Historically, middlemen or intermediaries have fulfilled an important role in getting goods from the manufacturer to the end user -- for example, from farmer to broker to distributor to food store to consumer, or from steel mill to manufacturer's representative to machinist. Because the intermediaries don't change the physical attributes of an item, the only function they seem to perform is moving information. Electronic commerce is supposed to reduce the cost of goods through the elimination of brokers and distributors, by automating the exchange of information directly between the producer and the user. In practice, however, it becomes obvious that the middleman does more than just repeat information. He or she interprets, translates, catches errors, and in many other ways facilitates the transaction. Many skilled suppliers, such as farmers or machinists, don't want (or can't afford) to become experts at electronic commerce systems. They will hire someone to do this work for them: a new middleman. Examples from such diverse industries as aerospace and photography are used to demonstrate the range of possibilities for middlemen in the new world of electronic commerce.
Business relationships are the combination of contracts, courtesy, confidence, and communication that allows the buyer and seller to exchange things of value with a sense of tangible net gain.
A business relationship has endurance, meaning it supports repeated transactions. It is fault-tolerant, meaning that it supports the resolution of mistakes and problems. It is flexible, in the sense that it can accommodate some amount of change in the things being exchanged. It enhances each participant's position by satisfying some tangible material or personal need of each participant.
A business transaction results in tangible gain to each party. Acts of charity, where one party gains a sense of well-being or promise of benefit after death, are not business transactions. Likewise, acts of aggression, where both parties know that one party will lose while the other will gain, are not business transactions. Acts of entertainment or enlightenment, where one party gains an experience but no financial or material gain, nor the means to acquire financial or material gain, are not business transactions. These and other transactions are outside the scope of this paper, although they are certainly occurring in the e-commerce world and frequently involve middlemen.
Superior business relationships exist when each participant believes that this relationship is the best available method for satisfying his or her material or personal need or needs.
When the first buyer and the first seller exchanged things of value, they did it face to face. When the first buyer showed her purchase to her neighbor, the neighbor asked her to buy enough for two next time and gave her some item of value. The buyer, acting on her neighbor's behalf, exchanged that item of value for the seller's item and brought it back to her neighbor. Although this transaction used an intermediary, it is not the first example of a middleman because there was no tangible value gained by the first buyer. Some time later, the seller lowered his price to the buyer, who continued to receive the same payment from her neighbor. Thus, the middleman was born!
A middleman is a person or organization that facilitates the business relationship between a buyer and a seller, and earns a tangible gain from its activity.
It is in this context that we explore the new role for the middleman that results from the technology and systems we call electronic commerce.
Much of what I know about middlemen comes from my own experience in the manufacture and distribution of electronic instruments and retail automation equipment. Most recently, I have had a unique opportunity to observe how electronic commerce affects the business models of small- and medium-sized manufacturers, as director of the Electronic Commerce Resource Center.
In my first professional job, I was an engineer in a circuit board fabrication facility for a large electronic instrument manufacturer. In those days, material requirements were determined on a monthly basis by taking the predicted sales of oscilloscopes and breaking that down to the number of each type of circuit board to be built. On a weekly basis, a report would be generated that would identify the changes to the monthly forecast. There was no automated way, however, to convert the number of individual circuit boards of all shapes and sizes into the number of 18x24-inch panels of copper-clad raw material, which is what the purchasing agent had to order. This conversion was a tedious, manual calculation.
Suppliers of the copper-clad raw materials were eager to sell their products, and their employees tried very hard to win and keep the account, including performing these calculations. The forecast and change reports were issued on paper, and of course, were considered company confidential. So, every Thursday, in the lobby of the Electrochem building, the father and son team that represented the laminate supplier would pore over the report for hours, in order to calculate the specific requirements so they could recommend to the purchasing agent what to order. If they had not done this work, the purchasing agent would have had to do it himself, and would have done it less frequently. By volunteering to do this tedious task, this supplier got better information sooner about what to order from their manufacturer. The purchasing agent got the benefit of a more carefully controlled inventory level than he could have achieved on his own. This supplier added value to the transaction by saving the company money in inventory-carrying costs.
