A B A C U S    Issue 2.06 - June 1996

No Network is an Island

By John Browning



If you think the Internet is growing fast - and it is - intranets will take your breath away. Intranets are little islands of Internet technology created for a company's private, internal use. Having seen how easy the Web makes sharing information - be it text, video, pictures or whatever - executives want the same technology for themselves. It's quite possible that all of the Western world's biggest corporations will have equipped themselves with intranets before the century is out. By 1998, according to Zona Research in California, the corporate world will be spending about US$8 billion a year on intranets, more or less four times as much as the total investment expected on the Internet. The excitement is palpable - and deluded.

Intranets are indeed important, and their growth is one of the most significant changes going on in the corporate world. But their significance lies in the fact that they will end up doing almost the exact opposite of what the people paying for them think they're buying.

The more simplistic attractions of intranets, first spotted by computer companies, are increasingly clear to all. The technology underlying the World Wide Web was designed to provide easy-to-use interfaces to all kinds of data, on all kinds of computer systems. In today's fast-moving markets, sharing information quickly and flexibly is the key to competitive success. So companies are seizing upon intranets as the answer to their fondest prayers.Telecoms giant AT&T, for example, publishes its internal directory of information about its 300,000 employees on its intranet - who's where, and doing what. Drug-manufacturer Eli Lilly uses its intranet to make instantly and ubiquitously available all of the complex regulations governing the trials of new drugs in over 120 different countries - so that its R&D managers can have all the facts to hand as they coordinate efforts to test their innovations. Publishing and distributing such information on paper would be expensive, and the data would change so quickly as to make accuracy nigh-impossible.

Because intranets borrow so much infrastructure from the Internet, and because they employ $20-a-user Web browsers such as Netscape as interfaces instead of $150-a-user proprietary groupware systems like Lotus Notes, intranets are cheap to set up. Lilly put the first 3,000 users onto its intranet for just $80,000, and it is now expanding the system to embrace 13,000 more people around the world.

As they install intranets, companies are rediscovering the joys of computers. These machines really do leap over barriers of space and time to improve communications. Most intranets contain some facility for threaded discussions - that is, email dialogues on specific topics, like private Usenet newsgroups. They can make a global corporation feel like a small, cooperative office, sharing ideas through electronic discussions and swapping solutions to common problems across thousands of miles. If the charm of the Internet is that it connects people with little in common, the practical attraction of the intranet is that it connects people with lots to discuss.

Best of all, as far as the executives who approve purchasing are concerned, intranets are controllable. While they allow information to flow faster and more freely within a company - so creating a more energised and effective work force - the "firewalls" of software that surround them do not allow precious ideas to leak out, or dangerous ones to leak in. In short, intranets seem to offer high-tech change without too much risk of digital revolution.

But this offer is illusory, the basic intranet error. The Internet is, and always has been, a network of networks. While some of its component parts have indeed been lawless, others are among the most tightly controlled networks in the world - like Milnet, the American military's computer network. So when companies set up intranets, they are not, as most now seem to believe, somehow setting themselves apart from the Internet. They are instead rushing to embrace it, with only the most arbitrary protections to ensure that the embrace remains chaste. Pretty soon, that protection will prove ineffectual.

Although executives haven't realised it yet, the standardised Internet technologies in their intranets mean that the passage of information from one company to another is made much easier. With everything in the same formats and protocols, all that is needed for information to flow across the boundaries between companies is a sound business reason. And those are going to spring up in abundance. The boundaries which have set companies apart from one another, and apart from their customers, will dissolve. Intranets will not preserve companies' integrity - they will blow companies apart.

The Coase is clear

To understand why, ask yourself what the point of a company is. Better still, ask Ronald Coase, whose answer won a Nobel Prize in 1991. Individuals, Coase reckoned, could in theory buy whatever resources they needed to be productive by bargaining among themselves in a free market. Entrepreneurs would be able to hire the most appropriate people or machines for whatever task lay ahead that day or that week. But the complexity of all this shopping would be overwhelming - and there would always be the risk that some crucial resource would be temporarily unavailable. Companies, argued Coase, exist to save managers from the paralysis of choice, and to ensure a pool of resources and people whenever one is needed. Companies are inefficient; not every resource will be fully employed all of the time. But the cost of the inefficiencies is far less than the costs and risks of gathering those resources from outside the firm.

