Web Usage Explosion By Charles H. Ferguson You think the Web is hot? You ain't seen nothin' yet. The Web is much more than an office distraction; it's a whole new software standard that makes the current software industry obsolete. In this article, consultant and entrepreneur Charles Ferguson argues that the Web provides a rich new landscape in which to plant entrepreneurial seeds, and that it poses the first real challenge to Microsoft's worldwide dominance. Suddenly, software has become interesting again, and with a vengeance. The World Wide Web has blossomed into the biggest revolution in information technology since the development of the computer, or perhaps even the telephone. Just 18 months ago, there were perhaps 5,000 Web services (aka "home pages") worldwide. Now there are over half a million, with about 5,000 new services being created every day. It's the fastest-growing technology in economic history. And that's just the beginning. Over the next five years, the Web will grow by at least another factor of 100. This implies tens of millions of new Web services annually. There will be, perhaps, 20 million Web authors, editors and publishers, and, potentially, 200 million Web users. Thus, while the frenzy over the Web has caused some excesses, the frenzy is generally merited. Despite all the publicity the Web receives, some of its most powerful implications are not yet widely appreciated. We are witnessing the stunningly rapid establishment of a new, open-architecture, global information infrastructure. This new infrastructure (based on the TCP/IP protocol, and the Web architecture) will not be limited to the public network that defines the World Wide Web today. It will also work for personal archives, CD-ROMs and and a variety of embedded, free, paid, open, restricted, private and corporate services. This new architecture arrives at an opportune time. Several current information technology markets are badly served by a lot of closed, incompatible and obsolete proprietary systems. Just as there was a large, pent-up demand for personal computers when they came on the scene to displace centralized mainframes, there is now a pent-up demand for a distributed, open-systems solution. Therefore, this new architecture is going to replace many inferior, proprietary architectures, such as Notes from Cambridge, Mass.-based Lotus Development Corp., and traditional on-line services. It will build a large new industry in their place. In a very real sense, the Web will drive the client-server reengineering revolution. This will be an industrial revolution whose scope makes the personal computer industry of the early 1980s seem like a small family business by comparison. It includes both full, Web-based information services (both public and private Web sites), as well as "webs," which are subservices or documents within those Web sites. Web services will become as universal as telephones, fax machines and e-mail. Basic Web capability is already becoming standard system software. People use browsers on the desktop as search and display engines, and Web servers as standard document servers or as personal archiving systems. Now, end-user Web development tools are available. Moreover, within 18 months, the infrastructure will be in place for full electronic commerce, directory services and navigation systems. That makes the Web a Very Big Thing, guaranteed to become the core of the global information infrastructure for the next several decades. And, if we do a few more things right, it can become Everything. Web infrastructure would be as interoperable as the world's telephone systems, and webs could then be easily created, universal, interoperable, transportable documents--in short, the basic units of the information economy. We would have a common, scalable architecture underlying all personal, corporate and public information systems. And, last but not least, we could have a software industry that is not controlled by Microsoft Corp. Of course, we're not there yet. The Web is still a 1990s architecture laboring under a 1970s infrastructure. That was tolerable when the number of Web sites developed each year was small, but it breaks when we add 10 million services per year. For the Web to continue to grow, its software model must become something more akin to shrinkwrapped, graphical PC software. Fortunately, that change is already underway as billions of investment and revenue dollars pour in, creating commercial-quality user interfaces, security, tools, applications and services. What follows is a survey of the Web revolution --------------------------------------------------------------------------- on-line/off-line Web-based systems will unify and replace conventional on-line services, workgroup systems such as Lotus Notes, personal archiving systems and embedded information systems. Conventional on-line services alone are a $15 billion industry, including both the business-to-business services, such as Bloomberg, Lexis/Nexis and Dialog Information Services,and the consumer services, such as America Online and CompuServe. However, the technologies and business models of the conventional on-line services are obsolete. They are closed and proprietary, and cannot operate with each other or the world at virtually any level. There are no cross-vendor, on-line directories, archives or search facilities. Every