importance, discussions of interactivity have been filled with a restrictive assumption that requires reexamination. This assumption is that reciprocal, two-way communication is a common desire of both the communicator and the audience. Rogers (1995) defines interactivity as "the degree to which participants in a communication process can exchange roles and have control over their mutual discourse (p. 314)." Rafaeli & Sudweeks (1997) describe interactivity as "a condition of communication in which simultaneous and continuous exchange occur, and these exchanges carry a social, binding force (p. 4)." These definitions, emphasizing "exchange" and "mutuality," assume the audience is interested in participating in conversations with the communicator. In a business setting, this implies that all consumers want to communicate with companies.

Studies of computer-mediated communication audience behaviors have shown this to be an invalid assumption. In electronic discussion groups, some members are quiet observers and "lurkers" who never participate while other members are active participants who contribute frequently to the discussion (Ha, 1995). Reasons for using the Web may be different from person to person. Some people are goal-directed and may want to complete a task through visiting a web site; others simply may be Web surfers who are only curious and relish the fun of finding out what is on the Web (Hoffman & Novak, 1996).

Companies are assumed to be interested in interactive communication with consumers. The advocates of relationship marketing contend that more communication between the consumer and the company will build the relationship between the two, and hence ultimately result in higher sales (Haeckel, 1998; Hoffman, Novak, & Chatterjee, 1995). This profit incentive, it is argued, will lead a company to emphasize interactivity.

The Five Dimensions of Interactivity

Discarding the unrealistic notion of mutual interest in two-way communication, this study proposes that interactivity should be defined in terms of the extent to which the communicator and the audience respond to, or are willing to facilitate, each other's communication needs. Thisefinition accommodates individual differences in communication needs. Sometimes, the audience wants only low levels of communication, having the freedom to navigate within a web site and the fun of selecting different options without direct contact from the company. Sometimes, the audience wants immediate assistance from a company, such as technical support information to solve a problem. Given these constraints, interactivity may be perceived to consist of five dimensions capable of fulfilling different communication needs: 1) playfulness, 2) choice, 3) connectedness, 4) information collection, and 5) reciprocal communication.

Playfulness. Play has been advocated as one purpose of communication (December, 1996; Stephenson, 1967). Play is an interlude from work and is a voluntary behavior (Stephenson, 1967). Information technology enhances and alters the entertainment experience of audiences (Bryant & Love, 1996). The Web, as a computer device, can perform many input-output functions with the click of the mouse to enhance playfulness and entertainment value. Such input from the consumers is considered the essence of interactivity by web designers (Digital Output, 1997).

Toys and games have been human companions since prehistory and are childhood phenomena of all cultures and civilizations. Indeed, the excitement of psychological gratifications in winning a game may explain why both adults and children spend a large amount of time in playing games, electronic and otherwise. The desire to demonstrate one's competence is an important motive in play and in games (Holbrook et al., 1984). By displaying mastery and control in games, one can reinforce self-esteem (Wilson, 1981) or sense of existence (Stephenson, 1967). Csikszentmihalyi (1975) describes the game player's feelings of satisfaction as a "flow experience," so engaging that he or she ignores everything else. After studying consumers' evaluations of five different web sites, Eighmey (1996) concluded that a successful web site must combine both entertainment and information to add value in the eyes of the consumers.

The games on many web sites are very similar to video games with which many individuals are familiar, and in which many people have indulged. The fun of the games is the sense of success enjoyed by the player. Moreover, curiosity arousal devices such as questions and answers on a web site are similar to popular quiz shows such as Jeopardy.

These games and curiosity arousal devices on the Web tend to be solitary games for individuals. Their presence provides a playful environment in which an audience member can communicate with himself or herself. As Stephenson suggests in his Play Theory on Mass Communication (1967), mass media content is play for the audience, not just information. Play is an inner talk or conversation within oneself that provides pleasure for an individual. Strictly speaking, the playfulness dimension of interactivity is within oneself rather than with another person. Yet the communication need of an audience member on many occasions represents a desire to communicate with oneself rather than with others. To the extent a communicator is able to electronically satisfy the self-communication needs of the audience, the games and other curiosity arousal devices on web sites qualify as interactivity devices.

Choice. The choice dimension of interactivity may be seen as consisting of the availability of choice and of unrestrained navigation in the cyberspace. Choice is closely related to the first dimension of interactivity, playfulness, because it also engenders an internal emotional sense of satisfaction. Steuer (1992) and Laurel (1991) share similar notions of choice as a dimension of interactivity when they discuss the "range" of interactivity as the amount of variation possible within each attribute of a mediated environment. Yet "range" is not a good descriptor of the choice dimension, because choices available to Web visitors may be discrete, rather than continuous.