As a production manager in a laser scanner factory, I implemented just-in-time approaches in my assembly areas and observed the adoption of vendor-managed inventory practices by the Materials group. In exchange for a commitment to purchase all of our electronic components from a particular supplier, the supplier set up and maintained an inventory stock in our factory, based on our production forecast. He kitted the parts into the particular assortment required to build a scanner and delivered them to the assembly area using a kanban system, which drew parts only when needed. This eliminated a few jobs in our warehouse, but the employees were hired by the supplier. In addition to saving my company the inventory carrying and kitting costs, we had 24-hour access to the supplier in case of quality problems or sudden changes in requirements.
I was hired by one of the laser scanner company's customers, a distributor of scanners to grocery and discount retailers. To my amazement, the retailers were willing to pay a third party (the distributor) to plan, schedule, ship, call for pickup, and repair scanners. The retailer was not avoiding a particularly difficult job by relying on us. Rather, the retailer was relying on us to do an excellent job on a trivial, noncore aspect of its business, namely, keeping the scanners running and expediting the conversion of nonscanning stores to scanning. What was an all-consuming business for us as a supplier was a mere detail to our customer. Our ability to perform this service consistently satisfied the customer's need for reliable, problem-solving service.
At the Electronic Commerce Resource Center, we have been observing, interpreting, and predicting the direction of business-to-business electronic commerce since 1994. Our objective is to help smaller businesses make sensible decisions about equipment investment, employee training, and subscriptions to electronic commerce services.  In 1995, a distributor of a data-over-copper-wire product came to our center and learned how to put up a Web page. Combining the World Wide Web's global reach with information about the manufacturer's approval as a supplier to the Army and their own less-than-book pricing, they were able to get orders from Army bases in Turkey, where previously their business had been limited to the Puget Sound area. Two years later, however, the distributor was somewhat disenchanted. Other distributors had followed their lead, leading to a price war that only stopped when the manufacturer intervened. The Army's electronic commerce initiative had served its purpose, lowering their average cost for this item by about 20 percent. The distributor had to retool his business to survive on lower margins, and was again differentiating himself from other distributors by offering rapid delivery, installation assistance, and so on.
This is an international conference, with representatives from all over the world. I have some personal knowledge of business practices in the European Union, South America, Israel, and Japan. I am well aware that what I don't know about how the world does business is much greater than what I do know. In this paper I have attempted to explain myself and use examples clearly enough so that those of you operating in different environments can recognize what applies and what does not.
As humans, we have developed a fairly standard response to conflicting pressures. When I am faced with the desire to make a deal, but fearful of being cheated, I seek out a mutually respected third party to assist in the transaction. Without some source of trusted assistance, the deal will not be made. By reducing the risk of the transaction, a trusted third party can increase the number of willing buyers as well as the number of willing sellers.
Fundamentally, middlemen absorb risk. In a slow, paper-based economy, they absorb currency foreign exchange risk. Where physical distance separates buyer and seller, they absorb transportation loss and damage risk. Where language and cultures differ, they absorb the risks of miscommunication and misunderstanding. Where buyer and seller are unknown to one another, they absorb the risk of bad intent. Unabsorbed, these risks reduce the number of willing buyers and willing sellers. With the middleman, these numbers increase, trade increases, and the compounding effects of the world economy generate benefits for all.
Vertical integration -- owning your sources of supply and of distribution -- was the historic alternative to middlemen. When the Ford Motor Company owned the entire automotive value chain, from iron ore deposits through showrooms and financing companies, it was fully vertically integrated. The danger is that a single company is unlikely to be equally good at each part of the chain. In today's competitive world, few companies can afford to carry all the risks of vertical integration.
Instant, cheap, accurate information transfer eliminates substantial amounts of risk, and sometimes eliminates the related need to hold stock or even to take possession of goods.