The Internet, however, changes completely the costs and the practicalities of reaching beyond the boundaries of the firm. Want to make your latest widget on a state-of-the-art production facility without actually building one ? With the Net, you can simply transmit your designs over the lines to somebody else's production line. There are already "fabless" chip makers that do just that. From tax advice to service with a smile (or at least a smiley, given today's bandwidth constraints), the Internet allows managers to pick and choose more easily than ever before. The more corporate information is held in intranets, the easier this will become - not because people will break down each other's firewalls, but because it will make good business sense for them to invite each other inside. After all, the firewalls are arbitrary, reconfigurable with a few lines of code - and a good idea.

Consider a couple of examples. McDonnell Douglas now sends out thousands of 20 to 25-page technical service bulletins each year to the 200 airlines that fly its planes. Each bulletin provides new information about how best to maintain the aircraft, and so it is important that they be received quickly and filed accurately. But the costs and complications of shipping and storing four million pieces of paper a year are huge. So McDonnell Douglas has instead begun distributing the pages via its intranet, in a form in which customers can file them electronically, and mechanics can print them for use in the service hangar.

So far so obvious. But Visa provides a more intriguing case. Visa processes credit card transactions for 19,000 member banks. Its intranet was originally created in part to replace the eight-inch-thick directory that helped Visa employees keep track of the appropriate contact at each bank. But now the banks want to turn the technology around. They want to use the intranet to gain access to Visa's own computers to check on the status of various accounts and transactions. Why, the banks wonder, should they have to phone a Visa employee, just to get him to tap a few keystrokes into the Visa computer, when they could do the same tapping faster and more easily for themselves ? Why, in short, should they be outside Visa's intranet ?

That question - or its equivalent for other companies and other industries - is a huge one, probably the toughest that executives will have to face over the coming decade. How permeable should which membranes be ? Get it wrong one way and the lack of information will turn customers off; get it wrong the other way and you will lose all purpose. To get it right, the executives will need to ask two basic questions. What sort of information is easily passed ? And what sorts of relationships matter ?

One type of information that can travel is price. In the twinkling of an eye, a computer can search the whole of the networked world for the lowest price on any given product. Management consultants at Arthur Andersen have built an "agent" to demonstrate just how effective such searches can be, and made it available at their Web site ( www.anderson.com). It will find, among other things, the lowest price on the Internet for any given CD. As Andersen cheerfully admits, electronic bargain hunting can rapidly drive profit margins to zero. People can choose to keep the agent from learning their prices - but how many of us are likely to think that the best hidden price is also the lowest ?

Such electronic bargain hunting is already having a global impact. At the end of March, Michael Fuchs, head of the German exporters' association, complained that it was cutting into the growth of German exports. Particularly in niche markets, he said, the Internet enabled customers to shop around for four or five bids - whereas even a few years ago it was cheaper and easier just to accept whatever price was offered by the established supplier. This price shopping has enabled new entrants from the developing world to gain business at the expense of German companies.

At first glance, you'd expect every company to be trying to push its suppliers into the electronic realm in order to price hunt more aggressively. But in some industries, at least, exactly the opposite seems to be happening. The motor industry is one of the world's most automated, and the big car makers have long linked their computer systems with those of their suppliers in order to facilitate just-in-time delivery of parts for assembly, among other things. But with increasing automation has also come a general reduction in the number of suppliers. Instead of bargain hunters, car makers have become relationship-builders, settling down with a few key suppliers for long-term relationships.

