Published in the Special Issue of "The Information Society" on Electronic Commerce, Vol. 13. No. 1., 1996.

The Emergence of Internet Commerce

The worldwide network of computer networks known as the Internet has grown explosively, from only 1,000 networks in 1990 to 488,000 networks in July of 1996 (Lotter, 1996). In 1995, International Data Corporation estimated that there were some 38 million users in 150 countries who are communicating, researching, being entertained via, and shopping on the Internet, a figure larger than the population of Canada. Global growth continues to surge at a phenomenal rate. For 1994, the growth of participating networks ranged from 19 percent in Asia to 40 percent in Eastern Europe (Lotter, 1995). Furthermore, this stellar growth continues to add value to the Internet since each new participating individual, organization, and firm adds incremental benefits to the network (Bakos, 1991; Metcalf, 1995). This provides a self-perpetuating incentive drawing more users and providers into the network. At this rate, Internet-facilitated commerce represents one of the most exciting and potentially significant emerging market opportunities available today.

The flexible, open, and standard infrastructure provided by the Internet is capable of transmitting a wide variety of digital data including live audio and video (Press, 1995:2). Much of the explosive growth can be attributed to the multimedia, hyper-text environment known as the World Wide Web. The Web currently accounts for the greatest increase in Internet network traffic and through 1995 was reported as doubling in size every two months (Semich, 1995). Commercial sites (identified by the .com extension) make up the largest percentage of Internet hosts, accounting for over one quarter of all hosts on-line (Lotter, 1995). Despite the wide press coverage, as Hoffman and Novak (1995:1) pointed out "virtually nothing is known about how to develop commercial Web sites to maximize profit impact." This is reflected in the initial foray of many firms into on-line marketing. Over half of the Fortune 500 companies have some sort of Web presence. However, a survey by CIO Magazine in March of 1995 found that over half of those firms characterized their current efforts as 'just playing or dabbling'. This suggests that few firms that are on-line are sure about what they need to do to become successful.

Forrester Research found that $240 million in sales were conducted through interactive media in 1994. Roughly three-quarters of this total was conducted through the on-line services, with the Internet accounting for most of the remainder. While this figure is still significantly smaller than retail and catalog sales figures, Internet sales have grown to this level from virtually $0 in sales in 1993. Furthermore, on-line sales are expected to surpass $5 billion by the end of this century (Chatsky, 1995). Firms which have established a presence on the Internet are engaged in a broad range of activities including advertising, product demonstrations, market research, sales, and support. According to Hoffman and Novak (1995:2) 'the Web has the potential to radically change the way firms do business with their customers by blending together publishing, real-time communication broadcast, and narrowcast.' The medium's ability to allow consumers to easily comparison shop, engage remote applications, and provide vendors with feedback (explicit and via analyzing usage patterns) will impact a variety of factors across a number of industries. These effects may lead to a realignment of profits among participants in the value chain (Benjamin and Wigand, 1995), a reduction in market imperfections (Press, 1995:1), a narrowing of profit margins and price differentials among commodity competitors (Bakos, 1991), an increase in product differentiation and mass-customization, and increased emphasis on individualized service (Kambil, 1995).

One commercial group which stands to gain from the explosion in Internet use is the small, emerging enterprise. A survey covering US, Japan, and Western Europe (IDC, 1995) revealed that such firms make up nearly 86 percent of all business establishments. In one of the few surveys of Internet-facilitated electronic commerce, Kennedy and Dietsch (1995) revealed that 'company size had little or no bearing on success on the Internet,' implying that success on-line does not necessarily result from benefits such as scale economies and a large staff. The cost of gaining a presence on the World Wide Web is remarkably low and continues to fall and as such, is well within the reach of many smaller firms. Once on-line, a firm across the globe may appear to an Internet user to be as close as one located across the street. Advances in low-cost global shipping make it possible for such firms to deliver goods to a worldwide customer base at a minimal cost. As such, market barriers which may have prevented smaller firms from gaining a global reach are now dissolving.

This paper will review literature and press accounts related to Internet-based electronic commerce and present the advantages as they exist for emerging enterprises. The results of a survey of a cross-section of small to medium-size businesses currently engaged in Web-based electronic commerce will be presented. The survey examines factors associated with the adoption decision as well as perceived barriers to success. Results from such a study will be of interest to individuals involved in all aspects of the Internet value-chain including Internet service providers, consultants, IS executives, and educators. Finally, suggestions for future research will be offered.

Drivers of Electronic Commerce

Electronic commerce is driven on both buyer and supplier sides by a number of factors. A survey of relevant literature and press accounts reveals a variety of benefits which may be achieved by firms utilizing the Internet as a commercial tool.

Access to an Affluent Customer Base

A variety of studies have explored the demographics of the average Internet user (GVU, 1995; Gupta, 1995; Yankelovich, 1995; Global Concepts, 1995). While methodologies for these studies vary, all agreed that the typical Internet consumer is better educated and more affluent than the average US resident. All of these studies placed the average annual household income of the typical Internet user within the range of $50,000-$80,000. This compares extremely favorably with the 1994 US Census reported median US household income of $32,264. A more affluent customer may translate into increased sales of higher ticket items. For example, direct marketer Dial-A-Mattress claims its average bedding and mattress sales to on-line customers are $557, 27 percent higher than the firm's average order figure of $450 (Executive Update, 1995).