If middlemen did not add value, they would cease to exist because buyers and sellers would find it more economical to deal directly with each other. The continued existence of middlemen, and the emergence of new forms, are evidence that middlemen add value.
In a complex or sophisticated environment, or when one of the parties is much more competent than the other, specialist technical knowledge is a key asset for middlemen. An example of an industry that would benefit from technically knowledgeable middlemen is the outdoor recreation equipment industry, which caters to climbers, hikers, paddlers, bicyclists, and skiers. This market has a huge and diverse pool of suppliers. Many of the suppliers of unique items are outdoors enthusiasts themselves, who have come up with an effective solution to a common problem and want to support themselves by manufacturing the product and selling it to retailers. The last thing these climbers and skiers are concerned about is inventory level or scheduled deliveries. The large outdoor equipment retailers are much more sophisticated, and rely on receiving goods on schedule in the ordered quantities and sizes.
Middlemen bring additional capital to the industry segment they serve. This capital is used to carry inventory during periods of low demand so that peak demands can be met without wide swings in production rates. Capital is also invested in warehousing facilities and trucking costs, and in promotion of the product to potential buyers.
When a supplier operates in one location and their buyers are located at a distance, middlemen who are located near the buyers add value in the form of market knowledge, personal relationships, and intimate knowledge of the prevailing culture and laws.
The key value added by the middleman is trust, the inverse of risk. We trust people who specialize in a field that is unfamiliar to us, who have experience, and who are referred by trusted acquaintances. In the United States, at least, trust is enhanced by simplicity of interactions, and by successive positive experiences that build confidence.
Today business processes are in one of three stages: Pre-Web, Web discovery, and post-Web.
Pre-Web business processes are essentially unchanged from their 1985 state. Information is exchanged on paper or by people speaking to one another. The pace is slow, despite an urgent need for action. The cycle time for asking a question and getting the answer is days, sometimes weeks. People refer to existing knowledge or limited information sources such as printed directories, publications, and personal networks.
Web discovery is a transient stage where each Web innovator sees every business relationship as a nail for his or her particular hammer.  Consequently, the scale of the technology and the scale of the problem being addressed are frequently mismatched: (1) The technology is not robust enough to support an unanticipated level of demand. (2) The solution only addresses the visible parts of the process, leaving cumbersome work to be done the old way. (3) The application is overengineered and unnecessarily complex. In any of these cases, the user is unable to benefit from the solution. What can be done has not yet been tempered by an understanding of what should be done.
As an example, consider the disappointing performance of electronic books, compared with the unbelievable popularity of electronically enabled sales of physical books. The people who predicted the demise of the printed book were tripped up by their focus on the costs and difficulties of printing, distributing, and selling books. They completely missed the nuances of our emotional attachments to books. I would guess that you still remember the first book that was your very own, and maybe some special rules for handling certain books. Books are profoundly tangible things. If a buyer pays a fee and downloads a book onto her computer, what has she bought? Not an actual book, which could then be treasured in her library, exchanged with a friend or at a used book store, but merely the right to read the book! Thus a book is transformed from a thing of value to an episode of entertainment. When e-books are marketed as entertainment, as an alternative to renting a video tape or seeing a show, then the technology will be aimed in the right direction.
According to Michael Dell, the middleman closes gaps between buyer and seller.  These gaps may be of knowledge or financing; of place, connections, or speed; or of coverage. Dell Computers exploits the fact that the Web can close the knowledge gap. On the Web, the cost of the next one-to-one user contact is zero. Without the Web, unknowing buyers and unknowledgeable sellers must rely on informed intermediaries. The Internet fundamentally changes the way companies do business through its ability to enable people to conduct low-cost, one-to-one customer interactions with very rich content.
