Phil Salin and AMIX

Phil Salin and AMIX – American Information Exchange

PHIL SALIN AND AMIX Phil Salin, who wrote the essay on costs and information in our last issue, died earlier this month of complications from liver cancer. This was a personal loss to his wife Gayle Pergamit and many friends; it was also a loss to the world at large, as the works he left behind indicate. The most significant of these is the American Information Exchange, which is starting commercial operations this January in the wake of his death. We extend sympathy to those he left behind, but we’d like to honor his contributions by discussing them for their intrinsic merit… American Information Exchange, started in 1984, is about to launch its service with the support of its parent Autodesk (see Release 1.0, 7-89). The basic idea is an information exchange, which allows people both to offer and to bid for specific information. The information can be either ready-to-use documents, software or data, or custom information services provided on request to a customer’s spec. Like most markets (listen up, Soviet units!), the result will be not just better, more efficient distribution of existing information products, but the creation of a feedback system, whereby more information is produced and tailored to users’ needs. Buyers and sellers can negotiate over contents, level of detail, prices, terms and conditions: exclusivity, timing, reuse and other issues. Buyers can also assess the qualifications of a supplier by inspecting authenticated references from previous customers or any other information the vendor cares to supply for free – including free samples, or the first few pages or a summary of a more detailed report. (Other sellers might provide a fee-based evaluation service.)

As the middleman, AMIX keeps order, maintains the market and arbitrates disputes. And of course, it gets about 30 percent on average of the revenues. Hovever, it makes no prohibitions against private sales, such as additional business or long-term subscriptions, once buyer and seller establish a relationship. It is not attempting to gain a monopoly on transactions among its customers, but simply to earn money for facilitating transactions that might not otherwise occur.

The point is to make more information more accessible on a semi-random basis, among buyers and sellers who do not know each other. In short, the credo is to reduce the transaction costs for information, as described in Salin’s essay in our November issue. How it works

