Every Manager's Guide to Information Technology:

Extract (1): Evolution of the IT Management Process

Extract's Table of Contents:

• The Data Processing Era
• The Management Information Systems Era
• The Information Innovation and Support (IIS) Era
• The Business Integration and Restructuring (BIR) Era

Many of the planning and accounting procedures, attitudes, and relationships in place today in organizations reflect a 40-year tradition of delegating information systems management and of handling IT costs and planning as budgeted overhead. The management process has not kept pace with technology-its uses and pervasive impact and the policy decisions needed to make it effective. As a result, professionals and business managers face one another across a gulf of unfamiliar language and culture and frequently experience mutual frustration trying to bridge it. A shared language helps, but more important it is a shared understanding of the issues that underlie the decisions. To merely define integration and architecture, for example, gives little insight into what these elements mean for the planning process or business managers' contributions to it.

In retrospect, we can identify four fairly distinct eras in the evolution of IT in organizations: 

• Data processing (DP) (1960s)
  
  
• Management information systems (MIS) (1970s)
  
  
• Information innovation and support (IIS) (1980s)
  
  

Business integration and restructuring (BIR) (1990s) 
Many of today's IS management practices were created in the MIS era and are now being adapted to meet the stresses and challenges of the IIS era. A relatively small number of firms are well positioned to use IT as a business resource that can help them rethink and restructure their organizations and business operations. (I estimate that perhaps 10 percent of the Fortune 1000 are at or almost at that stage, with a slightly higher percentage of smaller firms less constrained by old management practices of managing IT. These smaller companies began their deployment of IT as a business resource with the personal computer, not with large, centralized corporate data processing units.)

The Data Processing Era

Computers first became economically attractive for large and medium-sized companies in the 1960s. They were very expensive, very limited in application, and incompatible with every other computer until well into the 1970s. The word "incompatible" was assumed and therefore rarely used.

Developments in computer use were hardware driven, relying on improvements in equipment costs, capacity, and speeds; applications were built from scratch, as the third-party supplier of packaged software was not yet in existence. Telecommunications was confined almost entirely to telephones until the 1970s, when "voice" phone lines began providing access to computers from remote terminals. The transmission techniques for early data communications were slow and expensive. Because telecommunications was entirely regulated and choices of technology restricted, companies had no need to develop in-house technologies beyond operational telecommunications skills. Telecommunications was an operational backwater; the action was in data processing.

The economics of computer use during this era pushed firms toward automating large-scale clerical activities-hence the label "data processing." (Staff at a number of U.S. government agencies still routinely speak of ADP-automated data processing.) Payroll and accounting applications were natural targets of opportunity. If a computer that cost, say, $5 million could be justified only by the large numbers of fairly low-level administrative or clerical staff it displaced, then the natural starting points for data processing were clerical activities that involved repetitive, high-volume transactions based on applying automatic rules, calculations, and procedures.

Automating these functions was much harder than anticipated. Computer programmers with little if any business experience, interest, and aptitude had to wrestle with a new and complex technology, developing methods as they went along. The field was young, as were the people in it. Their value to the business-their analytic skills and infatuation with the technology-also became their limitation. They tended to be intellectually and attitudinally unable to put themselves in the shoes of "users," who were frequently afraid of losing their jobs or influence because of computers and of looking foolish because they did not understand them, or were annoyed at the condescension of programmers. The research literature in the IT field in the early 1970s frequently identified "resistance to change" as a major explanation for the frequent organizational failure of systems that worked technically. There is plenty of evidence that much of this resistance was to computer people, not to computers.

DP developed as a specialized staff function isolated from the business. The isolation was both psychological and physical. Many companies placed their DP units in new buildings designed to house large computers; those were frequently many miles away from any of the business operations; DP people rarely saw their users.

Even when they did, their main concerns were with the day's DP problems. Taming the technology, developing systematic project management methods, absorbing a flood of new programming languages, and managing operations consumed DP's resources and attention. The head of DP was almost invariably someone who had come up through the ranks in the programming field. He-there were very few she's at the top in DP, even though this was the first major business field in which skilled women were welcomed because of the often desperate shortage of capable staff-operated a factory dominated by systems development, central operations, and increasing investments in "maintenance," a somewhat misleading term for the work needed to keep existing systems functioning effectively and efficiently. When the tax law changed, the payroll system had to be modified; the firm could not say, "But we like the old tax laws; we'll keep them." Few if any experts in DP ever anticipated that maintaining old systems would consume more effort and resources than building new ones. In the 1990s, those old systems are a concrete block on the feet of Information Services desperately trying to meet the business demand for new ones.