As a result of perceived choice availability, an individual may feel empowered; able to choose from among several different available alternatives (Gilder, 1990; Pavlik, 1996). No obligation is undertaken when a navigator wanders through a web site and has the option of terminating the communication at any time. Choice may also be associated with minimizing effort in the achievement of a task. For example, by providing the option of choosing a particular language when navigating, a web site accommodates the native language of visitors. Providing a choice of text and graphic browsers allow visitors with different web browsers to access the full content of the web site. Thus, site visitors will not feel disadvantaged when they encounter the technical requirements of the top-end technology used by the website. As they encounter different alternatives during the navigation process, visitors are greeted with friendliness and feel respected. When the visitor is contented and feels empowered, he will spend more time at the web site, exploring alternatives and absorbing the materials.

Connectedness. For site visitors, hypertext in web sites can create a feeling of connectedness to the world by allowing visitors to jump, with little effort, from one point in cyberspace to another (Franks, 1995; Krol, 1996; Snyder, 1996). Such connectedness is the feeling of being able to link to the outside world and to broaden one's experience easily. With appropriate mapping of hypertext and images, visitors can interact with web site content as if physically present in a natural environment (Steuer, 1992). In a web site that simulates a showroom, for example, a visitor can feel virtually present and have questions answered with the click of a mouse. S/he is no longer confined to her/his study room or lab, but rather connected to the outside world. Robb and his colleagues (1997) describe this as a state of "rich content." Companies, by providing eye-opening connected experience at their web sites, fulfill individual information needs of consumers and engender trust. With minimal effort required to access bits of information, consumers can learn much faster through onvenient exploration.

The connectedness dimension of interactivity may accumulate over time. Walther and Burgoon's (1992) experiment compared computer users with face-to-face communication groups in completing three decision-making tasks. Their results suggest that the lack of non-verbal cues in computer-mediated communication can be compensated by computer users' accumulation of experience over time. They can understand their communication counterparts as well as in face-to-face communication after acquiring experience with the context of the communication. With the advantage of asynchronous interaction, more time is given to develop relationships among computer user groups than in face-to-face groups. In addition to textual information, video-clips, audio-clips, and animated graphics, the Web can enhance the feeling of connectedness by showing non-verbal cues such as action, facial expression, and tone.

Information Collection. Data gathering is primarily a communication need of the communicator. It is becoming more important to companies as they build databases about their customers and adopt the practice of integrated marketing (Blattberg & Deighton, 1991). With more information about audience, an organization can tailor messages to the interests and prior knowledge levels of the audience. In mass media industries, audience measurement is a term used to describe the process by which the communicator systematically collects data about individuals who consume media. Such information generally includes demographics, psychographics, and sometimes personality characteristics of the audience.

Information collection on the Web takes more varied forms than in traditional media. It can be in the form of admission requirements such as visitor registration, or it can be collected automatically without the awareness of the visitor as with cookie files (Dreze & Zufreyden, 1997). Cookie files are packets of data transmitted by a Web server to the hard drive of a user's computer. They store the user's ID or internet address when the user logs onto the Web server and provide information on the user's prior pattern of visits (Leibrock, 1997). Browsers only alert the user that the web site has a cookie file coming from the site's server. If the user refuses the cookie, he cannot continue the visit. Therefore, the information collection dimension of interactivity consists of audience's willingness to provide information or the automatic recording of audience data. Although it is possible for businesses to sell another institution information collected from consumers, it is beyond the scope of this study to address the ethical issues relating to such practices.

Reciprocal Communication. In the traditional mass media, communication is usually "one-way" - the communicator disseminates messages to a large audience with no expectation of feedback from the audience. Because of its interactive capability, a web site can be perceived as an invitation for visitors to do something (Sterne, 1995). A reciprocal unication loop can begin with the consumer's initiation of a conversation with a company by visiting a web site and sending a message to the Webmaster. Or it can begin with the company's provision of information and other content to consumers at a web site. The company expects response and feedback from the web site visitors in return for this content. To involve consumers and encourage feedback, a company must provide content useful to the consumer (Internet Marketing and Technology Report, 1997).

The more reciprocal communication between the site visitor and the web site owner, the more the site can respond to the particular needs of visitors. For example, if the visitor is only interested in compact cars, the site can be customized to the visitor's interest by sending only information on compact cars. Such dialogues between the communicator and the audience are called "collaborative communication (Robb et al., 1997)." At the extreme, it may be difficult to distinguish between the communicator and the audience, because both have the power to initiate contact and to receive messages, in a reciprocal relationship, there is an initiator who is expecting a return. In business settings, the initiator is usually, but not necessarily, a company.