service uses its own mutually incompatible browser, user interface, data formats and development tools, most of which are hopelessly obsolete. Some, like Bloomberg, even require proprietary hardware. All the services use centralized architectures (most use mainframes; CompuServe uses PDP-10s; and America Online uses fault-tolerant minicomputers), with an order of magnitude poorer cost/performance ratio than modern distributed systems. Their server software is not easily scalable or portable, so you can't move content or allocate services across different computer systems. They cannot easily access real-time or local data generated and stored by their content vendors or customers. That means they cannot be competitive for electronic commerce, where you need to know if someone's green widgets are in stock today in the downtown store. Content providers can't design, manage or cutomize the user interfaces or content in the services. And the on-line services often refuse to supply detailed customer information, or they charge high prices for it. The on-line services have prospered only because, when computing was centralized and complex, it was costly and difficult for new services to get started. Thus, despite the fact that everything was wrong with them, the established on-line services vendors could charge royalties of up to 80 percent on the content they carried. Now, Web services will destroy those feudal systems. The Web enables anyone to create his or her own service, offering the potential for millions of electronic publications and information services, and allowing their creators to keep 100 percent of their revenues. Such services can either be self-operated or operated by outside hosting services. Web services are evolving as a layered, open, interoperable industry. Any service can be linked to any other service. Independent vendors offer browsers that can be used to view any service on almost every major computing platform. Servers also run on most computers, and a few can support a service accessed a million times a month (more than enough for your average florist or graphic design firm!). The response of the proprietary services thus far has been a mixture of resistance, panic and accommodation. The smart on-line services vendors are converting to Web-based architectures. The on-line services companies not moving to the Web should be running scared. Following a $100 million acquisition spree, Vienna, Va.-based America Online Inc. has begun offering Web hosting services and software products alongside its proprietary services. Microsoft is repositioning the Microsoft Network to provide greater access to the Web. Still, the problem these companies face is that the market for on-line services logically consists primarily of potential competitors and frustrated content providers. Even with Microsoft's enormous power, the Microsoft Network, in its present form, will not dominate on-line services because it cannot destroy the Web. --------------------------------------------------------------------------- Workgroups and the Death of Notes For many of the same reasons, the next casualties of the Web will be workgroup applications, such as Lotus Notes and Microsoft Exchange. For a few years, Lotus Notes was the only game in town. It was the best way to provide some types of internal corporate services, if only because it could make use of modern server hardware. Notes flourished in the vacuum left by the failure of the proprietary, on-line services to make their server software available to corporate users. But Notes has many limitations, derived from its internal corporate focus. It is too difficult to install, use and develop applications for Notes, a reflection of the fact that it was designed in an era when MIS was a large, centralized function with armies of programmers and support engineers. When Notes was designed, personal computers were primarily corporate capital goods, and information architectures did not need to scale from corporate to public or consumer volumes. Notes databases are limited to 1 Gbyte, and each Notes server can effectively handle only about 100 concurrent users. Moreover, every new user must be registered by a system administrator, and it is difficult for end users to hop from one server to another. This effectively precludes anonymous use, self-registration and casual browsing, all necessary ingredients for large public services. Notes cannot easily access external database systems or real-time data. Consequently, it cannot easily be used for public services or electronic commerce. That will keep it isolated from most mission-critical systems. Now, server hardware is shipping at a rate of about 10 million units per year, exceeding the volume that forced the shrink-wrapping of PC software a decade ago. It is impractical for server applications to require custom installation and professional development and administration. Of course, the Notes support organization isn't standing still in the face of these burdens: it just put up a Web server. The other problem with Notes is perhaps the most fatal of all: It has been managed by Lotus, and now it is managed by IBM. I'm not just being cute; this is a serious issue. Consider: For the $3.5 billion it paid for Lotus, IBM could have purchased virtually the entire Web software industry, including