A semester-long e-commerce course the authors led in the spring of 1998 at The George Washington University in Washington, D.C., included a Web-based market system that allowed the students to electronically buy and sell goods and services while building an infrastructure approximating real-world technical, marketing, and record-keeping performance.
The students, who were from a variety of academic disciplines (see the sidebar "Course Background"), covered a range of e-commerce topics, including network infrastructure, economics, retailing, marketing and advertising on the Net, security and privacy, e-payment systems, the social effects of e-cash, legal implications, public policy, and online publishing (for a complete syllabus, see www.seas.gwu.edu/classes/ cs701/syllabus.html).
The students created a functioning online market that was the course's most remarkable and instructive feature. The market gave them firsthand experience using and creating e-commerce technologies, discovering how a market's various players really interact, and addressing the inevitable social and policy tensions. The market project was not a controlled economic experiment designed to predict real-world behavior, so we were surprised when its results mirrored the current state of real-world e-commerce as much as they did.
A market involves three phases: pre-purchase determinations, purchase consummation, and post-purchase activities. Before a purchase can be made, a consumer searches for information about a product or service, comparison shops and negotiates the price, and then places an order with a vendor. Vendors decide how to market their products or services and inform consumers about acceptable forms of payment. During the purchase transaction, the consumer contacts a vendor to authorize payment. If the payment is in e-cash or e-tokens, the vendor contacts the issuing bank or payment service to verify that the e-cash or e-tokens received from the consumer are valid. In credit or debit transactions, the vendor contacts the payment service or bank to verify that the payment can be made, then debits or credits the accounts of the various participants. Post-purchase activities involve delivering the goods or providing the service. In the course's market, the students were also required to decide how to build the market's infrastructure, including whether to adapt existing technologies or create their own systems for paying for and buying goods and services online
The students were divided into six groups, or companies, with at least one computer science student assigned to each group. One group established a bank to maintain consumer and vendor accounts and to serve as a certificate authority. Other students established three companies to develop e-payment systems based on credit, debit, and e-tokens; these systems could accept and fulfill requests for payment for consumer purchases and business-to-business transactions. The companies also negotiated with the bank to verify and authorize payments.
The students also established two marketing firms, one using push technology, one using pull technology. The push group provided vendor product information to consumers willing to supply preference profiles. The pull group offered information on products consumers could buy when visiting its Web site, which gave access to a database of vendor information.
All companies (see Figure 1) had to establish standards and prices for their products and services, each starting with $2,000 in class currency as seed money. The income received belonged to that group alone and could be divided up in any way among its members as decided by those members.
Separate from these six groups, each student also functioned as an individual student-vendor, opening an electronic shop (see Figure 2), selling a product or service via the Web to other students. Each student-vendor decided which forms of payment he or she would accept. In some cases, the vendor's Web pages also included advertising for products and services available from other student-vendors. The vendors were paid for their products and services by student-consumers, and some received income from advertising. However, student-vendors were not required to deliver actual products or services.
Each student was also a consumer, starting with $2,000 in class currency for buying products and services online from the vendors, using the various forms of acceptable e-payment.
The professors functioned as the "government," setting rules only when they felt intervention in the market was absolutely necessary; they set almost no rules.
By the fifth week of the semester, all student-vendors had established online businesses. By the eighth week, they had to determine which payment systems they would accept, incorporating them into their ordering and billing systems. By the tenth week, the market was functional and ready for alpha testing, with beta testing beginning in the eleventh week. The remaining four weeks were for buying and selling products and services in the market.
One of the market's first developments was the emergence of special interest groups. Almost immediately, the computer scientists in each of the six companies appointed themselves representatives to a technical-standards committee. In addition, one of the students established a "lobbying firm" to lobby the government (the professors) on behalf of merchants and other companies.
A formal lobbying or legislative process had to be established. Students were encouraged to post their proposed legislation on the class Web forum for review by the government and other class members. The government would then respond to the proposed legislation. Even after this legislative process was announced, a few students continued to lobby the professors and the teaching assistant privately by email. We weren't surprised, considering that some of the actions they proposed were for their own direct benefit, rather than for the benefit of the overall market.