Among the five dimensions of interactivity, one may consider some dimensions such as information collection and reciprocal communication as higher levels of interactivity, because they involve direct, two-way exchange of messages between the communicator (source) and the audience. These dimensions of interactivity are also source-oriented, because the source is the major benefactor of that interactivity. Dimensions such as playfulness, choice, and connectedness may be deemed self-communications that have no direct bearing on the source (company). They can be considered audience-oriented interactivity, because the audience plays a major role in the communication process: the web site provides the device necessary for the audience members to meet individual needs. To the company, the audience-oriented dimension of interactivity is bait to lead visitors to source-oriented interactivity. General consumers may appreciate audience-oriented interactivity more than source-oriented interactivity, because source-oriented devices involve such risks as infringement of privacy or disclosure of identity. Hence, in examining dynamics of interactivity in web sites, all these dimension must be considered.

Research Questions

The purpose of this study was to examine the dimensions of interactivity practiced in a sample of business web sites. Interactivity has been defined in terms of the five dimensions discussed above. Some dimensions of interactivity may be evident more often during a particular site visit than at other times. Because high-level interactivity involves risk and effort for the consumer, some opportunities may be considered more favorable by Web designers for inducing it. A business web site may score high in some dimensions but low in others. The score in each dimension may be influenced by factors such as the nature of the business and the intended function of the web site. For example, if a site is created for conducting electronic commerce, a high score on reciprocal communication and connectedness can be expected to allow visitors to conduct business transactions on the Web. Since we were unable to determine the intended functions of the web sites, we limited our explanation of the differences in interactivity dimensions to business types. Three questions relating to the dimension type and the timing for high levels of interactivity dimensions were posed in this study:

   [RQ.sub.1]: Which dimension of interactivity was more prevalent in early
   business web sites? How did business sites far in interactivity?

   [RQ.sub.2]: When did high levels of interactivity occur -- before or during
   a web site visit?

   [RQ.sub.3]: Were there significant differences between business types and
   their interactivity scores?

Method

A content analysis of business web sites was conducted in order to examine the five dimensions of interactivity. The web sites were categorized into eight formats according to the classification suggested Hoffman et al. (1995). An "other" category was added for sites that could not be categorized according to this scheme.

Three types of businesses were identified in this study: 1) manufactured goods, 2) services, and 3) retail outlets. Manufactured goods were defined as anything tangible produced under one company that resulted in ownership of a product such as Acura cars or Kelloggs' cereals. Services were construed to be anything intangible where the process of manufacturing and consumption occur simultaneously such as in the case of restaurants, insurance companies, banks, and so on. Retail outlets were identified as any businesses that involve the sale of different brands of goods and services to consumers for their personal, family, or household use. Examples would be institutions such as Barnes & Noble and Kmart.

Dimensions of Interactivity

Interactivity in web sites is made possible by interactive devices such as chat and games which require extra programming effort (Wacksman & Cohen, 1997). The interactive device must be available and visible in a web site. Therefore, the measurement of interactivity of a web site begins with the presence of interactive devices for each dimension of interactivity.

Playfulness. The playfulness of a web site was measured by the presence of curiosity arousal devices and games. Curiosity devices were defined as those which attract the attention of visitors and entice their participation during the visit. The question and answer format (Q & A) is an example of a curiosity arousal device, because when a company poses a question, site visitors can become eager to know the answers and to learn whether their answers are correct. Games are devices that must be played by rules and are based on visitor competence or skill to win a prize or to achieve a score. In the present study, sweepstakes (which are based on luck), were not considered games.

Choice. The presence of choice of color, speed, language, and other aspects of non-informational alternatives were counted as choices given to the audience. Only choices that provided clear directions on the web page were included in the study.

Connectedness. An interactive site should provide highly connected information about the product, the company, third-party information, and other content of interest to the visitor. This kind of information interdependency is a unique characteristic of the Web. Hyperlinks, which accommodate this connectedness, were defined as the underlined texts or highlighted items of a web page, when clicked with a mouse, open another web page. Four different types of hyperlinks were considered: 1) self-product related, 2) company related, 3) third-party related, and 4) other information. Hyperlinks within the same site were subdirectories of a home page. These could be read by scrolling down a page without changing the main URL address. For example, is a hyperlink of the main address Repetitive links to the same page were not counted. Hyperlinks to other sites were counted as links to sites with no relationship to the currently viewed page and having a totally different address. For example, under Rubbermaid's web site at one can find its division, Littletikes, as a hyperlink. When a visitor clicks on the hyperlink, he or she is transported to

Integratedness of the hyperlinks was measured by the sum of the presence of 1) self product-related hyperlinks, 2) company hyperlinks, 3) third-party product-related hyperlinks, 4) hyperlinks to the same sites, and 5) hyperlinks to other sites. Data were coded so that the maximum possible score was a five, and the minimum score was zero. Highly integrated sites were sites that scored four or above, which meant that they provided interconnected information not only from the company itself, but also hyperlinks to third parties to provide additional information to the visitor.