Except when a major systems development project affected their sphere of authority and responsibility, business managers had little interaction with DP. Generally, the people who worked with DP in specifying systems were lower-level supervisors. Frequently, busy business units were reluctant to release their best staff to work with DP, and the people assigned were those who were expendable. This hardly improved the quality of design.

Because fully half of application development projects badly overran their budgets, failed to perform satisfactorily relative to expectations and needs, or were abandoned before completion, business executives tended to view DP as a problem rather than as an opportunity. Furthermore, because it was both a new field and peripheral to the mainstream of business, it was not an attractive career option for most business managers. It had been neither part of their college and management education nor part of their personal development.

It was a frustrating situation for the best data processing professionals, who very often had outstanding analytic skills and were ferociously hardworking, self-motivated, and honestly committed to delivering first-rate systems. Their explanations of the frequent software development foul-ups highlighted users' inability to determine their needs and to decide on specifications, and their unwillingness to commit their time, knowledge, and prestige to the project. It was in this period that computer people began to be seen as "different" from business people.

They also were seen as expensive. Business managers came to view DP as an escalating cost, to be carefully controlled. This control was achieved principally by allocating costs on a formula basis to the departments that used the DP organization's services. Allocations were generally charged out so that DP expenses were fully recovered. Some firms ignored allocations, instead absorbing the costs centrally, with systems development projects paid for by the targeted business unit.

For many business units, "chargeouts" were frustrating; unit managers could not control IT costs, did not see much value in IT, and felt that IT professionals were not only too highly paid, but also far too ready to move on to the next lucrative job in the programmers' boom market of the 1960s through the mid-1980s. IT allocations were like a corporate property tax on business units.

The technology of the 1960s and 1970s was characterized by large economies of scale. A reliable rule of thumb was that a computer twice the cost of the one in place would provide four times the power. This naturally led to an emphasis on centralization and to planning based on hardware-driven decisions. Because it was expensive, hardware was treated as a capital item; the less-expensive staff resource became an annual expense.

Because development was not viewed as a capital item, few firms, even today, have any idea how much has been spent to create the software systems they use. The accounting system expensed salaries for development and maintenance staff as well as operators. As a result, firms did not track the relationships between these cost elements over the life cycle of a major system. Today, firms are discovering that most of the costs of IT are hidden. For instance, the costs of support and training to use a personal computer are far higher than the purchase price, and the cost of developing a system is often a fraction of the cost of operations and maintenance.

In the DP tradition, spending on computers is largely determined by an annual budgeting process. Next year's budget is set by aggregating business units' requests for new systems, comparing the firm's own expenditures as a percentage of sales with that of comparable companies, and capping the budget increase at X percentage. In a good year, X goes up; in a tough year, it is held constant or cut. The systems development and operations plan is fine-tuned within these parameters.

The heritage of the DP era remains a strong force in many firms. The following are among its remnants.

Naive chargeout mechanisms that block investment in longer-term infrastructures. If costs must be fully recovered and there is no central funding, the initial users must bear the total investment. Today, users can often find less-expensive alternatives in increasingly cost-effective personal computers, departmental communications, and off-the-shelf software. Cost allocations and recovery schemes are two of the major impediments to integration and architecture, because they discourage long-term investments in infrastructures that must be paid for by the early users.

A broad gulf between senior business managers and IT managers. Typical 40- or 50-year-old executives-for whom computers were not part of their education, not part of their move up the management ladder, and, in their new-found senior status, not a necessary personal tool-are ill prepared to play an active role in major IT policy-making and planning. "I am proud to say that I have worked in every single area of the company except computers and that will be true when I retire," the CEO of one Fortune100 company boasted publicly. More common is the manager who says, "I am too old to learn about computers."

A firm grounding in computing. At one time, computing provided the main body of experience, expertise, and authority for information technology; telecommunications played almost no role. But this scheme has shifted rapidly. Yesterday's add-on is today's infrastructure. Telecommunications today is the highway system over which the traffic of an increasingly distributed resource travels.

