Chapter 5. Infomediaries in e-Commerce: Case Studies

Introduction

New uncertainties in the digital age are transforming the role of the intermediary. Businesses and consumers now face problems of information overload, privacy concerns and lack of trust in online transactions. Like intermediaries in the Industrial Revolution, digital intermediaries must address these new uncertainties and reinvent their roles to remain viable. In particular, a new type of intermediary-an "infomediary"-has emerged. Because almost all online activities require searching for, obtaining, and processing information in some form or another, infomediaries step in to simplify the process, becoming information brokers.

In 1999, John Hagel and Marc Singer laid out a business plan for a specific kind of infomediary in their book Net Worth. Shortly after the book's release, several companies set out to launch infomediary services. This chapter examines these hopefuls, and seeks to determine why Hagel and Singer's infomediary model failed to make an impact in e-commerce. In addition, it seeks to determine why other digital-age intermediaries have had greater success.

To this end, this chapter first describes Hagel and Singer's infomediary model, and considers its potential to alleviate transaction costs and uncertainties. Second, it examines three businesses efforts-Lumeria, AllAdvantage and Passport-based on Hagel and Singer's plan and analyzes why infomediaries failed. Third, it considers two successful online merchants-eBay and Amazon.com-to determine how they act as intermediaries to generate trust and facilitate transactions.

Hagel and Singer's infomediary model

John Hagel and Marc Singer propose an infomediary business model in the book Net Worth: Shaping Markets When Customers Make the Rules (1999). They recommend that infomediaries act as a trusted third party to aggregate customer information, using their combined market power to negotiate with businesses on these customers' behalf, while at the same time protecting consumer privacy (Hagel & Singer, 1999). Accordingly, customers would receive a privacy tool kit providing them with an anonymous email address and filtering technology to block unauthorized vendors. Customers would use an infomediary-issued credit card, allowing the infomediary to track all of the customer's online and offline purchases, and to create a detailed personal profile of their purchasing behaviors. This anonymous payment mechanism would also prevents vendors from accessing personal data (Hagel & Singer, 1999).

By performing searches for the customer, the infomediary would alleviate problems of information overload. If a customer wanted a specific product, the infomediary would solicits bids from vendors by literally commodifying the customer's personal information and selling it to prospective vendors. To protect the customer's privacy, the transaction would be carried out anonymously, using email and targeted marketing. Assured of their security and privacy, individuals would welcome this targeted marketing (J. E. Campbell & Carlson, 2002; Downes & Mui, 1998; Hagel & Singer, 1999).

Under this system, both the infomediary and the customers would profit from the sale of personal information. For the consumer, the basic cost of membership-including the privacy and profiling services-would be free, and customers would pay a 2 percent agency fee when purchasing products through the infomediary. However, they would receive a portion of the vendor's payment for accessing their information profiles. Hagel and Singer estimate that, by revealing their information, a household would save approximately $1,400 per year through lower purchase prices and payments (Hagel & Singer, 1999).

Hagel and Singer contend that their infomediary model creates value not only for consumers but also for businesses. For, while customers have lower transaction costs and reduced uncertainties, businesses benefit by obtaining direct access to interested customers. For example, currently, 98 percent of customers ignore marketing messages such as pop-up ads and spam. By offering vendors direct access to customers that are known to be interested, businesses can save money through higher response rates, which in turn increase the value of the customer information (Hagel & Singer, 1999).

As we have seen from the previous chapter, for an intermediary to achieve a critical mass, customers must be able to trust them. Thus, to gain this trust, the infomediary must be "customer-facing" and not "vendor-facing". Customers must trust the infomediary to prevent unauthorized access and abuse of their personal information (Hagel & Singer, 1999). The infomediary would need to build trust through reputation and a positive feedback cycle, whereby it would then achieve market dominance and sustained growth.

To help achieve market dominance, Hagel and Singer's infomediary model would take advantage of information commodities. Although infomediaries can operate in any market, their prospects are greatest in information markets where market failures tend to generate high transaction costs. Thus, products and services with a high content of information, such as computer goods, mortgages and telecommunications services offer good opportunities for infomediaries (Hagel and Singer, 1999, 53). Because infomediaries would operate over the Internet, consumers would benefit from the infomediaries' capacity to search for and bundle information goods.