Information Collection. Information collection was measured by the presence of monitoring mechanisms. Monitoring mechanisms were classified as any explicit means by which a web site operator can record who has visited the site. One common monitoring mechanism is registration at web sites (CASIE, 1997). A registration requirement before admitting a visitor to browse a site was considered as monitoring before usage. A request for visitor's information for viewing particular portions of a site was categorized as monitoring during usage. Other monitoring mechanisms that were counted in this study included devices such as counters which displayed the number of visitor to a site or devices through which a visitor could retrieve the traffic flow statistics of that site. However, unobtrusive monitoring mechanisms such as cookie files were not measured in this study, because Netscape Navigator, the most popular browser at the time of the study and used by the coders, could not alert them to the existence of cookie files. Therefore, we could not assume users had the ability to detect cookie files.

Reciprocal Communication. The reciprocal communication dimension of interactivity was measured in terms of the presence of response mechanisms on a web site. Response mechanisms were any means through which the visitor could communicate with the web site owner. In this study, five response mechanisms were analyzed: 1) the e-mail address of the Webmaster or customer service representative as a hyperlink, 2) provision of a toll-free telephone number, 3) order or purchase mechanisms, 4) surveys or solicitation of information from visitors, and 5) other devices through which consumers could respond to the web site owner or discuss with other consumers such as chat rooms. Reciprocal communication differs from information collection in that in information collection, the visitor must provide the information either unaware of it (as in unobtrusive cookie files) or they will not get access to a certain portion of the web site as in registration. In reciprocal communication, the visitor can choose to give information and will not lose anything for not providing the information.

Since computer technology develops much faster than any researcher can write about it, some of the web sites analyzed in this study would have changed their content completely or perished by the date of this report. This study should be viewed as a benchmark study on early business web sites. Because the study was unable to randomly select web sites from a nationally representative sample, the results here should be considered preliminary. The present data provide broad perspectives about interactivity and its dimensions. Readers should be cautioned that the low level of interactivity found in some of the web sites may have been improved as a consequence of technological advancement. Whether the web site originated from computer engineers, marketing professionals, or corporate communications personnel may also affect the content of the sites. It is beyond the scope of this study to examine such issues.

Sampling and Procedures

The sample was a census of all 110 business web sites listed in the archives of the Web Digest for Marketers at between October 1995 to January 1996. The time-frame was chosen to represent the early stage of business presence on the Web as the trade press began to regularly devote special section to the topic of Web marketing. The sites included in the Digest were new business web sites of interest to marketers. The Digest is an information service provided by an online marketing consultant, Chase Online Marketing. At the time of this study, the Digest had no vested interest in promoting a particular site. It was the only available list at the time systematically tabulating newly launched business web sites. The archive consisted of weekly web site reviews. Each review listed the URL address of the business site and a brief description. The content analysis was conducted between February 1996 and April 1996. Trained coders were instructed to open the site location from within the archive rather than typing the address individually in order to minimize errors in typing the URL address.

The unit of analysis was the home page, because it served as the front door of the entire web site. Usually, the home page is named as or as a default page of the domain name such as Most visitors to a web site decide whether they will continue to browse a site based on their impressions of its home page. Moreover, web sites vary substantially in size. For example, Web Techniques (1997) estimated that web sites range from one page to 50,000 pages. Home pages therefore provide consistency across the sample, since all units were a single page. Further, coding an entire web site could be extremely time-consuming and introduces biases based on size. Although companies may put their interactive devices in their web sites other than in their home pages, we cannot assume that visitors have the patience to go through the whole web site (particularly if they contain over 100 pages). This study analyzed not the whole web site, but rather the key part of a web site represented by its home page. In analyzing the home page, coders opened all the hyperlinks in the forms of underscored text, icons, or pictures to determine the nature of the hyperlinks.

In light of the skills required in using the World Wide Web, seven independent coders already proficient in using the Web were trained to code the web sites over a period of three weeks. Four coders were from a large upper Midwestern state university while the other three were from a Western state university. The quality of collected data could be affected by fatigue and boredom resulting from the high volume of coding tasks. Therefore, each coder was assigned to analyze only a randomly selected list of 20 designated web sites (including two pretest web sites and three post-test sites).

The pre-test post-test coder reliability testing method used by James & Vanden Bergh (1990) was employed in this study. The seven coders all coded the same pre- and post-test sites. The major investigators randomly selected two web sites (Dunlop and BMW) as pre-test sites. Three other sites (New Balance, Rubbermaid, and Epson) were designated as post-test web sites. First, all coders collected data from the two pre-test web sites. These coding sheets were turned into the principal investigators for quality control analysis. Additional instructions were given to the coders after identifying possible areas of difficulty. All coders were instructed to print the home page of the web site before coding so that if mechanical failure occurred, coding could continue based on the printout. All the coding procedures were completed within two months after the pre-test.