Lower Information Dissemination Costs

Using the Internet, a firm can place brochure, catalog, corporate information, and support material on-line for a fraction of the cost of print counterparts. Kambil (1995) estimated the marginal cost of disseminating a 30 page document as less than a penny. Benefits to dissemination can be realized internally as well as externally via so called 'Intranets', internal networks which run Internet protocols and software, but which exist behind barriers to outside intrusion (firewalls). Netscape, the leading commercial Web server provider, claims that over half of its sales are to firms setting up internal servers (Sprout, 1995). Alburty (1995) reported that a large investment bank saved over $1 million by using the Internet to distribute documents (although in this case to a primarily internal audience). For smaller firms seeking internal document distribution, the Web can often provide an order of magnitude of savings over packages such as IBM's Lotus Notes (Timmins, 1995).

Lower Transaction Costs

A well-designed Web site can remove expensive human facilitation from the purchasing process. The customer ordering on-line 'drives' the purchase process by selecting products and filling out order information. Back end activities such as credit card authorization and inventory management can all be automated. Such a high level of automation allows 1-800-FLOWERS to charge on-line customers a surcharge of 45 percent less than the fee levied on customers placing phone orders (Executive Update, 1995). On-line orders can also eliminate some of the communication and recording errors often associated with phone orders. PC Gifts and Flowers pegs their error rate of on-line transactions at .02 percent, dramatically less than the floral industry average of roughly 5 percent (Tobin, 1995). IBM (1995) reports savings from Net-based catalogs of 25 percent in processing costs with an over 60 percent reduction in cycle-time. Other reports put the savings garnered from processing on-line orders compared with traditional mail order at greater than two thirds (Verity and Hof , 1994).

Broader Market Reach

An Internet presence may provide access to a rapidly expanding base of potential customers. Yankelovich (1995) estimated that while only 6 percent of American adults were on-line during the first quarter of 1994, that figure had grown to 14 percent by June 1995. Sims (1995) reported that 32 percent of Web users have already bought products over the Internet and 91 percent plan to do so in the near future. Given that international expansion rates are nearly all above 20 percent (Lotter, 1995), many firms are appealing to markets beyond national borders. Business to business reach is also broadened by the Internet. In a pilot of EDI (Electronic Data Interchange) procurement over the Internet, Lawrence Livermore National Lab and Wright Patterson Air Force Base found that because they had a broader reach with the network, they had a broader participation in the bid process and overall costs dropped 10 percent, a significant cost reduction for procurement (Medrich, 1995).

Increased Service

Search engines and electronic distribution of information can greatly reduce the flow time associated with traditional human-facilitated inquiries (e.g. Shneiderman, 1992). Such service provides an incentive for consumers to engage in pre-purchase, purchase, and product support activity through on-line channels versus traditional venues. For example, Internet-based music vendor CDNow! maintains a searchable database of artists and songs on its Web site. This service allows a customer to reference the source for obscure recordings with greater accuracy and less time than a traditional inquiry would take at a local record store. A much larger firm, GE Plastics, distributed some 1,500 technical documents, receiving 12,000 access its first month of operation (Anthes, 1994). This decreased the cycle time for answering information requests to a matter of seconds for those items available on-line, and it had the residual benefit of saving the firm storage, fax, mailing, and personnel costs associated with shepherding literature to inquiring customers. Federal Express and UPS both have Web pages which link customers to their package tracking databases, providing similar benefits by allowing customers to make frequent, direct inquiries without human intervention.

Additional Channels for Customer Feedback

Hoffman and Novak (1995:1) suggested that the interactive nature of the Web facilitates customer support and market activities to a greater degree than traditional media. Electronic mail links to Web pages can funnel customer input, suggestions, and inquiry directly to the appropriate staff. The absence of face-to-face contact with customers may produce a more frank and enlightened exchange, allowing the customer to sound off with suggestions more than one would attempt in person. Most commercial sites also record activity such as the number of accesses or 'hits' which particular portions of a site receive. Analyzing usage patterns can provide insights into general site activity as well as the effectiveness of copy, graphics, and other aspects of a site's user interface (Frentzen, 1995).

Consumer and Market Research

Firms are using Internet-based resources to access information previously unavailable or unknown to them. Government agencies such as the US Department of Commerce and Small Business Administration recognize the advantages of promoting trade information on the Internet and have launched a number of services related to domestic and global trade. Information on-line includes the ability to search a database of trade leads, review briefings on legal issues, examine documents related to starting and developing a business, and an option to receive cables from commercial offices abroad via e-mail. Thousands of small businesses are using the Net to conduct focus groups, design new products, gather demographic data and mailing lists, and test product ideas (e.g. Cronin, 1995; McWilliams, 1995).

