A year ago e-commerce exchanges and online marketplaces were an exciting new idea, full of promise and funding. During 2000 hundreds of these e-markets were formed to revolutionize how businesses buy and sell. Today, however, independent e-markets are going out of business at almost the same rate they were formed. All of the hype in the B2B space has shifted to consortia of large firms that have banded together to buy or sell as a group. What has happened to the promise of online exchanges? And what will it take for them to really transform business. Along with Accenture colleagues Jeff Brooks and Sue Cantrell, I began researching these e-markets at about the time they took off in late 1999. We interviewed many companies and surveyed over a hundred independent e-markets. We determined that two concepts explain much of the present and future of e-markets: aggregation and integration. Aggregation: Easy and Cheap What e-markets offer in spades is aggregation, the opportunity to do business with larger numbers of buyers and sellers. Most online markets and exchanges were created to aggregate supply and demand across multiple companies. Sellers would be able to sell their products to an aggregated set of customers; buyers would be able to choose among an aggregated set of suppliers. Exchanges also offered dynamic pricing and comparison technologies that made it possible to shop among multiple providers and get the best deal. Before long, everyone had the potential to offer aggregation, but there simply wasn't enough buying and selling to go around. Our research suggests that in the average independent e-market only 10% of registered participants actually become active traders. Those who actually trade do so sparingly. The median number of transactions per month was 175, with an average value per transaction of $6,500. This means typical monthly volume through a market of just over a million dollars. At a commission level of 5% or less, this amounts to actual revenues of less than a million dollars per year. In other words, aggregation became so cheap and easy that too many e-markets were competing for too few transactions and dollars. For example, more than 15 exchanges operated in MRO (maintenance, repair, and operations). Another aggregation-related problem is that suppliers are naturally wary of being aggregated, and with good reason. On most exchanges the only factor used in deciding among alternative suppliers is price. And when many prices are displayed before buyers, they tend to choose the lower ones. The online auctions conductor FreeMarkets Inc., for example, estimates that it is able to lower costs to buyers by 15% on average. As one online exchange executive put it, "It's not a supplier's best day when he is asked to bid his product in [our] auction." Because this aggregation is not in the best interests of suppliers, they have been slow to jump on the e-market bandwagon. Aggregation capabilities don't automatically lead to business for e-market participants for one other reason. Just because a buyer can choose among multiple firms to do business with, he or she may not necessarily trust all of them to become long-term, strategic suppliers. An online marketplace may offer few clues about whether the supplier can deliver in a pinch, how long the supplier will remain in business, and other prerequisites for long-term relationships. As a result, buyers may be willing to buy only commodities or unanticipated demand from suppliers they encounter in e-markets. They may not be confident enough in e-market relationships to risk endangering long-term, strategic sourcing relationships they have built with offline suppliers. Toyota Motor Corp. took this approach, for example, when invited to join Covisint, the automobile industry exchange. A Toyota purchasing executive commented that the company would buy office supplies and bolts -- in other words, commodities -- but would maintain its longtime, offline supplier relationships. The only markets that are likely to actually see large volumes of business are the industry-sponsored consortia e-markets such as Covisint. They already have a lot of business to transact, and they can simply move it to these e-markets. In 2000 over 60 of these consortia e-markets were formed. They are not just taking business from independent e-markets, they are taking them over. For example, Converge, a consortia e-market sponsored by the high-technology manufacturing industry (including Hewlett-Packard Co., Compaq Computer Corp., and Samsung Electronics Co. Ltd.) recently acquired an electronic components exchange from a href=http://www.verticalnet.com>VerticalNet Inc., one of the more prominent independent e-market makers. Integration: Worth the Effort The other concept that is key to understanding e-markets is integration. Integration means the ability to communicate easily across organizations and connect the business processes of the buyer with those of the seller. If e-markets and exchanges had high levels of integration, the information systems of a company wishing to buy something would send out a request to purchase. The exchange or intermediary would determine which vendor offered the best price and the best fit with the purchaser's requirements. Then the request would be forwarded to the vendor's information systems, and all of the purchase transactions would be taken care of automatically. This vision is appealing, but it is just a vision. Integration is difficult, expensive, and time-consuming. It requires that trading partners coordinate their business processes and agree on what to call particular components, products, and services across the industry. A "request for price quote" on a "35 mm stainless steel flange" must mean the same thing from one organization to another. Working out this level of process and information integration took years when companies did it through electronic data interchange (EDI). It may take even longer with the Internet, which allows a much broader range of information to be exchanged than was possible through EDI. Of course, technology can help with the hard work of integration. One such technology is XML, the Extensible Markup Language. It will become the primary vehicle for electronic interorganizational communication, and will ultimately lead to new worlds of information and process integration. Industry groups such as RosettaNet, an electronics industry standards group, already have made considerable progress in using XML to define information and process standards, but many participants have yet to implement the standards in their businesses. Enterprise or ERP systems are another key integration technology. As firms begin to make mainstream purchases that are integrated with their production processes, they ultimately will be forced to integrate purchase information with that of manufacturing and financial systems. They'll have to know "available to promise" inventory levels and the profitability impact of a sale. Presently these can be done only through an enterprise system. Because integration takes considerable time, technology, and money, it's likely that only the largest and most successful e-markets will be able to pull it off. It also will require substantial cooperation among competitors to establish standards. These factors suggest that consortia e-markets will be the most likely to develop high levels of integration. In our survey of independent e-markets, we found that all independent sites want to become more integrated, but it's unlikely they'll have the resources to do it. In an uncertain funding climate, independent e-markets may run out of time and dollars. E-markets promised something entirely different from the past, but they turn out to be continuation of familiar trends: supply-chain integration, inventory reduction, and cooperation with competitors and customers. They also continue the trend in which large, established companies are better able to afford investments in technology and process redesign. In other words, e-markets help the rich get richer. Thomas H. Davenport is director of the Institute for Strategic Change at Accenture (formerly Andersen Consulting), Cambridge, Mass.