The explanation for this seemingly contrary behaviour lies in a deeper look at the sorts of information that travels across computer networks. Computers are great at communicating facts about today's products, but they are bad at communicating the complex dialogue of instincts, hypotheses and prejudices from which tomorrow's products are born. As cars and the just-in-time processes of their production have become more complex, auto makers have formed longer-term relationships with suppliers in order to ensure that research is coordinated for mutual advantage. And this is a process in which computer networks play a surprisingly small role.

Academic sociologists studying computer-mediated communication (see, for example, Lee Sproull and Sara Kiesler's now-classic book Connections) have consistently found that computers encourage people to say what they know. Electronic discussions tend to be more vigorous and more egalitarian than those held face-to-face. But they also have a harder time reaching consensus - precisely because everybody is so busy stating their own opinion.

This is the sort of problem that managers, not technologies, have to solve. The first principle they should bear in mind is a simple one: new media complement - but don't replace - old media. Computers speed the transmission of well-structured answers to well-formed questions; face-to-face communication creates new questions. Both have crucial roles to play in day-to-day business competition. Indeed, companies that get the most out of electronic communication are typically those that have increased travel budgets in line with networking ones. Face-to-face meetings build consensus that electronic communication then acts upon.

It all comes back to relationships. In an odd way, the more information gets out - the more it is forced out by competition, by the demands of partners, customers and clients - the less the information itself matters. What matters is setting the context - determining the way the relationship is formed, the values upon which it is built. Take selling CDs. Though price competition may get tough, relationship competition will still be possible and profitable: our CDs may be more expensive, but we'll order anything you want, and we'll recommend new releases that might match your tastes. Those who don't form such relationships will forever find themselves judged by the criteria that others set for them.

Whatever happens, companies have to put more and more information out in order to compete, and take more and more information in to know what's going on. The Internet, with intranet nodes everywhere, makes that possible. Improved internal communications become improved external communications, too - except for those poor fools still convinced that their intranet is a private place, apart from the rest of the networking world. Ultimately, there is no real alternative to the Internet. The business world will be divided between those who realise it quickly enough to learn to live with it, and those who just die.

John Browning (jb@wired.co.uk) is executive editor of Wired.

In less than a year, telecoms giant AT&T has linked about 60,000 of its 300,000 employees to its intranet. The first thing AT&T put on to its intranet was its internal telephone directory. In keeping with the theme of simplifying the mundane, AT&T has also put online its internal ordering systems and library services. No need to revert to memos to order a floppy disk, or to trudge down the hall to get a technical report. It's all on the intranet.

But where AT&T's efforts get most intriguing is in its use of the intranet as a sort of virtual team room. Groups from far-flung parts of the company have to work together at short notice more and more often, given the pace of change in telecoms markets, plus AT&T's own plans to break itself into four parts. One of these groups' biggest head- aches is finding a way to transmit both their own ideas and their computerised files and records. By translating into the lingua franca of Internet technical standards, AT&T's intranet provides a common ground. Soon, AT&T hopes to build into the intranet a sort of internal electronic commerce system that allows them to track which department gets billed for what expenses in cooperative projects. That done, AT&T executives reckon they will be just about ready to start inviting customers into their virtual meeting rooms - for real electronic commerce.


For a glimpse of the ultimate fate of most intranets, look no further than www.fedex.com, Federal Express's Web site. The Web site turns FedEx inside out. By typing the number of a FedEx shipment into a form, any customer can find out exactly where that shipment is, right now, using FedEx's computerised logistics systems.

Historically, FedEx's logistics systems have been the key to its success. By making them directly visible to customers, FedEx can both more forcibly demonstrate the systems' superiority when things go right, and win some measure of understanding when things go wrong. Customers effectively start to view the world from FedEx's point of view.

As other companies come to copy FedEx's strategy, the fundamental nature of marketing changes. Communications become direct. Instead of advertising to convince consumers of the benefits of, say, a whizzo new logistics system, companies can demonstrate it. And instead of spending on market research to deduce what consumers want from their behaviour patterns, companies can simply ask them. Cathay Pacific, for example, regularly auctions off air tickets for frequent-flier points on the Web. So it gains a measure of whether its frequent-flier programme is too generous - or too miserly.