Problems and Barriers

Difficulty Monitoring Site Use

As mentioned earlier, the clicks a user generates on a Web site can be logged to a file by the server. This log offers insight on the activity and usage of portions of a site. However, such statistics lack the ability to track a specific user and offer little information on demographics or other aspects of use (Chao, 1995; Cutler and Hall, 1995). Hoffman and Novak (1995:2) suggested that requiring user authorization allows a site to collect and monitor site usage and even allows a firm to tailor a presentation to a specific user. However, many early efforts to require user registration have failed. Two highly popular sites, Hotwired and Time Warner's Pathfinder have abandoned their registration requirement in hopes of further increasing usage and revisit and in turn allow them to command higher ad revenue. Firms such as Nielsen and net.Genesis have developed a variety of approaches for trying to circumvent problems in gathering logs. While non-intrusive methods may eventually evolve, users remain wary of any direct attempt to monitor their Web use. For example, at one point Netscape had offered a pre-release version of the Navigator browser which allowed a Web site to identify and record each user accessing a server, however this feature was removed from subsequent releases due to heightened concerns over privacy from the user community. A similar problem with Netscape's Java script language prompted the firm to offer an option to disable this feature. The ability to introduce executable programs or applets into a Web page and the 'cookies' standard for improving Web client/server interaction may offer an alternative which allows a site to monitor the use of a page without compromising a user's privacy.

Security Concerns

In a recent survey, 56 percent of users polled agreed that security concerns are the main reason they do not buy on-line (Sims, 1995). Resnick (1995), Lyons (1995), and Kennedy and Dietsch (1995) offered examples of firms which experienced significantly less transactions via Web-based venues than from more secure presences on private networks such as Compuserve and America On-line. Greater security concerns over Net-based commerce versus the perception of greater security on private networks is suggested as a partial explanation for such discrepancies. Credit card and software firms have banded together in alliances to develop schemes for secure transactions. In principle, even the less powerful, export-approved, international version of encryption schemes require a prohibitively massive amount of computing power to crack (Stewart, 1995). However, early flaws in the implementation of encryption schemes in the Netscape browser have brought into question otherwise solid protection, requiring Netscape to immediately release a fixed version of their Web browser. Further mistakes may slow the user adoption process by offering users reason to question the integrity of Internet-facilitated transactions. However, a study by Forrester Research showed that concerns over Net security may be exaggerated (Noglows, 1995). Forrester estimated that the Internet commercial fraud rate is roughly $1 per $1,000 transacted. This compares with MasterCard's overall credit card fraud rate of $1.41 per $1,000 and fraud rates for cellular phone and toll call fraud at $19.83 and $16 per $1,000 respectively.

Bandwidth

Hoffman and Novak (1995:1) argued that a flexible and engaging hypermedia encourages exploration of on-line offerings by consumers while McCormick et al. (1987) claimed that visualization through graphics enriches the discovery process. Benbasat and Schroeder (1977) also found advantages to graphic displays. Furthermore, Ives (1982) provided evidence showing that graphics are an effective presentation medium where the objective is to either inform or persuade. Multimedia such as graphics and sound can be used to produce engaging, attractive and popular sites (e.g. Cronin, 1995), however the lower bandwidth available to the majority of shoppers requires firms to make choices which limit the experiential nature of their on-line content. A 1995 survey conducted by Yahoo, one of the Internet's most popular destinations, suggested that over 50 percent of Web browsers accessed the site at 14.4 Kbps or less, transmission speeds (bandwidth) which delay the presentation of complex graphics and limits the use of more engaging multimedia such as audio and video. Users accessing at higher speeds are likely doing so from school or the office, locations which are not conducive to shopping. This also implies that bandwidth divides users into groups and that groups accessing a site at higher transmission speeds could be more easily persuaded to make purchases than those accessing from low bandwidth environments (such as home). Some firms offer users an option to explore versions of their sites with less multimedia content to speed-up the on-line experience, however in addition to being less engaging, such an alternative imposes costs on the firm associated with designing, testing, and maintaining two distinct versions of a site.

Expertise and Access

While computer use is increasing, it still remains beyond the reach of the majority of the American population and is even more elusive for international users. Only 17 percent of Americans 16 years or older report having access to the Internet (CommerceNet/Yankelovich, 1995). Surveys suggest that 64 percent of Internet users have a college degree (CommerceNet/Nielsen, 1995) and the average age of Internet users is 35 versus 41 for the overall American adult population (Yankelovich Partners, 1995). Evidence suggests that factors such as age and education are related to the adoption of computer technology. For example, Baird et al. (1988) demonstrated that younger users are more likely to use new technology associated with a traditional task (in this case an information kiosk). As such, many firms seeking to appeal to certain demographic groups may fail at on-line sales efforts.

Decreasing Profit Margins

Many researchers have forecasted long term impact of a shift to widely available, consumer-driven, electronic commerce. Kambil (1995) suggested that creating and maintaining high profits in an environment of decreasing transaction costs will be one of the greatest challenges facing firms competing over the Internet. Benjamin and Wigand (1995), Kambil (1995), and Hoffman and Novak (1995:2) have predicted that retailers selling standardized goods will suffer as transaction costs associated with acquisition drop and the balance of power shifts toward the consumer who, using advanced technologies like intelligent agents and search engines, can search for products and take advantage of the lowest price. However, anecdotal evidence suggests that Net shoppers are more interested in gathering information on-line than in obtaining the lowest price. This may suggest that Internet activity may favor those products which do not face competition, but which have the purchase decision impeded by product complexity, lack of knowledge, and other factors which can affect buyer uncertainty.