A continuing bottleneck in software development and the burden of old systems. DP lost much credibility because it so rarely delivered on its promises. One leading authority, Frederick Brooks, who led the development of one of the largest civilian software development efforts ever when he worked at IBM, coined the term "the mythical man-month" to describe programmers' estimates of how long a project would take; typically, the figure was well over 100 percent too optimistic. Successes in systems development were overshadowed by delays, cost overruns, and bugs for most of the 1960s and 1970s. Large-scale software engineering is a difficult craft even today; the experiences of those years persist, but not for lack of effort, skill, or commitment on the part of development staff. The process is inherently complex and software productivity continues to lag behind other areas of information technology. It is growing in complexity, too. The new style of design called client/server computing combines every IT element-old and new, proven and unproven alike-to create an integrated, cost-efficient service and information base. This is the challenge for the 1990s.

The Management Information Systems Era

With most basic clerical and accounting processes automated by the mid-1970s, the focus of attention in Information Systems shifted to designing and building reporting systems to meet managers' information needs. These efforts were largely flawed for two reasons: limitations in the technology and the mistaken equating of "information" with "data." The technology of this period was still that of the data processing era-large, expensive, inflexible computers that generated enormous volumes of typically historical, highly detailed, accounting-based data. The rationale of MIS was to organize and present this data to managers, assuming almost by definition that the more data they had the better.

In practice, the data met few managers' needs; for example, because reports generally lagged events by several months, they were of little value for competitive analysis or for spotting trends and problems early enough to react constructively. Moreover, accounting data yielded only limited information for management decision making. Its primary value was for planning and administration. Changes in the technology gradually opened up new opportunities for turning data into information and tailoring systems to meet managers' real needs.

One important new tool for doing this was the computer terminal, which for the first time permitted flexible, occasional, and ad hoc access to central information stores and to "time-shared" processing on large computers. Time-sharing was an innovation that allowed a large computer to process many tasks (or jobs) simultaneously by giving each a small slice of time and going from one to the other very quickly. The result was that many people doing different work at the same time could be "logged on" to the computer. Given the economies of scale of computing, this encouraged the growth of time-sharing "bureaus" that could offer services individual firms could not afford. Time-sharing is still the base for today's airline reservation systems and automated teller machines, though more and more of the processing is shared between "client" software and hardware that requests services from "servers." With time-sharing, the central computer handled all processing. Now, a personal computer (client) can input your travel reservation information, check for errors and completeness, and send the information to the central time-shared computer, which stores data on seat availability. This distribution of tasks from mainframe to local devices and sharing of resources between clients and servers is the basic driver of every trend in the use of IT in organizations today-and will remain so for the next decade.

Another useful tool that emerged in the 1970s to help speed up the development of systems was packaged software sets of programs written by third-party developers. Most of these packages addressed needs that did not require company-specific software development, such as report generation, financial modeling, and general ledger accounting. The combination of time-sharing and flexible packages and modeling languages stimulated rapid innovation in what would be termed "decision support systems" and "end-user computing," a process that personal computers accelerated principally because the operating costs of timesharing were high and those of personal computers low.

Nearly all of this innovation occurred outside the central MIS units, mostly in the finance or marketing departments where small teams of business-oriented planners picked up the new tools and applied them imaginatively. Management Information Systems groups, like the mainstream technology they controlled, remained monolithic. The problem of balancing control and innovation led to continued crisis for MIS throughout the mid 1970s. When computers first came into companies, few if any control groups managed them, and the most effective units were those that encouraged experimentation. But to the extent that they managed to "sell" the wonders of the new technology, they created a demand that they could not reliably meet. Experimentation too easily became chaos. Costs escalated. Planning procedures were informal, at best. Software development was an ad hoc process, with undocumented, untested, and unmaintainable "spaghetti" code a frequent result. Testing and maintenance were neglected as the overcommitted DP group tried to catch up with a systems development backlog that averaged from three to five years of resources.

The MIS era saw efforts to introduce discipline and professionalism and to control MIS costs and resources more sensibly. Too often, control created bureaucracy. With MIS wrestling with major changes in technology, continued and growing problems in software development and project management, and senior executives interested in getting costs under control, the Information Systems function found itself increasingly on the defensive. MIS staff rarely understood the business, and few business executives understood MIS.

Most stereotypes of computers and computer people date from this period. Remember "Do not fold, spindle, or mutilate"? Were you among those who received bills for $00.00 and subsequent notice that legal proceedings would be initiated if payment were not forthcoming? Maybe you were one of the managers who received massive computer printouts of monthly reports that went straight into the trash. At the very least, you were probably informed in response to an inquiry that "the computer made a mistake."