Perreault and Leigh's (1989) Reliability Index(1) was used to compute coder reliability. The advantage of the Reliability Index is that it does not have a multiplicative chance agreement assumption and explicitly measures the level of agreement that might be expected by a true level of reliability (Perreault & Leigh, 1989). In this study, the pre-test coder reliability was .94 for business type, .91 for the presence of games, .91 for the presence of monitoring mechanisms, and .70 for the presence of response mechanisms. The post-test coder reliability was generally higher with .96 for business type, .94 for the presence of games, .87 for the presence of monitoring mechanisms, and .64 for the presence of response mechanisms. One possible explanation for the lower coder reliability of response mechanism among the five dimensions is that the response features were displayed subtly in some web sites making them difficult for coders to identify.

Results

All 110 web sites were successfully accessed by the coders, although some required a second or third attempt. Among the sites, 55% were for manufactured goods, 32% were for services, and 14% were for retail outlets. Information sites were the most popular format among business web sites (30%), followed by incentive sites (26.4%), mall sites (14.5%), and online storefronts (12.5%). Different business types displayed different preferences for web site formats. Manufactured goods (33.3%) and services (31.4%) tended to have a much higher proportion of information site formats than retail outlets (13.3%), while retail outlets appeared most likely to employ a mall format (33.3%). The low percentage of image sites (2.7%) and flat ads (2.7%) show that very few companies used the Web for traditional image advertising.

Playfulness and Choice in Business Web Sites.

Not many business web sites were playful. Curiosity arousal devices were more commonly used than games to create a playful environment. Almost one quarter of the web sites contained some sort of curiosity arousal devices to attract visitors. Only about one-fifth of the sites contained some sort of game for consumers to play. In terms of choice, business web sites generally scored low in providing options to consumers. Very few sites provided a choice of text or graphics browser, for instance. Color was the dominant choice feature (43.6%), followed by speed (14.5%). The least common choice offered to visitors was language (9.1%).

Connectedness in Business Web Sites.

One interactive device unique to the World Wide Web is the hyperlink. Many business sites analyzed in this study contained more than 13 non-repetitive hyperlinks just in their home pages. But when the nature of the links was further investigated, most of them were self-promotional. Links within the same site were highest in number (M=12), followed by links about the product or service itself (M=5), and links about the company (M=3). Links to third-party information about the product and other information were rare with no more than two links per site average.

Information Collection in Business Web Sites.

Both integrated marketing communication and database marketing require rigorous consumer information collection. Although they are increasingly popular as marketing practices, our study found that 81% had no explicit monitoring devices at all. Eighty-six percent of the web sites had no registration procedure. Among those that did employ such procedures, registration before visits were two times more frequent than registration during visits. The higher occurrence of registration before a visit was no surprise, because after admission to a site, consumers have little incentive to reveal their identity or disclose other information.

Reciprocal Communication in Business Web Sites.

Despite the ease with which a web site can be used as a two-way communication tool between company and visitor, this study found that 38.2% of web sites had no explicitly displayed consumer reciprocal communication devices. Among those that did provide consumer reciprocal mechanisms, the most commonly employed device was the e-mail option (46.4%). Not many companies saw their web sites as direct marketing tools. Only 24.5% provided a toll-free phone number for ordering or inquiry. Some 22.7% had order mechanisms on the Web. Cross-media reference was seldom used. Surveys about consumers were not common as only 11% of web site home pages ask consumers to complete a survey for them. Among the few web sites that used other response mechanisms, visitors were asked for specific information or were invited to participate in some kind of online chat device such as the Hyundai World Cup in Korea site's "fanletter" and Prudential's "communication e-mail" to other Prudential customers.

Among the web sites considered in this study, the most prevalent interactivity dimension was reciprocal communication. This occurs in slightly more than half of the web sites. The other 40% of web sites did not contain such mechanisms. Only one-fifth of the web sites had explicit information collection devices (19.1%)

To understand how different types of businesses fared on interactivity, an interactivity score was computed for each web site by summing the presence of games and curiosity arousal devices (playfulness), choice, highly integrated hyperlinks (connectedness), monitoring devices (information collection), and response devices (reciprocal communication). One-way ANOVA was used to examine the differences between these interactivity scores among manufactured goods, services, and retail outlet web sites. Results show that the three types of businesses were similar in one another in entertaining and engaging consumers: no significant differences were found between the sites in their playfulness and choice dimensions. Products, services, and retail outlets differed significantly on connectedness, information collection, and reciprocal communication dimension of interactivity. Specifically, manufactured goods (M=.26) were more likely than retail outlets (M=.20) and services (M=.06) to contain monitoring devices (F=3.25, p [is less than] 0.05, ([[Omega].sup.2] =0.06). Manufactured goods (M=.72) also were more likely than retail outlets (M=.67) and services (M=.46) to contain response devices (F=3.1, p [is less than] 0.05, ([[Omega].sup.2] =0.06). Manufactured goods also exhibited better integration of hyperlinks (M=3..35) than retail outlets (M=2.73) and services (M=2.86).