Difficulty in Obtaining Reliable Searches

Bakos (1991) argued that an increase in market participants and a reduction in search costs can benefit the consumer. While the Internet may at first seem to reduce information search costs, the phenomenal growth and turnover of content on the Web make reliable searching difficult at first pass. The more information which exists on-line, the more difficult it is becoming to get a reliable search. As such, the Internet is subject to consumption externalities. A solution to the search problem may be the introduction of header information as suggested by Negroponte (1995), however such a scheme would likely be complicated by reliably and regularly classifying vastly divergent and fluid content. As Bakos (1991) indicates, the best strategy for a buyer in such a market is to determine a 'price and inconvenience' threshold and keep searching until a satisfactory product is located or the threshold is reached. Consequently, customers with lower search costs become more demanding and are willing to make fewer compromises concerning their ideal product.

Exploratory Study of Small Businesses

Small and medium-size businesses were selected from the October 1995 list of on-line stores in the popular free listing service known as the iMall (http://www.mecklerweb.com/). It should be noted that small to medium-size enterprises are deemed to possess specific attributes that distinguish them from the large enterprises most often studied in regard to information systems for competitive advantage. On one hand, these enterprises are more vulnerable because of their lack of financial and human resources (Noori, 1987) as well as information resources that are needed to sufficiently understand and master the organization and its environment (Van Kirk and Noonan, 1982). On the other hand, small and medium-size organizations have specific advantages compared to large organizations because of their greater flexibility (Crawford and Ibrahim, 1985), their quickness in implementing decisions, and their capacity for adaptation and short-term orientation (Julian and Lafrance, 1983).

Our intention was to limit our search to firms dealing primarily in goods rather than services. Firms with a working Web site and e-mail address were taken as potential recipients of a mailed questionnaire. The mailed questionnaire was chosen over electronic forms because it is quicker/easier to fill out and because it allowed us to send the instrument to the proprietor who may not be the administrator of the site. All firms in the following iMall classifications were considered: food, beverages, crafts and hobbies, collectibles, gifts, kids' stuff, toys-games-novelties, publications, book publishers, on-line music shops, auto, boat, and bicycle. A total of 487 firms were identified. Of those, only 349 had a viable e-mail address and were subsequently contacted to solicit their participation in our study. Almost 50 percent, 164 firms, agreed to participate and 141 completed a questionnaire.

The questionnaire was divided into four parts. The first part consisted of 19 items dealing with the characteristics of the firm's on-line presence such as the types of on-line services offered, the frequency and nature of site updates, the number and cost of products purchased on-line, and the percentage of total sales originating from the Net. The second part asked respondents to rate the importance of 18 expected benefits of an on-line sales presence on a 5-point Likert scale. These expected benefits were identified from the literature and from informal conversations with individuals involved in electronic commerce. The third part dealt with some of the barriers that may prevent firms from realizing additional benefits from their on-line presence. Once again, respondents were asked to rate each of 9 potential barriers on a 5-point Likert scale. The fourth and final part inquired about some of the characteristics of the respondents and their organizations such as type of industry, total annual sales, and the number of employees.

Characteristics of the On-Line Sales Presence

General Characteristics

In an attempt to better understand what types of firms are participating in Web-facilitated commerce, general characteristics of the firms surveyed were collected along with those factors influencing and hindering adoption. Table 1 provides the reader with a snapshot of mean responses on firm, service, site management, and purchase-related characteristics. Annual sales of $500,000 was used as a dividing point to classify firms into two groups containing a similar number of organizations. The two groups, which will be referred to as larger firms and smaller firms, included 67 firms and 66 firms respectively (8 firms did not supply the information required for classification). These two groups were then compared using a 2-tailed t-test to examine the effect of firm size on the individual variables. Table 1 reports the variables for which significant differences were found. The first column identifies the variable under examination. The second and third columns list the means for larger firms and smaller firms respectively. Finally, the fourth column gives the probabilities for the 2-tailed t-tests. An alpha level of .1 was used to identify significant differences. Though this level is higher than the commonly used .05, it is nonetheless easily justified due to the exploratory nature of the study and the small sample size. Moreover, the p values for the significant variables are, for the most part, smaller than .05.

Of the 141 firms that responded to the questionnaire, approximately 91 percent were based in the United States. The remainder of the participants included organizations based in: Canada, England, Italy, Holland, South Africa, and Taiwan. Unfortunately, the number of international respondents was not large enough to perform cross-national comparisons. However, this strong bias toward US-based firms was anticipated, since the majority of Internet hosts are located in the US (Quelsh and Klein, 1996). Total annual sales ranged from $1,000 to $40,000,000 with average sales of approximately $1.2 million. The number of employees varied from 1 to 300 with an average of about 10 employees per firm. More than half of the firms, 56 percent, considered themselves direct mail businesses while only 32 percent operated physical facilities (i.e., a store front). The majority of products sold were consumer tangible goods, either durable or non-durable, with 56 percent of the respondents selling the former and 44 percent selling the latter. Most on-line sites, 88 percent, had been in operation for 1 year or less and had been instigated internally (85 percent). Therefore, it appears that on-line sales for our sample of small to medium-size businesses is a relatively new phenomenon and one that originated within the organization.

