B2B commerce entails using new and emerging Internet-based technologies to support business-to-business transactions among trading partners. B2B commerce systems differ from traditional Electronic Data Interchange (EDI) systems in many ways. First, B2B commerce systems go beyond document interchange (the main objective of EDI) and take a holistic view of business models while integrating business processes. Second, B2B commerce systems can rely on electronic exchanges (e-exchanges) that provide intermediary services for B2B transactions (Phillips & Meeker, 2000). Exchanges provide tools for supplier and price discovery as well as for establishing business arrangements "on the fly." This contrasts with traditional EDI systems, which are governed by predefined contracts and business arrangements existing between trading partners. Third, in comparison to traditional EDI systems, B2B commerce systems use standards-based Internet technologies and open systems architecture. EDI systems, while following the guidelines of standards bodies such ANSI (X.12) and EDIFACT, are often implemented on top of proprietary services providers such as the Value Added Networks (VAN) and proprietary software technologies. Factors that impede the widespread adoption of EDI include the focus of EDI on just data exchange formats, the limited reach and scope of VANs, total cost of ownership of proprietary systems, and the lack of flexibility to provide dynamic end-to-end B2B services.
E-exchanges as intermediaries in B2B transactions play a key role in (1) matching buyers to sellers that include determination of product offerings, search and price discovery; (2) facilitating B2B transaction by providing logistics, settlement and trust services; and (3) providing all the institutional infrastructure in terms of legal and regulatory framework and support necessary for conducting B2B commerce (Bakes, 1998). E-exchanges support various market mechanisms such as auctions, reverse auctions, automated Request for Proposal (RFP) systems, negotiations and clearing (Beam et al., 1999; Liang & Huang, 2000). E-exchanges form the hub of "the business Web," a phrase that denotes a network of suppliers, distributors, service providers and customers that facilitates the creation of value for end customers and each other (Tapscott, Ticoll, & Lowy, 2000).
Many factors determine the adoption or non-adoption of e-exchanges by enterprises. Among these, the cost of connection enablement to an e-exchange can be a key factor for non-adoption especially when enterprises have to incur significant effort and investment in integrating with an e-exchange. Other factors such as trust, privacy and anonymity, risk aversion and the support of key players play a major role in adoption (Hsiao, 2001; Koch, 2002) of e-exchanges. E-exchanges can be privately owned by enterprises or owned and managed by independent third party service providers. The ownership and management of exchanges has lead to the evolution of three basic types of electronic markets: buyer centric, seller centric and neutral. In the buyer-centric markets, few major buyers host an e-exchange to transact business with a large number of suppliers. In this case, the buyers have sufficient transaction volumes to justify hosting an e-exchange. Examples include K-Mart's Retail Link and Covisint. When there is one buyer and many sellers, the e-exchange serves as a dedicated procurement hub for the buyer. Buyer-centric markets are generally biased towards the buyers with the buyers exhibiting significant clout in the electronic markets. In the seller-centric markets, few major sellers host an e-exchange to trade with a large number of buyers. In this case, the sellers have sufficient transaction volumes to justify hosting the e-exchange. Examples include Grainger.com and GE Global exchange. Seller-centric markets are biased towards sellers with the sellers exhibiting significant clout in the market place. In neutral markets, an independent third party hosts an e-exchange where buyers and sellers conduct business transactions. A major advantage of a neutral market is that participants in the exchange assign a higher level of trust to the functioning of the electronic market compared to buyer-centric or supplier-centric markets. Therefore in theory, neutral markets would attract more participants than biased markets. Examples of neutral markets include NASDAQ and Arbinet.com.
Electronic markets can also be viewed in the context of vertical and horizontal industry domains. Electronic markets for a vertical industry domain serve a particular industry by providing content as well as processes to support that specific industry. Examples include Transora (Consumer products and services), Covisint (Automobile), Cordiem (Aviation), and Converge (Technology marketplace). In contrast, electronic markets for horizontal industry domains, serve a large number of industries by providing services that are common to them. For example, Grainger.com, MarketMile (by American Express and Ventro), and ICGcommerce provide maintenance, repair and operations (MRO) supplies used by many industries.
Many different vendors such as Ariba (www.ariba.com), Commerce One (www.commerceone.com), Ventro (www.ventro.com) and I2 (www.i2.com) provide e-exchange software. The current exchanges are second-generation exchanges. These exchanges have evolved from the first generation exchanges, many of which have failed in their endeavors due to the lack of realization on the part of exchange providers that exchanges as intermediaries, hindered long term relationships between trading partners where quality and trust often mattered more over price.
While the core objectives of e-exchanges across vendor products are similar, there are no standardized methodologies for implementing services, interfaces, and the exchange architecture. Every exchange implements a given exchange functionality differently, using a variety of enabling technologies. For example, some exchanges require suppliers to use an X.12 EDI for transactions, whereas others require proprietary formats. The standards and specifications used for integrating business partners also vary widely such as FTP, HTTP, XML-RPC, S-MIME, JMS and SOAP. In some cases proprietary middleware are used. This hinders buyers and sellers to connect to exchanges "on the fly," as they have to use exchange-specific custom interfaces and enabling technologies. Thus, exchanges and the participants in the exchanges are being forced to invest in expensive application and systems integration efforts, by using software, managed services or both. The first generation exchanges underestimated the efforts required for supplier and buyers to participate in the marketplace. Supplier enablement can be attributed as one of the causes as to why many exchanges could not achieve a critical mass of participants.
To enable large numbers of buyers and sellers to participate in an e-exchange "on the fly," it is essential for exchanges to support a variety of interfaces, data formats, and trading mechanisms and styles. The dynamics of the market place requires e-exchanges to have highly flexible and scalable architectures. One way to accomplish this is to use intelligent software agents (described later) to build e-exchanges and enterprise systems for B2B commerce. The following section describes the information technology infrastructure requirements for intelligent enterprises to conduct B2B commerce.