Netscape, Spyglass, Vermeer and NetManage, plus one of the major independent Internet access providers, such as Netcom Online Communication Services, PSI Holdings or UUNet Technologies. Moreover, IBM's server platforms are losing market share rapidly to Unix, Windows NT and NetWare. As a result of IBM's acquisition of Lotus, we will now watch a $70 billion company run by a financial engineer (a very gifted one, clearly) trying to use an obsolete product to compete against both Microsoft and an entrepreneurial wave that is growing by 20 percent per month. The result is a foregone conclusion. IBM will eventually have to switch its focus to the Web, probably once again via acquisition. Ironically, by focusing on their respective proprietary systems, IBM and Microsoft are, for once, in the same position--making the same mistake. Future business workgroup applications will instead be based on Web servers. The Web architecture is very well-suited to personal bulletin boards, discussion groups, workgroup document sharing and other collaborative business applications. In fact, Web-based services are now about evenly divided between public services on the Internet--that is, the World Wide Web itself--and private services on internal networks. The basic architecture of the Web is enormously more flexible than Notes, CompuServe forums or Microsoft Exchange. Web servers and browsers are scalable and portable, and can be used on the public Internet or any private network, embedded service or personal archive on PC hard drives or CD-ROMS. Any web can be linked to any other web, and fully integrated with corporate databases and legacy systems. It will be easy to set up temporary, project-by-project Web services and applications sharing information between vendors and customers. Lotus and AT&T Corp. have applied a few band-aids via Network Notes, but the effort is doomed. In fact, AT&T is already entering the Internet access business and will surely become a large Web-hosting vendor as well. So although Notes is a moderately good 1980s architecture and will probably reach five million users before it declines, it cannot possibly compete. Embedded sytems represent still another market. These applications include on-line documentation, information kiosks, point-of-sale terminals, telemarketing and help-desk systems, and many vertical applications. Just as many accounting packages have an embedded DBMS engine underneath them, invisible to the user, so too will an increasing number of information applications have Web servers, browsers and authoring tools underneath the covers. Finally, the Web scales all the way down to personal services. That has created the newest, fastest-growing category for the Web. Personal services can be both private or public, and are already being used for personal libraries, archives, CD-ROMs and reference services. Some are created for fun by students or hackers; others are developed by professionals for such business purposes as posting documents to workgroups and companies or organizing their personal archives. Having a Web service is now as imperative for cool professionals as was having a cellular phone in 1985. These personal Web services will be of two very different types, both soon ubiquitous. The first is outward-looking personal services, which will permit anyone to post information to the outside world, create discussion groups, refer users to other services, and so forth. The second is inward-looking, allowing everyone to create and manage their own personal archive, library, catalog, encyclopedia, reminder system, subscription service and "window" onto the larger corporate networks and the public webs they use. Think of your Rolodexes, telephone directories and bookshelves. At the same time, another revolution is underway in the use of public systems for personal publishing. This is generating new forms of art, literature, communication and social interaction, from newsletters to cool sites to "zines." In the future, webs will become normal documents, created and transported routinely by end users for everything from academic papers to multimedia presentations. --------------------------------------------------------------------------- Opportunities Large and Small So what opportunity does this portend for new and existing companies? There are several that I predict will provide varying degrees of success. One is Web-hosting services. Although it could become a commodity business, it might be a profitable one, and will probably oscillate for several years between destructive competition and high profits. If hosting becomes both stable and profitable, it will be as a result of increasing barriers to entry. It may require very large computer and communications infrastructures, analogous to those required by long-distance telephony. Another opportunity that is receiving a great deal of attention (and deservedly so) is security, payment and transaction systems. There are a large number of vendors entering this business, from Netscape to MasterCard, from Microsoft to several banks. Interestingly, however, major consumer-oriented banks, such as Citicorp and Bank of America have been very slow to move--possibly at their peril. But these systems will not be easy to construct, and there will probably be many more entrants and an eventual shakeout before the industry stabilizes. The winners should do very well. Browsers and Web server software might be another matter. They will be mostly low-margin, commodity businesses, most likely based largely on nonproprietary standards and bundled free with software, computers and on-line services. This brings us to the question of Mountain View, Calif.