Services Offered and Types of Ordering Methods

Respondents were asked to identify the types of on-line services that their organizations offered from a list of seven such services, which included: brochures/catalogs, newsletter, on-line ordering, e-mail support, contests/games, links to other sites, and encrypted (secured) credit card orders. The results indicate that a wide variety of on-line services were offered by our sample of businesses. Over two thirds of the firms offered brochures/catalogs, on-line ordering, and e-mail customer support. Furthermore, more than half, 55 percent, included links to other Internet sites from their Web page. Adding such links is likely aimed at making the firm's site a central resource for Internet information of likely interest to customers. However, less than 12 percent had encrypted credit card capabilities despite consumer concerns with security. Though there were no differences between larger firms and smaller firms for individual services, the larger firms offered a significantly greater total number of services than the smaller firms (see Table 1). This is to be expected, since larger organizations have more resources available to offer a wider range of services.

Most firms offered a number of ordering methods. In fact, each of the five ordering methods included in our survey: on-line form, e-mail, fax, phone, and postal mail were utilized by at least 65 percent of the respondents. The two most popular methods were phone (83 percent) and postal mail (78 percent). The least popular method was on-line form (65 percent), which indicates that organizations and consumers may still prefer the more traditional methods of ordering. Once again, the larger firms offered a significantly greater total number of order methods than the smaller firms.

Responsibility for the Development of the Site

As mentioned previously, the idea and motivation to establish an on-line sales presence usually originated from within the organization. The next set of variables asked respondents to identify who was responsible for certain aspects of the development of their sites. Four categories of responsibilities were included: text, graphics, HTML code, and server administration. The level of internal responsibility decreased from category to category. That is, 79 percent of the firms were responsible for text development, 63 percent for graphics development, 50 percent for HTML code development, and only 26 percent for server administration. This decline in the level of internal involvement can be easily explained by the increasing level of complexity of each category, which would require the firm's personnel to have highly specialized skills. Furthermore, the larger firms had significantly greater levels of outside involvement (i.e., less internal involvement) for text, graphics, and HTML code developments than the smaller firms (see Table 1). Therefore, it appears that the larger organizations in our sample tended to farm-out the development of their on-line sales presence more frequently than the smaller organizations. This difference may be explained by the greater monetary resources available to larger firms to hire consultants and by the greater involvement of employees (usually the owners) in all facets of the organization for very small businesses.

Purchase Characteristics

Most items sold on the Net were relatively low price items. In fact, 65 percent of the participants responded that the average cost of items sold on the Net was less or equal to $50.00. Furthermore, the number of items sold per transaction was normally three or less as indicated by approximately 84 percent of respondents. However, the larger firms differed significantly from the smaller firms on both the cost of items sold and the number of items sold per transaction with the larger organizations having a higher level of both. These two differences may be explained by the significantly greater percentage of larger firms involved in industrial (i.e., business-to-business) transactions, which usually consist of larger orders.

The percentage of total sales and total orders originating from the Net varied greatly among firms. More than half, 54 percent, had less than 5 percent of their total sales coming from the Net while 24 percent had more than 50 percent of their sales from the Net. Furthermore, the smaller firms possessed significantly greater percentages of total sales and total orders coming from the Net than the larger firms. This indicates that the smaller firms conducted smaller transactions than the larger firms, but were more dependent on the Net for their source of revenue. Smaller firms may seek a higher initial relative advantage to the innovation (Rogers, 1993) than larger firms due to their increased likelihood to concentrate on niche markets or diversified products and services. This greater focus on niche markets or diversified products is brought about by the more limited financial resources of smaller businesses versus their larger counterparts (Noori, 1987) or by an attempt by smaller businesses to move rapidly ahead of larger and less nimble competition (Julian and Lafrance, 1983).

Interestingly, larger firms obtained a significantly greater percentage of their Net sales from outside of their home countries. This may suggest that foreign customers prefer to buy from larger organizations. As such, the impact of name recognition of more established (hence larger) firms may be stronger among international buyers. Also, international shipping may provide additional cost and complexity barriers for smaller firms.

Finally, almost 63 percent of firms planned to increase their investment in on-line marketing over the next year while another 31 percent planned to keep it the same. This strongly suggests that, in the near term, on-line marketing should keep on growing and that organizations are satisfied enough with the performance of their on-line sales presence to increase or maintain their investment. In fact, less than 3 percent responded that they would cease their involvement in on-line marketing. The next section will discuss the factors that influenced the adoption of an on-line sales presence.