Further into the Web discovery stage, entrepreneurial practitioners of a pre-Web process recognize the power of applying a Web technology, and tune it for their particular niche. These innovators are better equipped than the pure technologist to create a match between need and solution, because they are better informed about the whole environment. However, what was a parameter in pre-Web days may have no relevance once the Web is applied. Even seasoned professionals in an industry can create a process that simply doesn't work. For example, a manufacturer may devise an elegant approach to presenting appropriate product pricing to a visitor to a Web site. Based on the home address of the visitor and the relevant country-specific distribution and pricing agreements, a price is displayed. The manufacturer swiftly learns that artificial price differentials are not sustainable in a Web environment where everyone can shop everywhere. As a company adopts electronic commerce, the makeup of its organization will have to change. Information intermediaries can expect that information technology costs may be 40 percent of revenues at maturity (even higher to begin with) and information technology personnel may be 50 percent of the total staff.
Post-Web processes are marked by simplicity, robust economics, and startling effectiveness. Customers or users can serve themselves information and action from the same systems that internal people use. I have not yet seen a process that I would categorize as post-Web. The system configurators on computer-selling Web sites come closest. Other apparently mature operations, such as Internet Service Providers (ISPs), and the use of credit cards as the standard payment method have not achieved simplicity, robustness, or startling effectiveness.
ISPs as we know them are independent entrepreneurs growing into mid-size businesses by leveraging sweat equity and relatively little competition in smaller markets, or huge corporations leveraging existing connectivity. The selection, implementation, and maintenance of the technology that supports instant information flow is not trivial. The cost of this technology will be borne by some combination of public financing and self-funding, somehow resolving the question of how self-funded assets will be shared with other self-funders. The best way we have found to build infrastructure is through public financing, relying on relatively long payoffs at lower-than-market rates, for the common public good. I think the post-Web ISP will be a public utility.
Credit cards as the standard payment method for Web transactions is a typically U.S. assumption. Since we each have two or three credit cards for each debit card in our wallet, we think everyone else runs their personal finances the same way. But the rest of the world predominately relies on debit cards or stored value cards for convenient cashless payments.  The post-Web payment method will be cashless, of course, but not a credit transaction, as the costs of the credit card infrastructure will become significant and prohibitive, relative to those of the debit or stored value infrastructure.
Post-Web processes will probably not be some totally new way of trade. Trade is, at its core, a human endeavor, and we have had thousands of years to express it and evolve it. The challenge we face in building post-Web processes lies in seeing which parts of the existing process are fundamental and which are only artifacts of the available technology.
In general, a middleman succeeds by lowering the barriers to action, increasing the number of willing buyers, increasing the number of willing sellers, or some combination of these three actions.
Varda Lief, of Forrester Research, identifies three types of intermediaries that will merge, over time, to become infomediaries (a term originally coined by John Hagel of McKinsey). 
Aggregators primarily serve buyers in fragmented markets, providing updated prices and a single point of contact for purchase from many suppliers. The sale of photographic images is an example of an aggregator opportunity.
Auctioneers primarily serve sellers with perishable or surplus goods and services, and bargain-hunting buyers with low risk aversion. The sale of limited appeal goods at auction is an example of the auctioneer opportunity.
Exchanges create liquidity in otherwise fragmented markets, lowering the average stock levels by matching bid/ask prices, and serving as a neutral third party enforcing market rules and settlement terms. A holding company that serves formerly competing metal parts fabricators in the aerospace industry is an example of an exchange.
Forrester's Lief does not mention a fourth type of infomediary, which I call the knowledge broker. Extracting and re-deploying successful business practices is an example of a knowledge broker.
Each of these examples is described in detail below.
The pre-Web business process for the sale of photographic images included completing studio work or field expeditions, turning the film over to a processing house, reviewing contact prints to select images, turning the marked-up contact prints over to a printing operation for enlargements, then judging the final work and (frequently) repeating the markup and printing process. To market the image, the artist made a transparency from the final print and sent it to a stock photography house. The stock house would decide whether or not to add it to their collection, and would pay for the image's use if and when an end client elected to use it. The artist lost control of the image and its final form very early in this process.