Infomediaries in practice

Shortly after the release of Net Worth in 1999, several new companies that used Hagel and Singer's infomediary model emerged. These newcomers included Lumeria, Privaseek, AllAdvantage, Firefly, PrivacyBank.com, Popular Demand and Enonymous. Three of these cases are discussed briefly below.

i. Lumeria

In 1999, Lumeria was poised to become the most successful of the new infomediary businesses. Lumeria was founded on the principle that a customer's personal information must be safely guarded, yet easily accessible and anonymously shared with others (Lumeria Inc., 2000b). In keeping with Hagel and Singer's model, Lumeria adopted a "customer-facing" strategy. Customers stored data in a "SuperProfile," giving Lumeria permission to sell that information to a network of vendors. The vendors could then perform highly targeted, permission-based marketing to those people who had expressed interest in the merchandise. Vendor fees would be paid directly to consumers, with Lumeria taking a small percentage (Lester, March 2001; Lumeria Inc., 2000b).

Closely following Hagel and Singer's plan with its "Sunshine Platform" technology, Lumeria promised to "dramatically improve the security, efficiencies, and effectiveness of any online transaction, whether for commerce or content, for business or for individuals," (Lumeria Inc., 2000a). The Sunshine technology offered the following services:

In the words of Lumeria's founder, Fred Davis, "basically, we created a new piece of Internet infrastructure for the secure communication and authentication of transmissions across the Internet. It took us a few years and millions of dollars to develop it, but now it's here, and it's pretty cool,"(Lester, March 2001). Unfortunately, Lumeria was not cool enough. As of 2002, Fred Davis was seeking a buyer for the company (Delevett, 2002, February 19).

i. AllAdvantage12

Another online infomediary, AllAdvantage, also followed Hagel and Singer's plan. Launched in March 30, 1999, AllAdvantage allowed individuals to sell their attention and personal information to advertisers while maintaining their personal privacy. To accomplish this, AllAdvantage provided a software program -the "ViewBar"-that allowed customers to view targeted advertisements on their personal computers. AllAdvantage paid customers fifty cents and hour (up to forty hours per month) for viewing the advertisements, while businesses paid AllAdvantage for this targeted advertising opportunity. To increase participation and build a critical mass, AllAdvantage encouraged its members to make referrals. For example, it paid its members small fees (based on a descending scale of four levels) for the time that a referral, or referral of a referral, etc., spent searching the web. In addition to its advertising fees, AllAdvantage intended to capitalize on the information stored in members' browsers, which documented keyword searches and sites visited. By exploiting this information, AllAdvantage could better target ads, and thus charge a premium for advertising space (Beal, 2002).

Although AllAdvantage successfully generated a critical mass, the overall business model was unsuccessful. In 2000, AllAdvantage grew exponentially in popularity and membership rates. Within five months of its launching, 10,000 to 20,000 members signed on each day. By July of 2000, total membership to AllAdvantage surpassed seven million (Beal, 2002). Nonetheless, problems mounted. As AllAdvantage's founder, Jim Jorgenson, explained, "It was a chicken and egg problem," (Beal, 2002). Because the market for targeted advertising was still young, few businesses were ready to subscribe to AllAdvantage's service. Furthermore, due to software constraints, only 2 to 3 percent of the ads were actually targeted to members. Another fatal flaw in the business model was customer behavior. Receiving money for merely looking, most members did just that-only rarely did they purchase the advertised merchandise. Some even blocked the ads on their computer screen with duct tape. As a result, AllAdvantage was unable to add value for their advertisers. Moreover, they were obliged to pay members, regardless of whether or not they sold advertising space. Because membership rates were so high, payments to members reached the millions, and AllAdvantage quickly burned through its investor money. In the wake of the dot.com crash, funds disappeared, and AllAdvantage folded (Beal, 2002).

iii. Firefly/Passport

A third infomediary, Firefly, predates Hagel and Singer's infomediary plan. Firefly began in 1995 as a program called Helpful Online Music Recommendations (HOMR), using intelligent software to recommend music and bands (Oakes, 1999, August 12a). Later, Firefly developed an infomediary service called "Passport", which collected demographic and psychographic information from consumers. Passport also used customer information to offer personalized suggestions and reviews on items of interest (Kannan, Chang, & Whinston, 2001).