Discussion

This study's five-dimensional view of interactivity recognizes theiversity of personality and communication needs of Web users. For self-indulgers and Web surfers, the playfulness and choice dimensions of interactivity fulfill self-communication and entertainment needs. For task-oriented users, the connectedness dimension can fulfill information needs. For expressive users, the information collection and reciprocal communication dimensions of interactivity allow them to initiate communication with the web site representatives or people of common interest online. In this study, the generally low use of interactive devices by Web designers reveals a discrepancy between the interactive capability of the medium and the actual implementation of interactivity in business sites.

The most prevalent dimension of interactivity in business web sites was reciprocal communication (61.2%), a source-oriented interactivity dimension. The next most frequently found dimension of interactivity is availability of choices (52.7%), an audience-oriented interactivity dimension. The prevalence of the reciprocal communication dimension can be attributed to the common presence of e-mail addresses in web sites. Choices in business web sites is a false sense of empowerment, because consumer choice is still defined by the company. The company decides what choices will be provided to the consumers who make use of the options on the company home page. The choices given to visitors were mostly of product-based choice or visual preference choice, rather than choices that could appeal to a more diverse audience (such as language choice or browser choices).

In examining the hypertext links used by the business web sites, the study noted that many sites were not well integrated in terms of their hyperlinks. Third-party information, either about the product or about other relevant information, was rarely found. Hence, the connectedness dimension of interactivity was an illusion in business web sites. Most companies only wanted to confine their visitors to the connected world of the company and not to other third-party sources. Web sites shielded consumers from information from other sources by controlling access to competing and conflicting messages.

If two-way communication between consumers and a company is treated as the highest level of interactivity, then this study shows that manufactured goods and retail outlets want to interact with consumers more than service providers. The reason that goods and retail outlets exhibit similar levels of response and monitoring mechanisms may be the difference in consumer decision-making processes between manufactured goods and services. Consumers can easily make a purchase decision about goods from a web page because goods are tangible. Verifiable information can be obtained to help reduce perceived risk. Consumers can immediately inquire about the goods being offered and use the response mechanisms provided in the site. Services may be much more difficult for the consumer to evaluate based on web page information alone. The site serves more as a showcase of the accomplishments of the ervice company rather than as a business transaction device. Even accepting this rationale, services have failed to capitalize on the power of the Web to generate sales and inquiries by using few response and monitoring mechanisms.

The low levels of explicit monitoring device usage across all business types is another finding of this study. This result may be explained by the technical limitations of developers or the unwillingness of companies to hire a traffic audit service. Unobtrusive monitoring devices, however, can obtain only on physical responses of the visitor. Important demographic and psychographic information is not available without reciprocal communication devices. Another possible explanation for the low usage of explicit monitoring devices could be a concern about the privacy of Web users. In fact, this concern has caused many visitors to enter false information during site registration (GVU's 7th WWW User Survey, 1997). Perhaps companies do not want scare visitors away with monitoring devices such as visitor registration. To enhance reciprocal communication and information collection, a company must gain the trust of their visitors. Companies could gain that trust by giving assurances to consumers that collected data will not be misused.

Theoretically, these five dimensions of interactivity can be applied to traditional media. For example, TV can provide a dimension of playfulness for its audience, as with the CNN news trivia segments. Nevertheless, traditional media have physical and technical constraints that make it difficult to achieve five-dimensional interactivity with the same efficiency as the World Wide Web. For example, the production cost, physical space requirement, and weight of printed paper will make the connectedness dimension difficult to achieve in traditional print media. It requires effort from the reader to cross-reference materials even when they are well indexed and contained in a single volume, because the reader must flip through pages. The bandwidth limitation, production cost, and technical skills required to produce and store TV programs and signals make it nearly impossible to incorporate real-time reciprocal communication and information collection. In many cases, when traditional media need to interact with the audience, they use cross-media support such as telephone and mail. On the Web, the seemingly difficult interactive tasks in traditional media disappear with the ease of retrieving, copying, and distributing messages through cyberspace. Every interactive feature is at the fingertips of the computer user through the click of a mouse.