Factors Influencing the Adoption of an On-Line Sales Presence

The relative importance of 18 expected benefits of an on-line sales presence was measured using a 5-point Likert scale with 1 meaning that the expected benefit was not at all important and 5 meaning that the expected benefit was extremely important. Principal component factor analysis with Varimax rotation was used to simplify the interpretation of the results. The analysis yielded a six-factor solution, which explained approximately 64.4 percent of the total variance. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy was 0.688 and the Bartlett test for sphericity was highly significant. Though the KMO is low, it is still acceptable. Variables with loadings less than 0.5 were eliminated from the analysis so that the final analysis was performed on 17 variables (only one variable, to help develop new products, was eliminated). The results are summarized in Table 2. The first column contains a description of the variable. The next five columns report the factor loadings for each factor. A sixth factor (factor 2 in terms of weighted average), to experiment with a new marketing tool, is not included in the table since it contained only one variable. However, the results of this factor will be discussed completely in the following section. The factors are listed in order of importance based on the weighted average of the variables that make up the factor. The weighted average, the percent of variance explained, and the Cronbach's alpha for each factor are included in the column headings. The sixth and seventh columns list the means for larger firms and smaller firms respectively. Finally, the last column reports the probability for the 2-tailed t-test comparing larger firms to smaller firms. The symbol 'ns' is used to indicate that there were no significant differences at an alpha level of .1 between the two groups.

Low Development and Maintenance Cost

The first factor was labeled low development and maintenance costs. This factor had a very high Cronbach's alpha (.95), which indicates a high level of internal reliability. The factor includes only two items and explains over 11 percent of the variance. The factor also possesses the highest weighted average, 4.00, of any of the factors in our analysis. This strongly supports the notion that low development and maintenance costs play a critical role in the decision to go on-line, especially when combined with the high individual ratings of the variables. Therefore, it may be suggested that the low costs of development and maintenance of a Net presence is a low-risk strategy and thus, very appealing to small and medium-size businesses, which may lack the financial resources to use other methods of marketing communication. The positive reaction toward the Web's low cost of entry reinforces Rogers' (1983) claim that triability is a significant factor in the adoption of an innovation. The high weighted average of this factor implies that firms perceive these lower development and maintenance costs as an advantage which allows for experimentation and reduces the risk associated with adopting a new technology.

To Experiment with a New Marketing Tool

Factor 2 (not listed in Table 2) consists of only one item, to experiment with a new marketing tool, and thus is not included in the table. The item has a mean of 3.90, which indicates that it is relatively important in the decision making process. It also reflects the innovative nature of our respondents who are willing to experiment with a new method of marketing their products. The significance of this factor also reinforces the attractiveness of the triability attribute of an innovation as suggested by Rogers (1983). Furthermore, it highlights the fact that a Net presence is a relatively new phenomenon and as such, there is a degree of uncertainty in the likelihood of achieving any anticipated benefits.

A Promotional and Image Building Tool

Factor 3 consists of two items, promotional tool and image building tool, and explains 6.5 percent of the variance. The weighted average indicates that this factor is third in importance. Furthermore, the mean for the promotional tool benefit is one of the highest for our sample at 4.06. This suggests that the promotional benefits of a Net presence are very important in the decision to go on-line. It may also indicate that our respondents perceive the Internet as a new and innovative method of promoting their products. In effect, the Net may be viewed by some of our respondents as a new medium for advertising to supplement the traditional media of television, radio, and print.

Our results also indicate that the image-building benefits of a Net presence are considered more important in the decision to go on-line for larger businesses than for smaller businesses. In fact, this benefit shows the most significant difference between the larger and smaller businesses in our sample. Furthermore, the weighted averages for the two variables in this factor are also significantly different among smaller and larger firms (p=.027). This demonstrates that the larger respondents perceive that an Internet presence reflects positively on the firm's image, possibly creating a halo effect through perceived technology leadership.

Financial Considerations

Factor 4 has been labeled financial considerations. This factor contains four items related to increasing sales or cost reduction. The Cronbach's alpha is relatively low (.58), but nonetheless acceptable considering the small sample size. The weighted average is 3.53, which places the factor fourth in importance. However, the weighted average is misleading in this case, since two of the variables, to increase sales and to expand geographical coverage, are the highest rated variables for our sample with means of 4.48 and 4.18 respectively. This suggests that financial or revenue generating benefits are the most important when making the decision to go on-line. This is certainly not surprising, since business decisions must be financially viable. However, the apparent lack of importance of the cost reduction benefits of a Net presence was surprising. The literature has often reported the importance of cost reduction, especially in the distribution channels, of a Net presence. Therefore, it may be that small to medium-size businesses are not as concerned with these reduction benefits, but focus more on the revenue generating potential of the Net presence. These findings are supportive of statements made earlier that smaller firms are more dependent on what Rogers (1983) would refer to as a high relative advantage to the innovation and that smaller firms, given their more limited financial resources, have a lower threshold for costly experimentation.

The importance of increasing sales also differed between the larger firms and the smaller firms in our sample. The smaller firms placed a significantly greater emphasis on increasing sales than the larger firms. This complements the interpretation of the previous factor, which showed that the larger firms tended to stress the promotional and image-building benefits (i.e., non-financial benefits) of a Net presence more than the smaller firms.