In 1998, Microsoft acquired Passport so it could bundle the software with its XP operating system and accelerate the delivery of its privacy-enhancing services. Microsoft's Passport Wallet, which collected a user's credit card and password information in a database, was designed to simplify e-commerce transactions. However, the purchase met opposition from privacy advocates who accused Microsoft of becoming "Big Brother," and trying to control customer profile databases (Luening, 1998).

Then, in October of 2001, the Passport software exhibited a serious flaw, which would expose users' password, credit card information and other personal data (McWilliams, November 2, 2001). In 2001, privacy advocates filed a complaint with the FTC, claiming that Microsoft's Passport had threatened the privacy of the 100 million users by deceptively collecting consumer information (Manjoo, August 16, 2001). In 2002, Microsoft settled the charges with the FTC. According to the complaint settlement, Microsoft had falsely represented that:

Passport software is still available for Microsoft users, but the security and privacy flaws have prompted many users to opt out of its services. The case illustrates that, even for companies that have tremendous resources at their disposal, the infomediary model has some inherent flaws.

Shortcomings of infomediaries

Notwithstanding the myriad uncertainties in the digital age, infomediaries built along the lines of Hagel and Singer's model have yet to succeed. The NASDAQ crash of 2000 and the subsequent withdraw of investments from e-commerce contributed to the demise of Lumeria and AllAdvantage. But equally, if not more, important was the fact that a large majority of businesses and consumers rejected these privacy-based infomediary business models (Bodorik & Jutla, 2003). As we have seen in chapter four, a successful infomediary must follow seven rules for success. Hagel and Singer's model followed two of these rules: it took advantage of information commodities and helped organize and bundle large amounts of information, and it attempted to offer consumers advice on e-commerce purchases in specialized industries. As described below, by neglecting five of the seven rules, infomediaries failed to reduce uncertainties for consumers and businesses.

1. Provide value for consumers and businesses

Infomediaries did not provide value for both businesses and consumers. For customers to trust infomediaries with their personal information, the infomediaries must be totally "customer facing" (Levin, 1999). However, by focusing on consumers almost exclusively, infomediaries were unable to provide enough value for businesses. To make a profit, infomediaries must have the participation of a vast network of businesses. But businesses will only participate given a significant positive net gain. Moreover, an infomediary must be more efficient than businesses that act independently to bring buyers and sellers together. In the case of AllAdvantage, businesses did not benefit from targeted advertisements to interested customers, leading to its failure (Chen, Ganesh, & Padmanabhan, Fall 2002; Oakes, 1999, August 12b). In some cases, working through an infomediary can decrease net gains for businesses. With direct access to an increased number of customers, businesses can poach on their competitors' customers. Thus, infomediaries can unravel, meaning that no business will gain a net profit from joining, even if the access is free (Chen et al., Fall 2002).

2. Promote trust in online transactions

The infomediary's inability to establish a high level of trust also helped to account for the model's failure. Although the infomediary is a third party, there are no guarantees that it will be a 'trustworthy' third party (Shearin & Maes, no date). Especially in a rapidly changing e-commerce environment, in which most companies are unstable, consumers are unlikely to trust so much personal information to any company, third party or not (Muller, 2000). For example, customers who invested time creating profiles on Lumeria must be wondering now what has become of their personal information. Thus, unless consumers are offered substantial benefits for revealing their personal information, they will be uncomfortable with computer programs tracking their every move, and unwilling to transact through a single infomediary (Bodorik & Jutla, 2003; "The rise of the infomediary," 1999, June 24; Vora, December 1999).