Students were frustrated by the government's laissez-faire approach to its responsibilities, lobbying for more intervention in such areas as infrastructure, technical standards, privacy, security, and unsolicited email, or spam.
In developing payment systems, simple-and-easy-to-implement standards, such as secure email transactions, won over such nominally more technically advanced solutions as Secure Sockets Layer (SSL), a protocol providing a more secure version of the ubiquitous hypertext transport protocol , and Secure Electronic Transactions (SET), a protocol designed for e-transactions supported by many credit-card companies, banks, and other organizations interested in e-commerce . These choices were made in part because of the course's time constraints, along with the companies' nontechnical members, who wanted to keep the systems simple enough for nontechnical people to use and understand.
Although the class's credit company, called "Secure Credit Payment System" (see Figure 3 for its home page), and its debit-payment company, called "QuickSafe," began as two separate competing enterprises, they eventually merged. (Operating de novo, the market had no antitrust law to prevent their doing so.) The new credit/debit conglomerate also considered merging its accounting system with the bank's accounting system but decided to keep its system autonomous because its role in the market might be displaced if it shared its primary assetits accounting systemwith the bank.
The competing e-token payment system, called MON-E, differentiated itself from the credit/debit system by guaranteeing anonymity in purchase transactions. Student-consumers and student-vendors found the MON-E system more difficult to understand and use.
Eventually, the Secure/QuickSafe credit/debit system achieved a nearly 90% market share, while MON-E achieved only slightly more than 10%. This was due in part to the credit/debit company's marketing strategy, which tolerated a small amount of fraud, but focused the combined companies' efforts on being ubiquitous in the market. Most students chose to trade off anonymity in favor of simplicity and ease of usebehavior typical of utility/security compromises.
The student-vendors' Web pages represented a cross section of the evolving world of Internet commerce. Students "sold" both common and niche products, including flowers, food, carpets, and antique maps, as well as a range of services, including insurance (against losses in e-commerce transactions), theater tickets, food, gambling, and adult entertainment (with age-appropriate click-through screens, mirroring real-world concerns and satisfying the university's computer-use policies). (See Figure 4 for the home page of "Blackstone's Pet Embalming.")
Several students planned on translating their classroom-market experience into real-world online business ventures. One of them, who offered degrees from a fictitious university in exchange for donations, used this experience to convince a real university to hire him to develop an online fund-raising project. After the course was described in national online media , another student added "As seen in The New York Times" to his resume and used his Web site to carry his own "situations wanted" ad for a summer job. It worked.
One student-vendor received a real order from a real live online consumer, during alpha testing, for one of her oriental carpets. The possibility of interacting with the outside world prompted other student-vendors to quickly add disclaimers saying their Web pages were only part of a class project.
The student-vendors found that niche businesses were more successful than those selling commonly available items. For example, when the two students selling flowers online realized each had a direct competitor in the market, both tried to distinguish their products, such as by offering overnight delivery service.
The student-vendors grappled with how much consumer data they should collect and retainaware of the competitive advantage they could gain by offering strong consumer privacy protection. But because acquiring and retaining consumer data meant they could trace problems and detect fraud more readily, they had to calculate the trade-off between privacy and the customer service they could offer. Students mentioned that given more time, they would have explored ways of providing better customer service while still protecting consumer privacy. Approximately 90% of the students' Web sites posted privacy policies, most added later in the course, after strong advocacy by several guest speakers. However, posting privacy policies didn't ensure the policies were always adhered to. If the market had continued, enforcement mechanisms may have evolved.
Another important lesson for student-vendors was that it was very difficult to change negative first impressions about their products, services, and Web sites. For example, the MON-E payment system became easier to use over time, but because of negative first impressions, consumers, as well as vendors, were reluctant to give it a second chance. Similarly, student-consumers who had a negative experience dealing with a particular vendor's Web site were less likely to return.