Obtaining and Disseminating Information

Factor 5, which has been named obtaining and disseminating information, includes six items related to information gathering (e.g. acquiring mailing lists and soliciting customer feedback) and to information dissemination (e.g. providing customer support). The factor explains almost 23 percent of the variance and has an acceptable Cronbach's alpha of 0.75. The weighted average of the variables in the factor is 3.07, which is relatively low compared to the other factors. This indicates that this factor is not as important in the decision to go on-line as most of the other factors. However, the individual means of the variables suggest that low cost and rapid way to disseminate information (3.94) is relatively important in the decision making process. This particular benefit of an on-line presence has been repeatedly suggested by the literature and should be especially important for small to medium-size businesses. In fact, it appears that it was important for our sample of businesses.

Only one variable in this factor, to provide customer support, shows any difference between the larger firms and the smaller firms in our sample. The data indicate that larger firms place significantly more importance on providing customer support when they decide to go on-line. Furthermore, the weighted average for the factor is also significantly different for the two groups with a higher rating for the larger firms (p=.015). This may suggest that larger organizations are more willing to utilize their on-line presence to supply their customers with additional services besides simply selling their product. It also indicates that larger organizations place a greater level of importance on the information benefits of a Net presence. However, the difference may also be explained by the nature of the products sold. The larger firms in our sample tended to sell more expensive and more industrial products which may require a greater level of customer support.

Competitive Considerations

Factor 6 included three items and is related to competitive pressure or considerations. The factor appears to be relatively unimportant with a weighted average of only 2.66. Furthermore, the Cronbach's alpha is low, .47, which indicates some weaknesses in the reliability of the factor. However, one of the items, to be one of the first firms in the industry on the Net, obtained a high importance rating of 3.90. This may indicate that the individuals in our sample, mostly the owners of the businesses, regard being technological leaders or innovators as relatively important. The high desirability of being the first on-line implies that firms see an advantage in their early entry into a specific market. This behavior may reflect a belief that first-mover advantage can result in significant market share gains (Urban et al, 1986), and such gains may be particularly strong in electronic markets (Bakos, 1991). Therefore, our sample of firms are not simply trying to match the competition, as indicated by the low score (2.50) on that item, but are attempting to move ahead of the competition by becoming an early leader in a new and expanding medium.

Barriers to Obtaining Additional Benefits from the On-Line Sales Presence

The last section deals with some of the barriers that may prevent our sample of businesses from obtaining additional benefits from their Net presence. The respondents rated 9 items on a 5-point Likert scale with 1 meaning that the item was not a barrier at all and 5 meaning that the item was a very large barrier. Once again, factor analysis was used to simplify the interpretation of the results. The same methodology was utilized as in the previous part. The final analysis included 8 variables and yielded a 4-factor solution which explained 69.6 percent of the variance. The KMO was relatively low at .60, which is at the lower bound of the acceptable range while the Bartlett sphericity test was highly significant. The results of the analysis are summarized in Table 3 and follow the same format as Table 2. Factor 4, products not well suited for the Net, contained only one variable and is not included in the table although its results are discussed later in the section. One variable, security concerns, was eliminated since its factor loading was below 0.5. The lack of significance of security concerns is surprising, suggesting that vendors perceive security concerns to be less significant a barrier in realizing benefits than may have been highlighted in the trade press.

Difficult to Analyze and Promote the Site

Factor 1 consists of two items and explains over 17 percent of the total variance. The weighted average for the factor is by far the highest of any factor (3.02). Furthermore, the apparent difficulty in promoting the site is the only barrier which received a rating greater than 3. The high rating on this item suggests that our respondents find it relatively difficult to attract potential customers to their sites. This may be due to the large and growing number of commercial sites available on the Internet as well as to problems associated with on-line searches. This may also be an area where small to medium-size businesses, with relatively limited financial resources, are at a disadvantage over the large and well-known organizations (e.g. GE or Microsoft). The difficulties involved with analyzing the use of the site are also well documented in the literature and appear to be recognized by our sample of respondents. It was the second highest rated barrier and may become even more problematic as the number of users increases.

Company's Resources

The second factor, problems with product/medium mix, had very low reliability as indicated by its Cronbach's alpha of only .37. Therefore, it will not be discussed, but the information is included in Table 3 for completeness. Factor 3 includes 3 items, which explained 25 percent of the total variance. The weighted average of the three variables is relatively low, 1.93, which suggests that this factor is not considered to be a serious barrier by our respondents. Two of the variables, requires additional personnel and the length of time required to analyze and respond to customer feedback are perceived as significantly greater barriers by the larger firms than by the smaller firms in our sample. However, the overall means for the two items, 1.76 and 1.88 respectively, suggest that neither of the items are considered a major barrier to obtaining additional benefits. Nonetheless, the differences between larger and smaller businesses may be explained by the greater volume of information and inquiries that must be handled by the larger businesses. This larger volume may bring about more demands on the organization and force the firm to acquire additional personnel.

Products Are Not Well Suited for Net

The fourth and last factor included one item, products are not well suited for the Net, and thus is not included in the table. The mean for the variable was only 1.86, which clearly indicates that our respondents feel that their products are appropriate for on-line marketing. However, our sample included only a few types of products, which may bias our results. Therefore, it would be of interest to determine if some types of products are better suited for on-line marketing.