3. Respect information privacy

Ironically, although infomediaries were conceived to protect individual privacy, they ultimately generated additional privacy concerns (Dix, no date). Controlling personal information, infomediaries determine the perceived level of privacy. And how can consumers be one hundred percent certain that these agents will act entirely on their behalf? As the case of Microsoft's Passport illustrates, once such comprehensive personal profiles have been put together, there is a significant risk of malicious misuse by the infomediary or an outside agent (Bodorik & Jutla, 2003).

4 & 5. Achieve critical mass, and gain a first-mover advantage

As we have seen in chapter four, network effects are essential to an infomediary's success. Although Lumeria heeded many rules for success, it failed to achieve a critical mass of users and a first mover advantage, so it could not sustain its efforts. In contrast, AllAdvantage achieved a critical mass of users, but the expense involved, together with the fragile economic environment of 2000, caused it to lose its edge. AllAdvantage also failed to take advantage of positive network effects to the detriment of its vendor network. Unable to properly analyze its customers' browsing and purchasing records, the company could not infer needs and preferences for the benefit of its vendors (Vora, December 1999).

These failures to gain a critical mass raise the question of who is better suited to become an infomediary: a startup company or an incumbent. Newcomers must convince millions of consumers to join, promising them benefits down the road. Given their existing customer database and a critical mass of customers, existing companies might have better luck. But herein is the paradox. An existing company is unlikely to be a successful infomediary, because customers would be wary that the company would act on its own behalf. Even if a startup infomediary were to form a joint venture with an existing company, it would not likely solve the problem of trust. If the infomediary ran into financial trouble, there would be no guarantees to prevent the sale of personal information to un-trusted third parties (Hagel & Singer, 1999; Muller, 2000; Rosenblatt, July 1999).

The case studies described above suggest that the business model of the infomediary is not financially viable. The authors of the plan estimated that an infomediary based on a joint venture would require an up-front investment of $380 million. According to their estimates, positive returns could be expected in the eighth year, with revenues of $5 billion by the tenth year. These high fixed costs and low profit margins are unattractive to venture capitalists, who want to see more immediate returns on their investments. Even Hagel and Singer admit that for a startup company, the investment would be significantly more (Hagel & Singer, 1999; Muller, 2000; Rosenblatt, July 1999).

Alternatives

Although no company has yet to develop a successful infomediary on the order conceived by Hagel and Singer, more modest versions of infomediaries, taking the form of search engines, communities of interest, and corporate sites, do exist (Grover & Teng, 2001). Amazon.com, eBay, Expedia.com and PriceLine.com are forms of infomediaries. They work to reduce search costs and uncertainties associated with e-commerce. The main difference between these digital intermediaries and Hagel and Singer's infomediary is that the latter do not deal with digitally related uncertainties by commodifying consumer information and privacy. Rather, these companies succeed because they establish a social contract with their customers. Social contracts define collective privacy policies that e-commerce businesses adhere to, and they constrain the behavior of companies while protecting individuals (Kaufman & Powers, 2002).

i. eBay

eBay, the largest Internet auction site, exemplifies a successful infomediary. Unlike the physical auction houses of the 19th century, eBay brings buyers and sellers together in cyberspace, allowing millions of goods to be auctioned off simultaneously. It does not act as a principal, because it has no ownership interest in the goods sold. As an infomediary, its only interests are to protect the integrity of the online auction process by providing information links. eBay profits by charging fees to list items, and collecting a percentage of the profit of transactions (Bunnell & Luecke, 2000).

Although eBay was not the first online auction site, it was the first to achieve critical mass, capturing a dominant share of buyers and sellers. By achieving critical mass, it gained the first-mover advantage necessary for success. Thus, eBay benefits from positive network effects, because each additional member benefits both buyers and sellers (Bunnell & Luecke, 2000).