Scott Bourne is a Tacoma, Washington, photographer who is exploiting the Web to create a worldwide enterprise. He has vertically integrated all post-processing operations, eliminating the printing and distributing middlemen. The processing house now delivers transparencies and digital files obtained by scanning them. (Digital cameras cannot yet capture the range of contrast and light required for the best photographic images.) The artist crops and manipulates the image on the computer. A compressed file is posted to the Web site for discovery and review by potential customers, and final prints are printed on demand by the artist using archival paper and inks.
With all this capability readily available to the individual artist, it may appear that there is no opportunity for a new middleman. But the post-Web version will take into account the fundamental nature of creating art: the fact that most artists have no interest at all in anything other than the creation of their art work. They don't want to invest in computer equipment; they want to buy new brushes or lenses or clay, or go to Costa Rica to see the beauty of the rain forest. They don't have time to learn a computer program or resolve incompatible printer drivers or build Web site; they have painting, sculpture, music, photography, weaving, and creating to do. It drives them, and it drives an opportunity for intermediaries.
This reluctance on the part of the seller to engage in anything other than the creative act has created an existing middleman structure that is highly artificial and highly intermediated. The "high art" world is full of mystery and charisma; gallery owners and curators entice wealthy status-seeking patrons to acquire the work of the latest and greatest "discovery." Personality and opportunism, not talent or vision, are what makes an artist successful in this marketplace. Economically, the whole structure is terrifically inefficient. The post-Web business process will rely on middlemen who are trusted by the buying public, provide the cachet of the "high art" market, easily deliver services to nontechnical users, and offer better returns and distribution to the artist. By creating a sales channel that bypasses the existing structure of agencies, gallery owners, and artists' agents, this new middleman will aggregate both supply and demand.
My experience as both a buyer and seller on eBay, as well as discussions with numerous participants, leads me to assert that a trusted third party can increase the number of willing buyers and the number of willing sellers. Gear.com and the Defense Reutilization and Marketing Service (DRMS) are other examples. 
With respect to the consumer-to-consumer, one-of-a-kind transactions that eBay supports, the old business processes were operated through pawn shops, garage sales, second-hand dealers, classified ads, donations to a charity, or even disposal in the trash. These old processes leave the seller convinced that he or she
eBay is a public marketplace with fair rules evenly and swiftly enforced, where buyers and sellers have similar rights and advantages. eBay allows both parties to see if similar items have sold recently, and for what price. The seller sets his or her lowest acceptable price and learns if anyone wants to buy and at what price. Potential bidders can ask questions about the seller, the item, and the terms before making an offer. Each party can see brief comments about the other, based on prior eBay transactions.
In my opinion, eBay's greatest contributions are the way it serves fragmented microniches of trade and how it allows us to convert sunk costs into available cash. The market for classic acoustic guitars is an example of such a microniche. The second-hand market is small, usually full of guitars from large manufacturers, and served by a network of dealers. Prices of guitars in good condition are typically in the $400 to $600 range, not much money for a powerful emblem of your youth. Pre-Web. eBay has connected passionate aficionados, and the going price for some of these guitars is now over $1,500.
Consider the set of all owners of a particular item. Some use the item, some are simply storing it, some wish they could dispose of it, and others are actively selling it. The set of all nonowners of a particular item is composed of those with no interest in it, those who are unaware of its availability, those who dream of owning it, those who are on the lookout for it, and those who are actively seeking to buy it. By increasing the visibility of willing buyers, nonusing owners can be convinced to become sellers. By increasing the accessibility of sellers, interested nonowners can be convinced to become buyers. This marketplace creates additional discretionary spending because it presents a real opportunity to recover prior expenditures. Thus auctioneers, easily accessible from your home, increase the turnover of goods and cash.
Is eBay a post-Web business process? I think it is coming close. What is still missing is a Web-based payment method, perhaps tapping into the telephone billing system for both buyers and sellers, to eliminate the requirement for trips to the post office or bank for money orders.