The service begins this month with several markets: pc industry trends; technical information on NetWare, MS Windows and desktop publishing and multimedia; and executable software for sale (with license restrictions as defined by sellers), including full applications and Smalltalk, C, C++ and Microsoft Windows modules and components. The choice of subjects reflects the fact that people who want computer-oriented market information are the most likely both to be online in the first place, and to feel comfortable with a service such as AMIX’s. Each information “market” has a market manager, someone whose job is to find and enroll sellers and buyers, make sure that the information offered matches the market profile, and handle complaints and problems of market participants. If a market overflows with information, the market manager decides how to subdivide it, or into which market to redirect participants. Would-be buyers pay $50 to join for six months and $5 per month thereafter, plus $3 to $9 per hour for connect time (including long-distance costs), billed to a credit card. Buyers get the requisite client software. Sellers pay to AMIX 40 percent of the first $50 of each transaction down to 10 percent of any amount over $500. They also pay storage fees for posting their wares or descriptions of them (which discourages them from cluttering the system with items that are mispriced or inappropriate). The information purchased will generally be delivered through the system, but it could also be delivered by mail, Federal Express, or in an agreed-upon consulting session by phone or in person. The seller can handle simple sales automatically; for example, anyone who wants to pay the price can automatically down-load a given report at a given price. Not just another text filter Although it does address the problem of information overload, AMIX is not just another automatic text filtering tool that lets you select certain items by category or keywords (although certainly you could program an agent to do so). The system isn’t a database (or even a document server) to be searched; it’s an electronic intermediary between living, conscious people or firms. Juan can leave a message for Alice to negotiate a fee, or if he wishes, he can pick up the telephone. AMIX’s role is to facilitate transactions, not to interpose itself. Beyond this, two features make all the difference: Each item has a specific price, which may be negotiated; you’re not paying a blanket fee for a bundle of information you want only a portion of. Second, some items don’t yet exist. The seller is offering a capability, and will provide the information or service on request for a fee. Thus AMIX will have an impact not just on who gets the information that is produced, but on what gets produced. And, if you believe in the efficiency of markets, more of precisely the right stuff will get produced, and the low-quality redundant stuff will diminish. By and large, most of our information purchases are through long-term relationships. It’s as if we had all been saying up to now, Well, I guess I’ll wear whatever Ann Taylor or Levi brings out this month.” Imagine the state of the apparel industry – and the variety of people’s clothing. Now, in effect, you can pick and choose from all the suppliers, and get someone to custom-make your clothes. In other words, consumer feedback becomes more fine-grained: This information is interesting; that is not,” instead of “Well, this service or periodical has more of what I want than anyone else.” Our own vested interest Our personal interest in AMIX is threefold: as a user, to find information we need (without rummaging through old PC Weeks, calling up IDC or asking an industry friend); as a vendor, to find a broader market for the information we produce; and as an observer, watching the fruition of the information marketplace of the next decade. As a vendor, we distribute this newsletter to a subscriber base of approximately 1300. Most of them probably read one or two issues thoroughly, and pass some on to others who might be interested. And of course people make copies – a practice we acknowledge but do not permit – either on a regular or irregular basis. At the moment the full text of the newsletter is available – with several months’ delay – through Ziff-Davis’s Information Access online service and its Computer Select CD-ROM. So far, that service has meant little to us; there’s no feedback loop. We receive an annual fee from Ziff-Davis, but it’s related to Ziff’s overall on-line and CD-ROM revenues, with no relationship to how much people read our stuff as opposed to anyone else’s, let alone which particular items they might read. From another angle, however, we know of only two cases where people might have decided not to subscribe because they figured they could get the information electronically (and many more where people don’t subscribe because they get it “from a friend,” or an investor, or whatever). In general, however, we assume we get exposure this way that leads people to subscribe. Part of the purpose of a newsletter or any strong editorial product is not just the information but the selection of the information: “Hey, this is something you should know about, even though you might not be looking for it.” But the audience that wants everything we write is relatively small certainly compared to the audience, perhaps ten or even a hundred times larger, that wants one tenth or one hundredth of what we produce. Those people, moreover, are probably willing to pay more than one tenth or one hundredth of our subscription fee of $495 per year for those particular articles. If only they could find us – or we could find them and sell a subset of what we write. With systems such as AMIX, this will happen. However, it will be a while before we overcome the closed-market problem: How do we reach the people not on AMIX – 100 million pc users minus a few hundred or thousand at best in the first few years of AMIX’s life? Even assuming that the potential market is much smaller, just the people in the computer business, leaves a lot of the potential market unserved. That’s why we need standards such as Wide-Area Information Server (see Release 1.0, 4-91) and broad interconnection of all these systems. Markets and relationships From AMIX’s point of view, what of the danger that market participants will trade outside the market? (This is an issue for other markets, of course, such as the New York Stock Exchange, which is seeing its business going to third-market brokers and foreign markets.) First, AMIX properly sees its role as facilitating transactions among (relative) strangers and lowering transaction costs for searches, not as a way to tax long-term relationships between sellers and buyers. In the end, the buyer comes to the market to find the information; the seller, who pays the fees to AMIX (although of course it’s built into prices) comes to the market to find buyers he can’t find directly. Certainly there will be transactions conducted off-market, but that is business that doesn’t belong at AMIX anyway. Sellers who habitually abuse the system – offering items for sale and making side deals will eventually be delisted from the system, since AMIX can monitor activity, and wants only sellers who are active on its system. The philosophy behind AMIX is that there’s enough real need for its services – facilitating transactions – that it has no need to intrude on resulting relationships. The real relationship it wants is between itself and its buyers and sellers who come to its market to meet each other without the costs and inflexibilities that come from continuing relationships (such as subsidizing the production of information they don’t want). Thus, AMIX’s initial seller base will be not so much the Gartner Groups and IDCs, although they may ultimately use it for incremental revenues, but the free-lancers, specialists and newsletter-writing oddballs who need a better way to find their potential markets. And it will certainly enable some people to leave their day jobs at large research organizations or other kinds of businesses for more fulfilling free-lance work doing only the assignments they want – or electing to do those of high value to customers because the price is right. See you on the net!

COPYRIGHT 1991 EDventure Holdings, Inc.

COPYRIGHT 2004 Gale Group