By outsourcing some of its core services, eBay maintains a vertically disintegrated business model. The most important outsourced activity, web hosting, is assigned to AboveNet and Exodus Communications. In addition, eBay maintains relationships with PayPal, its preferred secured payment service, and Mailboxes Etc., which provides a 10-15 percent discount for eBay users. By relying on others to perform these key tasks, eBay can focus on its core business practices (Bunnell & Luecke, 2000).

eBay's culture has served to enhance an image of trustworthiness. eBay was founded on core values of honesty, openness, equality, empowerment, trust, mutual respect and mutual responsibility. This foundation has prevented it from becoming opportunistic, decreasing uncertainties for its members (Bunnell & Luecke, 2000). In online transactions, trust is established through 'feedback'-reviews submitted by buyers and sellers, vouching for the quality of the product, ease of payment processing and arrival time of the good purchased.

Nonetheless, eBay's trusted system is subject to fraud, raising concerns about eBay's validity as a model infomediary. With computer technologies, it is easy to cheat on the Internet, so online companies must continuously struggle to keep swindlers at bay. In fact, according to the Federal Trade Commission, online auction frauds are more prevalent than any other type of online fraud. In 2003, they accounted for 48 percent of all fraud complaints filed (Krebs, January 23, 2004). Yet eBay itself claims that only one-hundredth of one percent of the 20 million items that it lists at any given time are fraudulent (Hafner, March 20, 2004).

To cope with the possibility of fraud, eBay complements the feedback system, acting as its own intermediary to build trust and provide protection. To this end, eBay takes the following security measures:

Although eBay began as a person-to-person channel of exchange, small businesses increasingly use its services to attract customers. In 2000, an estimated eighty percent of eBay's revenues were generated by twenty percent of its users, most of which are businesses using the site as a portal to their own website. Thus, by simplifying the search and transaction process, eBay benefits businesses and consumers (Bunnell & Luecke, 2000).

To address privacy concerns, eBay adheres to a strict privacy policy. It does not sell or rent personal information to other businesses without explicit consent of its members (eBay, 2004). The information that eBay does collect allows the company to customize its services. Thus, eBay uses cookies to track web flow, measure promotional effectiveness, and promote trust and safety. It aggregates personal information and discloses it to advertisers and third parties for marketing and promotional purposes. However, this information is disclosed in a way that precludes any personal identification (eBay, 2004). By forming privacy-based social contracts with its members, eBay has decreased the uncertainties entailed in e-commerce.

ii. Amazon.com

Amazon.com, "the earth's largest bookstore," is another example of a successful digital-age infomediary. Some refer to Amazon.com as an example of disintermediation, because it allows customers to bypass traditional brick-and-mortar bookstores to purchase books online. But Amazon.com does not disintermediate-it does not produce books, it only distributes them. Amazon.com.com is in fact an intermediary, bringing buyers and sellers together, establishing trust and reducing search costs associated with e-commerce. To support its activities, Amazon.com embeds filtering technology that enables personalized recommendations to be shown based on buy and search history, thus digitizing its expertise in the book market (Chircu & Kauffman, 2001; A. L. Shapiro, 1999).

Amazon.com benefits handsomely from its first mover advantage, which in turn provides positive network effects. As more users participate, everyone benefits from Amazon.com's search functions, recommendations, and user reviews. Amazon.com's first-mover advantage has allowed it to amass market value over thirty times that of BarnesandNoble.com (Chircu & Kauffman, 2001). This market dominance has enabled Amazon.com to expand its selection to music and merchandise as well.

Like eBay, Amazon.com vertically disintegrated by forming partnerships and alliances with other firms. Amazon.com follows three business models: retailer, a platform provider for other retailers (such as Circuit City) and an auction marketplace. It forms partnerships with manufacturers (such as the Disney Store), smaller businesses and individual sellers (E-Business Strategies, 2001, October). Furthermore, Amazon.com employs other companies to perform functions such as shipping, filling orders, analysis, marketing, processing payments and customer service (Amazon.com, 2004)

Amazon.com is a trusted web merchant, building a reputation through branding and service. In fact, consumers trust Amazon.com so much that they frequently buy from the site even when it does not have the lowest prices, confirming the importance of trust in e-commerce (M. D. Smith et al., 2000). Yet Amazon.com ran into trouble in 1999 when the public discovered that publishers were paying more than $10,000 to have their books featured with accolades such as "New and Notable" and "Destined for Greatness." By allowing publishers to pay for placement, Amazon.com put the needs of the publisher above the consumer. Customers lost faith in the site as a result. To overcome the problem, Amazon.com quickly changed its policy, promising to disclose when publishers had paid for placement. Still, the incident unfavorably affected the perceived level of trust in Amazon.com as an unbiased intermediary (A. L. Shapiro, 1999).