Conclusions

Our study examined some of the issues associated with the adoption of an Internet-based sales presence by small to medium-size businesses. Our empirical research yielded a number of interesting results. First, small to medium-size businesses appear to be relatively sophisticated and offer a wide variety of on-line services and order methods to their customers. Furthermore, the on-line sales presence appears to be internally driven and supplies the firm with a pool of new customers. Most items sold on-line are relatively low cost items, which may indicate that consumers are still reluctant to spend a large amount over the Internet.

Six factors that impacted the decision to go on-line were identified: low-development and maintenance costs, an interest in experimenting with a new marketing tool, the desire to promote products and build the company's image, financial considerations, benefits in obtaining and disseminating information, and competitive considerations. Our analysis indicated that among these factors low development and maintenance costs was revealed as the most significant according to the surveyed firms. The low financial commitment required implies a high level of perceived triability among would-be Internet store fronts, an important consideration since smaller firms tend to have more limited financial resources. Previous studies of electronic commerce have noted the high cost involved in the startup and maintenance of electronic commerce systems (Malone, Yates, and Benjamin, 1987; Bakos, 1991). As such, the World Wide Web as an innovation is significant in that small firms perceive that the financial obstacles to their participation in electronic commerce have been reduced to the point where they can experiment without an overly burdensome commitment.

Among those individual variables examined, increasing sales and expanding geographical coverage were cited by participants as the most important individual variables in the decision to go on-line. Both of these variables suggest a perception among surveyed firms of a high relative advantage of anticipated benefits derived from the on-line sales presence. The emphasis on the financial benefit that would result from the Web presence is also reflective of smaller firms' greater sensitivity to financial considerations. The results of this study indicate that small businesses which have made the decision to go on-line perceive that deploying a Web-based sales presence is a relatively low risk, high reward strategy. Most executed their sites due to an internal motivation as opposed to being prompted by external consultants or existing customers. Larger firms in our study showed a higher level of interest in improving the firm's overall image through use of the Web site, while smaller firms emphasized the importance of increasing sales to a greater degree. Overall, our respondents seemed to be relatively satisfied with their on-line sales presence as indicated by their willingness to increase or maintain their investment in on-line marketing.

Limitations and Areas for Further Research

Our study suffers from a few weaknesses. First, our sample was relatively small and limited to only a few industries or types of products. Therefore, the results may be of only limited generalizability. Second, our sample may suffer from a strong element of self-selection, since it included only firms that were on-line and thus did not include firms that made the decision to not go on-line. Furthermore, almost half of the firms that we contacted either declined to participate or did not respond to our request. This may have produced some non-response bias. However, it should be noted that nearly 50 percent of those firms contacted electronically agreed to participate and over 85 percent of those firms which were mailed questionnaires returned them, a very high response rate for a mailed survey.

Our analysis was also based on perceptual data and not behavioral data. Though using perceptual data is not optimal, Venkatraman and Vasudevan (1987) found a strong degree of convergence between the two measurement methods. Therefore, the perceptual nature of our measurements, though not optimal, should yield valuable information. Finally, the study was overwhelmingly biased toward US-based firms. This may be problematic since Parente and Prescott (1994) found that that the barriers to adoption can differ across national boundaries. Similarly, Phillips, Calantone, and Lee (1994) stated that cultural affinity has a significant influence on Rogers' Technology Acceptance Model (TAM). As such, we suggest that there are opportunities for executing broader, global studies to examine the motivation and adoption barriers of Web technology to see if there are general factors associated with the phenomenon as well as unique characteristics particular to various cultural and socio-economic classifications.

This research can be furthered by exploring perceptions of participants over time. While the Web has lowered the participation barriers to electronic commerce, it remains to be seen if participating firms will be satisfied with benefits received from the medium over time. Our initial study was exploratory in nature. A larger study with more stringent operationalization of variables, perhaps with multiple indicators offering non-perceptual measures would supplement and further the work offered in this paper. A larger study may also allow for the consideration of the impact of industry and product mix on both perceived and realized benefits.

Being a relatively new field of study, Internet commerce offers a wide range of other research opportunities. A broad number of research topics are worth of further consideration. These include identifying the impact that an on-line sales presence has on firms and the impact of Web participation on industry structure and competition. Researchers and practitioners should also be interested in determining those characteristics associated with a more successful Internet presence. Key variables such as timing, firm reputation, industry, services, interface, product mix and price, among others, must be further explored. Such research into site effectiveness must also consider the consumer side of the equation, surveying user attitudes and identifying methods used to select a purchase medium and a firm within an electronic commerce medium.

As stated earlier, one limitation of this study is that it concentrates on those firms which have already made the decision to go on-line. It should also be determined if the decision making process differs (i.e., the criteria involved in the decision making process) for firms that decided not to invest in on-line marketing. Lastly, cross-cultural studies must be executed to determine if firm and consumer behavior in electronic commerce environments varies internationally. 

Please feel free to send your comments to:

Pat Auger or John Gallaugher