Over time, students also developed innovative ways of marketing and advertising their products and services, crossing media boundariesfrom sending email to distributing flyers and product samples (such as roses and Mexican food) in class. A lucrative market in Web-page banner advertising also developed, as banners competed for space on student-vendors' Web sites. One company, "Agora" (the pull marketing firm), used its weekly five-minute progress report to present in-class infomercials.
Student-vendors also found that push marketing and spamming increased sales. Many of them used the push marketing firm to advertise by sending unsolicited, as well as solicited, email; some of them simply spammed the rest of the class on their own.
Student-consumers demanded privacy, but not anonymity, in their transactions. Most student-vendors and student-consumers opted to use a more convenient payment system than MON-E, even though it was not anonymous. Consumers also reacted negatively to unsolicited email. Some opted out of receiving promotional email by unsubscribing themselves from marketing lists. Others lobbied unsuccessfully for an anti-spam decree from the government.
One rogue student-vendor committed credit fraud, using consumers' credit-card numbers to buy products from other vendors. The bank detected the fraud, but there wasn't enough time before the market was scheduled to close for the credit/debit company (or anyone else) to determine how defrauded student-consumers should be compensated or how to take action against the rogue vendor.
Consumer protections were weak and unregulated. Consumer-protection groups never managed to organize, since most of the semester was devoted to building the market, rather than to participating in it. Consumer groups may have evolved if the market had a longer run. Unknown is whether this protection would have evolved in the form of self-regulation by vendors or in the form of "government regulation."
Lessons for Future Courses
Some aspects of the market were clearly successful while others needed to be done better, or at least differently.
Level of instruction. One challenge in teaching an interdisciplinary topic to an interdisciplinary audience is determining the material's level of difficulty. How do you teach at a level all students can grasp, while still challenging advanced students? We provided an easy-to-understand text as introductory material for the students who needed it . For the students who wanted to delve deeper, we offered recommended readings and allowed them to design projects as challenging as they wanted. We asked our invited speakers to assume that our students were sophisticated and already understood many of the concepts. No student was left unchallenged, and all, including the instructors, came out of the class with much more knowledge than they started with.
Incentive structure. The incentive structure, or grading, for the market project was based on the quality of the students' participation, not on how much profit they made. But many students did not participate for as long or as actively in the market as we would have liked. One possible solution would be to include a measure based on the profit they and their companies earned as part of their grades.
Computer resources. Because of the high technical overhead involved in the online market, we now recommend advance planning and budgeting of computer and system administration resources. We did not anticipate a number of problems, including:
- Unexpected system crashes while students were finalizing their projects.
- No clear procedure or chain of command for interfacing with the university computing center's technical support staff.
- Objections by students to the university policy of not allowing direct access to the common gateway interface (CGI) bin. Being denied access meant the students could not, for security reasons, test their CGI scripts without human intervention from employees in the computing center.
- Limited Web access. When an independent server was finally available for the class, it provided Web access only for students dialing in from home, work, and other domains.
- Computer distractions. Computers in the classroom were both a help and a hindrance. Aside from providing access to the Web for in-class demonstrations and assignments, computers helped make the lectures easier to follow. We posted most of the invited speakers' presentation slides on our Web site, so students could follow their lectures on the PCs in front them. But, at times, computer use was mildly distracting, prompting us to ask the students to turn them off so they could give full attention to the speakers' presentations.
- Not enough time. The appropriate balance has to be established between the time for developing the market and the time for participating in it. Some students suggested the course should be spread over two semesters, the first for building the market, the second for using and enhancing it.
- A closed market. We received requests from various sources, including other members of the university community and reporters, who wanted to test or participate in the market. Opening it to outsiders might better simulate a real commercial market and provide greater opportunity for experimentation and analysis.
The professors, guest speakers, and students all generally enjoyed and learned from the course, finding it quite rewarding. The students gained hands-on experience in e-commerce technology, legal and policy issues, economics, marketing, advertising, and online business skills. The guest speakers uniformly noted how well the course reproduced current developments (and uncertainties) in real-world e-commerce. The students overall reaction was overwhelmingly positive, as reflected in their evaluations, high attendance rates, and requests for a follow-up course.
©1999 ACM 0002-0782/99/0800 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 1999 ACM, Inc.