-based Netscape Communications Corp. Netscape has attracted enormous publicity, as well as extravagant amounts of cash at stratospheric valuations. But while Netscape is undeniably a very impressive, serious company with excellent people and technology, I'm worried about its business model. The bottom line is that Netscape is a high-cost company, competing (and still losing money) in browsers and servers. It has now built itself to over 300 people, with a very high structural burn rate. Netscape has responded by developing enhancements to the core Web server and browser architectures. But where Netscape has developed value-added functionalities, such as SSL security and HTML enhancements, it has done so in a rather proprietary way, with preferential support for its own products, rather than open, industrywide connectivity. I believe vendor-independent, open architectures should and will prevail in the Web arena. Moreover, it's bad for the industry. As the Web architecture and its tools are enhanced over time for such uses as database access and electronic commerce, four criteria should be met. These are: end-user ease of use; compliance with nonproprietary standards for basic communications protocols and data formats; maximum interoperability; and vendor-independent, platform-independent, open APIs. If the Web's standardization and interoperability deteriorate, both users and vendors will suffer. And in the end, the only large beneficiary would be Microsoft. There are other important opportunities to enhance both browsers and servers. One is putting more intelligence into browsers, to offload processing from servers. Hot Java, from Sun Microsystems Inc. in Mountain View, Calif., is an example of this. Programs written in the Java language, which is rather like C++, can be downloaded from a Web server and executed inside the browser. Netscape has stated that it will support Java in its browsers. The problem is that Sun is commercializing Java in a proprietary way. Netscape has also agreed to support the Acrobat technology from Mountain View, Calif.-based Adobe Systems Inc. for viewing PDF (essentially PostScript) documents. This is less exciting, and possibly a mistake. Acrobat will have a useful niche for high-end graphics but will not otherwise be widely used on the Web, because it is not editable, requires enormous processing power and is visually inflexible on computer displays, due to its printer-derived heritage. Systems integration also represents a large opportunity on the Web. Presently, webs and applications are constructed by small consulting firms using baling wire, pliers and glue. Soon, multinational firms will require global, Web-based systems of great complexity. Currently, there is nobody they can depend upon for that. As with other areas of the Web, it will be interesting to see whether this market, a large one, will be dominated by the likes of Electronic Data Systems Corp. of Plano, Tex., and Andersen Consulting of Chicago, or by startups. The area of development tools is where I'm biased. Until now, webs have been quite hard to develop and manage (although still easier than Notes or services from White Plains, N.Y.-based Prodigy Services Co). It has been necessary to write raw HTML (very nasty) and, for interactive functions such as text searches and discussion groups, to write custom programs in C or in arcane scripting languages (even nastier). Moreover, it has also been very hard to develop services remotely, interactively or collaboratively across a network. There are several reasons for this, including the fact that certain features of hand-coded webs (e.g., access control, image maps and scripts) have not been easily transportable from one system platform or Web server to another. This is a rather absurd situation, given that the Web is all about networked information access. Still, if it becomes possible to fix these problems, everyone will be able to create Web documents and operate Web services. Web services will then easily become as ubiquitous as word processor documents or spreadsheets. In such a world, there will, literally, be tens of millions of new webs developed per year, from sophisticated personal home pages to classified advertising services to dramatically different electronic publishing. Imagine, for example, a journalist creating an article as a text-searchable web. Multiple articles could be integrated into a complete business page, which would be a bigger web with a table of contents and a discussion group facility. The business page could be combined with other pages across the Net to assemble a whole newspaper or magazine, which would be a still bigger web. The final document would then be installed daily or weekly on the production Web server on the public Internet, with appropriate access controls for various classes of subscribers. A number of vendors and products are moving in this direction. Many good WYSIWYG HTML editors are becoming available. NaviSoft, a subsidiary of America Online, has marketed a good visual tool, although it requires