Suggestions for Future Research

The transient nature of the Web argues for employing a longitudinal perspective to study the evolution of web sites. The current study is only a snapshot of the interactivity of the Web during its early stages. To understand why some dimensions of interactivity were underutilized inweb sites, site executives could be interviewed. They may be able to explain whether the current practice was due to a lack of know-how or a deliberate corporate strategy intended to maintain low interactivity levels between the company and the consumers.

Future research should explore the site visit experience of the consumes. Are consumers aware of the interactive features of the Web, and do they like these features? Do they appreciate source-oriented dimensions? How do they determine whether or not they will use the interactive devices during their visit? What are the effects of the interactive devices on the visitor's product knowledge, consumption behavior, and purchase behavior? Cultural differences between company and consumer may account Cultural differences between company and consumer may account for different content and interactive devices on the Web. Future studies can compare web sites originating in different countries to account for the role of cultural differences in interactive Web communication.

Table 1
Dimensions of Interactivity of Commercial Web Sites

Interactivity Dimension                    Percentage of Sites

Choice                                           52.7
Playfulness                                      19.1
Connectedness:
  Integrated Hyperlinks (score 4 or 5)           37.0
Information Collection:
  Presence of monitoring mechanism               19.1
Reciprocal Communication:
  Presence of response mechanism                 61.8

n=110

Notes

(1.) Perreault and Leigh's (1989) Reliability Index (Ir)

Ir = [{[(Fo/N) - (1 - k)][k/(k - 1)]}.sup.5]

where: Fo = Number of judgements on which the judges agree

N = Total number of judgements made by each judge

k = Number of categories

References

Activmedia (1997, June 12). The 1997 real numbers behind `Net profits [online]. Available: http://www.activmedia.com/ Netprofits.html#1.1.3 Web Avenue.

Balthazard, P. & Koh, C.E. (1997, February 18). Electronic commerce on the World Wide Web: Part II: A study of the use of the web as a business tool [online]. Available: http://www/uncg.edu/~cekoh/report2.htm.

Berthon, P., Pitt, L.F., & Watson, R.T. (1996). The world wide web as an advertising medium. Journal of Advertising Research, January/February, 43-54.

Blattberg, R.C. & Deighton, J. (1991). Interactive marketing: Exploring the age of addressability. Sloan Management Review, Fall, 5-14.

Bryant, J. & Love, C. (1996). Entertainment as the driver of new information technology. In Dholakia, R. R., Mundorf, N. & Dholakia, N. (Eds.), New infotainment technologies in the home: Demand-side perspective (pp. 91-114). Lawrence Erlbaum Associates, Mahwah: New Jersey.

CASIE (1997, June 12). Casie Guiding Principles of Interactive Media Audience Measurement [online]. Available: http://www.commercepark. com/AAAA/bc/casie/guide.html#bibliography.

Csikszentmihalyi, M. (1975). Beyond Boredom and Anxiety. San Francisco: Jossey-Bass.

Digital Output (1997). Interactive? Not by a long shot. Integrated Marketing and Promotion, October/November, 50-51. Dreze, X. & Zufreyden, F (1997). Testing web site design and promotion content. Journal of Advertising Research, March/April, 77-91.

December, J. (1996, June 15). Unit of analysis for Internet communication. Journal of Computer Mediated Communication [online], 1(4), 20 pages. Available: http://www.cwis.usc. edu/dept/annenberg/vol1/issue4/vol1no4.html.

Eighmey, J. (1996, June 18). On the Web: It's what you say and how you say it [online]. Available: http//eighmey.jlmc.iastate.edu/WIP2.html.

FIND/SVP (1996, June 15). Home use of the Internet and commercial online services [online]. Available: http://etrg.findsvp. com/financial/homeuse.html.

Gilder, G. (1990). Life After Television. Knoxville, TN: Whittle Direct Books

GVU's 7th WWW User Study [online] (1997, June 15). Available: http://www.cc.gatech.edu/gvu/user.surveys/survey-1997-04/.

Ha, L. (1995). Subscriber's behaviors in electronic discussion groups: A comparison between academics and practitioners. Proceedings of the First Annual Conference on Telecommunications and Information Markets, 27-36.

Haeckel, S.H. (1998). About the nature and future of interactive marketing. Journal of Interactive Marketing, 12(1), 63-71.

Hahn, H.H. (1996). The Internet Complete Reference, 2nd edition. Berkeley, CA: Osborne McGraw-Hill.

Hoffman, D.L., & Novak, T.P. (1996). Marketing in hypermedia computer-mediated environments: Conceptual foundations. Journal of Marketing, 60 (July), 50-68.

Hoffman, D.L., & Novak, T.P. (1995, June 12). Business scenarios for the Web: Opportunities and challenges. Journal of Computer-Mediated Communication [online], 1 (3), 19 pages. Available: http://www.usc. edu/dept/annenberg/vol1/issue3/vol1 no3.html.