Like eBay, Amazon.com adheres to a strict privacy policy based on social contracts. Amazon.com uses cookies to track information about its customers, using this information to personalize the shopping experience by recommending future purchases, one-click purchasing, personalized greetings, wish lists and shopping carts. Amazon.com does not sell personal information to other businesses, but it does share its customer information with its subsidiaries (Amazon.com, 2004). Consumers do not mind that Amazon.com collects information, because it allows the site to offer personalized recommendations.

Conclusion

Hagel and Singer laid out a complex model for infomediaries in their book Net Worth: Shaping Markets When Customers Make the Rules. Their business plan attempted to resolve many uncertainties of the digital age: information overload, information asymmetry, privacy concerns, consumer data collection and trust in e-commerce. By controlling and aggregating personal consumer information, infomediaries commodified information privacy. Consumers then sold their personal information in exchange for targeted marketing messages from approved vendors. At the same time, infomediaries used the consumer information profiles to make recommendations and anticipate the needs of their customers.

In 1999, infomediaries were the 'next big thing' in e-commerce. Companies such as Lumeria, AllAdvantage and Passport closely followed Hagel and Singer's plan. Yet, no infomediaries created after the release of Hagel and Singer's book still exist today. Although Passport, now owned by Microsoft, is still functioning, it faces serious allegations over the misuse and violation of personal consumer information.

There are several reasons for the failure of infomediaries. First, Hagel and Singer introduced their business plan before the dot.com bubble burst. Investors clamored to sink money into potential e-commerce companies. The demise of AllAdvantage and Lumeria corresponds with the crash of the stock market, signaling that it may have been bad timing.

Second, in placing so much value on the consumer, infomediaries failed to create enough value for the business. This lack of balance was especially evident in the case of AllAdvantage; its business plan unraveled when businesses became disaffected, and the company still had to pay millions of dollars to their customers. As we saw in chapters Three and Four, an intermediary performs best when it meets both business and consumer needs equally.

Third, infomediaries failed to garner enough trust from their customers. A high level of trust is crucial in e-commerce, as without it transaction and agency costs are too great to sustain a market. However, the benefits of having an infomediary did not outweigh the risk entailed in revealing personal information. In Lumeria's case, customers were unwilling to trust an unknown company to handle all of their private information.

Fourth, infomediaries overestimated their ability to reduce privacy concerns in e-commerce. They failed to take into account the fact that consumers desired a perceived level of control over their own personal information, and that entrusting that information to an infomediary did not equate with control. This problem suggests that infomediaries cannot resolve uncertainties relating to information privacy. Rather, privacy concerns should be resolved through transparent privacy policies and social contracts.

Fifth, infomediaries failed to take into account key aspects of a networked economy. To succeed, infomediaries needed a critical mass of users and a first-mover advantage. Without a critical mass, infomediaries, most notably Lumeria, could not sustain their high fixed costs. Although AllAdvantage had a critical mass of users, it was unable to take advantage of the ensuing network effects.

The failure of Hagel and Singer's infomediary plan does not suggest that all infomediaries are doomed. In fact, several infomediaries have emerged in recent years to become dominant market players. eBay and Amazon.com are most notable in this regard. Shunning the personal information and privacy business, these companies are remarkably distinct from the model business proposed by Hagel and Singer. They are intermediaries in a more traditional sense- bringing buyers and sellers together, stepping in to reduce search costs, transaction costs, information asymmetry and information overload in e-commerce. They have succeeded by adhering to the seven rules of success for digital age intermediaries. They resolve tensions in consumer information collection and privacy by entering social contracts with their customers, thus restraining tendencies for opportunism.