NaviSoft's proprietary Web server, database system or Web hosting services. My company, Vermeer Technologies, has created an end-user, client-server visual development environment that provides server- and vendor-independence. There is also substantial opportunity for Web-based service businesses that support use of the Web itself. The Web is now rather like the telephone system in 1900: rapidly growing, obviously right, but chaotic and messy. It's difficult to register or announce yourself if you're a vendor, and difficult to find the vendor or information you want if you're a user. Services such as Yahoo and the Commercial Sites Index are a start, and many specialized directories are being created, but we need much, much more. Given that conventional Yellow Pages publishing is a multibillion-dollar business and that newspaper classified advertising generates over $20 billion per year, the opportunity for Web directories must be large. The embryonic, shrink-wrapped Web software industry both reinforces and depends on these developments. These applications will typically be distributed systems, either client-server or peer-to-peer products, and available for multiple systems platforms (both PCs and servers). They will also be an enormous industry. If Lotus Notes has attracted 5,000 applications and a large aftermarket with a poor product and two million users, then think of what the Web will bring. It would be stunning if we do not see at least a thousand Web-based applications available as shrinkwrapped products a year from now. They will provide every combination of workflow, discussion groups, bulletin boards, packaged content, electronic commerce, version control, archiving and electronic publishing. --------------------------------------------------------------------------- The Microsoft Question But there's still a risk to the development of this new industry. Commercial work is outracing the sluggish committees responsible for Web standards, and vendors aren't cooperating enough to compensate. Some standards, such as HTML and security, are already fraying, placing at risk the compatibility critical to the Web's power. In the face of those problems lurks Microsoft. Its bid to control on-line services would gain enormously from the Web's failure to maintain interoperability. The Microsoft Network, either in its current form or as a Web-based successor, backed by Microsoft's huge market power, is the only system that could rival the Web in providing information services. Will Microsoft, the most powerful, brilliant, predatory and ruthless company in the entire technology sector, stifle or dominate the emerging Web industry? It will certainly try. For the next year or two, Microsoft will continue to emphasize its proprietary efforts, including the Microsoft Network (MSN) in public on-line services and Microsoft Exchange in corporate workgroup applications, in combination with defensive, hedging entries into Web-based software. The Microsoft Network is a fabulous buggy whip--state-of-the-art technology for its time, unfortunately used to implement the world's very last proprietary, closed, centralized, obsolete on-line service. MSN might indeed dominate the proprietary, on-line services industry, particularly if Microsoft is permitted to bundle it with its operating systems. But I doubt that Microsoft's potential dominance in proprietary services will stop the growth of the Web. The real danger is that MSN's dominance of proprietary services, combined with Microsoft's position in everything else, would gradually enable it to dominate the Web industry and electronic commerce. I have no doubt that, at some point, Bill Gates will become serious about Web systems. Already, Microsoft is entering the Windows NT Web server market, and is bundling a browser, licensed from Spyglass, into Windows 95. Certainly, only Microsoft could aspire to enforce a proprietary system standard in the face of the open, largely nonproprietary Web architecture. Microsoft's logical strategy would be to bundle Web browsers and Web servers with its operating systems, just as it is doing with the Microsoft Network. Its goal would be first to kill the emerging industry, and then lock consumers into a sequence of proprietary Microsoft additions to the Web architecture or into linkages between Microsoft Web products, proprietary electronic payment systems, the Microsoft Network, Microsoft Mail and Microsoft Exchange. The fate of Microsoft's other major architectures, especially Windows NT, SQL Server and MSN, will affect its chances of success. Windows NT, in particular, plays a major role because it is unlikely that Microsoft will support Unix, OS/2 or NetWare with its Web-based software. If non-Microsoft operating systems remain healthy, Microsoft's control will be limited. If NT takes over the world, however, it will give Microsoft a powerful stage on which to play its usual tricks. My sense is that Microsoft can become truly dominant in Web software only if it is permitted to employ highly unethical, predatory strategies. I would support legal measures to prevent any predatory behavior, especially massive bundling. Interestingly, this time, Microsoft is not the upstart fighting musclebound giants. For the first time, it is one of the proprietary, reactionary incumbents with too many agendas, facing a rapidly emerging