Holbrook, M.B., Chestnut, R.W., Olivia, T.A. & Greenleaf, E.A. (1984). Play as a consumption experience: The role of emotions, performance and personality in the enjoyment. Journal of Consumer Research, 11, 728-39.

Internet Marketing and Technology Report (1996). Is a web page advertising or is it content? And why should we care? 2(3), 5-6.

Internet World (1997, March 10). First year of PC Meter Web measurement reveals: If you promote it (and offer value), they'll come [online]. Available: http://www.npd.com:80/pcmpr23.htm.

Kosters, M. (1997, July 1). InterNIC update [online]. Available: http://rs.internic.net/presentations.markk.97

James, E.L. & Vanden Bergh, B.G. (1990). An information content comparison of magazine ads across a response continuum from direct response to institutional advertising. Journal of Advertising, 19(2), 23-29.

Krol, E. (Adapted by Klopsenstein, B. 1996). The Whole Internet, Academic Edition. Sebastopol, CA: O'Reilly & Associates Inc.

Laurel, B. (1991). Computer as theater. Reading, MA: Addison-Wesley Publishing.

Leibrock, L.R. (1997). Privacy, surveillance, and cookies. In Peterson, R.A. (Ed.), Electronic Marketing and the Consumer (pp. 155-161). Thousand Oaks, CA: Sage.

Maddox, L.M. & Mehta, D. (1997). The role and effect of web addresses in advertising. Journal of Advertising Research, March/April, 47-59.

Magill, K. (1996, September 29). TV-based online service ad targeting. DM News, p. 19.

Morris, M. & Ogan, C. (1996, June 15). The Internet as mass medium. Journal of Computer Mediated Communication [online] 1 (4), 11 pages. Available: http://www.cwis.usc.edu/dept/annenberg/vol1/issue4/vol1 no4.html.

Negroponte, N. (1996). Being Digital. NY: Vintage Books.

Nelson, T.H. (1990). The right way to think about software design. In Laurel, B. (Ed.), The art of human-computer interface design. Reading, MA: Addison-Wesley.

Pavlik, J.V. (1996). New media technology: Cultural and commercial perspectives. Boston, MA: Allyn & Bacon.

Perreault, W.D., Jr. & Leigh, L.E. (1989). Reliability of nominal data based on qualitative judgements. Journal of Marketing Research, 26(May), 135-48.

Rafaeli, S. & Sudweeks, F. (1997, June 15). Networked interactivity. Journal of Computer Mediated Communication [online], 2 (4), 17 pages. Available: http://www.207.201.161.120/jcmc/vol2/issue4/rafaeli.sudweeks.html.

Relevant Knowledge (1998, March 15). Males continue to dominate the Web [online]. Available: http://www.relevantknowledge.com/Press/articles.html.

Robb, J.M., McCarthy, J.C. & Sheridan III, H.D. (1997, February 1). Intelligent interactivity. The Forrester Report [online], 1 (12), 18 pages. Available: http://204.179.229/fish/reports/fish9702it.asp?Uname=LHA.

Rogers, E.M. (1995). Discussion of Innovations, 4th ed. NY: Free Press.

Rogers, E.M. (1983). Discussion of Innovations, 3rd ed. NY: Free Press.

Snyder, I. (1996) Hypertext: the electronic labyrinth. Melbourne University Press.

Stephenson, W. (1967). The play theory of mass communication. Chicago: University of Chicago Press.

Sterne, J. (1995). World Wide Web marketing: Integrating the Internet into your marketing strategy. NY: John Wiley and Sons.

Steuer, J. (1992). Defining virtual reality: Dimensions determining telepresence. Journal of Communication, 42 (4), 73-93.

Wacksman, B. & Cohen, A. (1997, July 21). How to make your web site truly interactive? DM News, p.22.

Walther, J.B. & Burgoon, J.K. (1992). Relational communication in computer-mediated interaction. Human Communication Research, 19(1), 50-88.

Web Techniques (1997, June 17). Recent web statistics. Cyberatlas [online], 2(5), 1 page. Available: http://www.cyberatlas.com/

Wilson, R. (1981). The courage to be leisure. Social Forces, 60, 282-303.

Carol S. Lomicky (Ph.D., University of Nebraska, Lincoln, 1996) is an Assistant Professor of Journalism at the University of Nebraska, Kearney. Her research interests include First Amendment and communications law and policy issues.

Charles B. Salestrom is Director of Public Information at Mid-Plains Community College, North Platte, NE. He received a master's degree in journalism from the University of Nebraska, Lincoln in 1995. His research interests center on broadcasting law and regulation.

The authors would like to thank Kurt Siedschlaw, associate professor of Criminal Justice, University of Nebraska, Kearney, for his help in preparing this article.