Aerospace manufacturers rely on many small specialized manufacturers for high-quality and long-lead-time precision machined metal parts. In the pre-Web business process, the aerospace company's buyers know the parts requirements and the qualified suppliers, and place orders as needed by phone, fax, or letter. The seller receives the order, checks the available capacity, schedules the work, and notifies the buyer when to expect delivery. The machining company then builds and ships the parts, and issues an invoice. The aerospace company receives the parts and pays the invoice. Machining capacity is limited by the capacity of the shops' specialized equipment but even more by the limited supply of skilled machinists. Many of the highest quality machining operations are two- or three-person shops. The paperwork of an order, with its technical content (specifications, materials, design features) takes up the owner's time, which is really better utilized in setting up and operating a machine.
During the early Web discovery stage in this industry, there was great interest in electronic marketplaces. The machining company registered its specific capabilities: number and type of machines, quality certifications, types of materials usually used, total production capacity, and so forth. The buyer posted requirements and technical data on specific parts. The machining companies then could negotiate combinations of capacity with each other and quote prices and delivery to the aerospace company. The architects of this marketplace expected that the free exchange of information would result in better capacity utilization, quicker fulfillment of requirements, and lower costs of procurements. Although over 150 companies registered, and several bid opportunities were posted, the marketplace never met expectations. The reason? The owner-operators of the machine shops simply were not comfortable using the Web to mediate deals. They used phones, faxes, even email, but never adopted anything other than one-to-one communications to strike a deal. Another reason was that the marketplace was not accessible to the buyer's existing purchasing software, making its use a completely additional activity.
Is the situation hopeless? No. At least one effective alternative has emerged that transforms the role and power of the small machine shop in the aerospace supply chain. Transtar Metals  is a new supplier to aerospace. It is a holding company consisting of three independent metal-working operations. The holding company is the interface to the customers for administrative functions: ordering, scheduling, bidding, electronic data interchange, billing, and electronic funds transfer. The holding company's staff, experienced in the aerospace industry, are responsible for the information systems' compatibility with customers' ordering systems, and negotiate any new electronic commerce requirements on behalf of the shops. The shop owners can devote their attention to the technical aspects of their business, which is where they excel.
Transtar Metals is acting as a type of exchange, more fully utilizing and creating more liquidity in scarce precision machining capacity. Is this a post-Web process? Quite possibly. Although the scale is somewhat limited at present, it seems relatively straightforward to increase the number of participating shops and customers.
An HP advertisement  advises companies that operational processes may be a revenue-generating asset. Specifically, there is value in the information gained or methods developed in the course of doing core business activities. For example, billing methods, inventory management techniques, or manufacturing processes may be just what another company wants to outsource or replicate. A company's knowledge of raw materials, their suppliers, and the nature of the raw material market may be critical information for investors in that market. Information about clients' needs, which the company itself is not positioned to address, and access to client decision makers may be just what a supplier of a complementary (or even an unrelated) product needs to be successful.
Jack Stack, president of Springfield ReManufacturing Corporation (SRC), has created high-margin services that pay for needed equipment. The remanufacture of intracity truck engines is a precise, low-margin, high-volume business. Mr. Stack transformed SRC from an inefficient division of a poorly performing conglomerate to a thriving, enduring business by implementing Open Book Management.  Mr. Stack does not believe manufacturing will ever again command the stock price multiples or respect that dot.com companies currently enjoy. He recommends instead that companies provide value-adding services to their suppliers and customers, and even to businesses in completely different niches. His recommendation is backed by SRC's experience. SRC has begun a bank; it conducts in-plant workshops on Open Book Management; and it earns fees and royalties from Mr. Stack's lectures, consulting engagements, and writing.  Mr. Stack is a pioneer. Other pioneers will try to emulate his success; most will succeed only when a facilitating middleman makes it easy to recognize the value embedded in their operating processes, package it for sale, and deliver it to interested parties. This middleman would be a knowledge broker.
In conclusion, I suggest six essential attributes for enduring success as a post-Web middleman:
By demonstrating these attributes, value-adding middlemen will not just continue to exist, they will thrive.