industry. Microsoft now has both the benefits and difficulties of enormous size, visibility, complexity, huge wealth and industrial power. Everyone fears it or should. This gives Microsoft leverage, but it also creates difficulties as more companies and industries--from software vendors to banks to telecommunications firms--begin to think through the implications of dependence upon a potential competitor for their information infrastructure. Moreover, Microsoft now has serious internal politics to contend with. The right way to do Web systems is to make open, multiplatform products, with graceful interactions to all industry-standard operating systems, database systems, security technology, mail systems and the like. Microsoft can't fully support the Web without damaging many of its own proprietary efforts. I do not mean to encourage complacency, either within the Justice Department or within the Internet industry. But abject terror is not required, either. I would bet that the architected, layered Silicon Valley model will predominate, with a very important systems integration tier as its top layer. There are just too many good ideas, smart people and opportunities, and too much of the basic Web architecture is embodied in nonproprietary standards, for any single firm to dominate them all. --------------------------------------------------------------------------- A Global Transformation This new multibillion-dollar industry, moreover, will generate a profound, wrenching, fabulous transformation in the global economic, political and educational scene. By accelerating and unifying the provision of electronic information services, the Web is bringing into collision a number of industries that have eyed each other warily for years. In a well-structured, on-line world, there is very little difference-technologically or strategically--between white pages, yellow pages, classified advertising, reference services, traditional on-line services, news and publishing, and Web hosting. The Web is also starting to carry real-time voice, fax and multimedia transmissions, bringing the prospect of a sharper conflict between Internet access vendors, cable systems and the traditional telecommunications industry. That this is occurring at the same time as a massive deregulation of telecommunications is being debated politically only heightens the competitive pressures these industries will feel. Many observers have (correctly) pointed out that the Web enables disintermediation on a grand scale, both within companies and between companies and their customers. In many industries, intermediary functions--such as wholesaling, distribution, advertising, retailing and customer service--collect a much higher fraction of total costs and revenues than they intrinsically need to. The Web will reduce barriers to entry and increase startup activity in industries where the principal scale effects and entry costs derive from marketing and distribution. Software and publishing may be such industries. Some of this will be painful. It will cause a substantial compression in wages, employment levels and margins in several distribution-based industries. But there will also be large economic benefits. The Web, for example, will enable U.S. manufacturers to bypass collusive and restrictive arrangements in Japan. It is difficult to prevent access to Internet-based information. China just approved the creation of 200,000 domestic Internet access accounts because, in effect, it had to; controlling them will be impossible. Finally, the Web will accelerate the general pace of business activity, will continue the automation of many information and service functions, and will further the globalization of markets. It will also make many business processes much more transparent, both within and between companies. Federal Express Corp. already lets customers check shipment status on the Web by accessing FedEx's internal tracking systems. Wells Fargo Bank lets customers check their account balances on the Web, while many mutual funds offer their prospectuses over the Web. Here again, there will be some industries (I would bet on on-line services, publishing and financial services) where this effect will be large enough to produce a new generation of startups who will challenge the established vendors. Similar information access and distribution patterns are transforming internal corporate workflows. The creative use of Web-based systems may therefore generate a much more real and useful reengineering revolution than the "reengineering" trend ever did. To see how, look at the emerging Web industry itself. The Internet software industry is already using Web-based systems intensively and imaginatively, to refer customers to resellers and consultants, to collect bug reports and answer customer questions, to distribute software products and documentation, to conduct market research and so forth. Once again, the information technology sector is proving to be a laboratory not only for technology, but for the management and technology strategies appropriate to the control of complexity, uncertainty and rapid change. There can be no more spectacular example than this one. Charles Ferguson is the author of Computer Wars: The Fall of IBM and the Future of Global Technology, and chairman of Vermeer Technologies Inc. Contact him at charles@ferguson.com.