Given that during this era most innovation occurred outside the MIS area, it is not surprising that MIS professionals became frequent resisters of change. The widespread early opposition of most, though by no means all, central MIS staff to the introduction of personal computers in the early 1980s was a case in point. MIS staff wrote off the new "micros" as machines for amateurs, viewing themselves as the "experts" on matters relating to computers; they worried, justifiably, about problems of testing, documentation, and security.

If this summary of the MIS era sounds negative, it is because little of a positive nature occurred during the period. The field of MIS ossified at a time when office technology, end-user computing, and decision support systems began to flourish. MIS became stuck in an organizational, technical, and attitudinal rut; it failed to build a constituency (few business units viewed the traditional MIS group as more than a necessary part of operations overhead); it had no strong management sponsor.

Although the problems MIS faced were sufficiently complex to persist today-notably large-scale software development and maintenance of old systems-MIS was not without substantive achievements. Its principal technical development was the shift from transaction processing to data base management. In the DP era, the main challenge was computer programming, which relied on quirky minds to develop immensely detailed and precise instructions to carry out even simple transactions or generate reports. Data files-weekly payroll records, customer history files, or purchase orders-were organized for processing efficiency, and data were frequently duplicated. A bank, for example, would store customer names and addresses in each pertinent master file for such business applications as checking accounts, mortgages, and savings accounts. Changes of address had to be entered in each file.

Inconsistency, redundancy, and duplication of data became commonplace and contributed to the bureaucratization of the MIS function, adding considerably to the data administration burden of user departments and customers.

The MIS era saw the creation of the concept of data-base management systems (DBMS), an innovation that has evolved slowly but consistently over the past fifteen years. With a DBMS, data are organized in much the same way as books are in a library. Information, such as a name and address, is stored in only one place, on a computer tape or disk. The DBMS software includes the equivalent of the library's index of authors and subjects. A transaction processing or report program requests the customer record, which the DBMS retrieves, just as a borrower uses the library's card catalogue to locate the shelf and unique reference number for a book. Any number of programs thus access the same data items, instead of storing them many times over in separate files. When information such as a customer's address is changed, the DBMS locates the relevant item and updates it. The DBMS will also check for errors and handle such aspects of security as providing access to information only to authorized users.

A library catalogues books by title and subject, but not by chapter. The evolution of DBMS has been toward the most detailed level of indexing-the equivalent of indexing a book by phrase or word. Initially, data were organized in records; a customer master record might contain many data items, including last name, first name, middle initial, social security number, home phone number, and so on. Early DBMS accessed information at the record level. "Relational data-base management systems" now access it at the item level and allow very complex searches through data bases in order to answer such questions as, "Show me all the customers in ZIP code 20007 with outstanding balances of more than $500 who own their own homes."

Data-base management software incurred substantial processing overhead and, for most of the MIS era, was too expensive for many applications and completely impractical for major transaction processing systems. But in shifting the focus from programming to data and data management, the genesis of data-base management systems marked a turning point in the role of Information Systems organizations. More and more, IS has become a cornerstone in the balanced coordination of key, shared, corporate infrastructures, and decentralized use of IT. With this shift, the old centralized data centers housing mainframe computers have been turned into information libraries. The computerless office has equally turned into a data center, with PCs, telecommunications, software packages, DBMS, printers, and disks. This "distributed" resource works cooperatively with the central libraries. The technical, organizational, and economic challenges to manage this resource ironically depend on many of the old skills of the DP professional.

The Information Innovation and Support (IIS) Era

The history of IT in large organizations has been a pendulum, swinging between extremes of innovation and discipline and thus of emphasis on decentralization and centralization. As the pendulum swung back from the freewheeling days of early DP, it met the discipline of MIS.

Indeed, the MIS era marked the extreme of centralization in computing. And there the pendulum hung, through the ossification of traditional MIS, before swinging back toward innovation and a deliberate lack of over- restrictive discipline, first in the early use of personal computers and then in the transformation of both the technology and its uses that subsequently made innovation the norm. We call this position of the pendulum the era of Information Innovation and Support (IIS).

The swing away from over-discipline, which began in the early 1980s, was fueled primarily by changes in the technology, most obviously office technology and personal computers. These created an entirely new market for IT. The term "information technology" replaced "computers," "information systems," and "data processing," as telecommunications became the access vehicle for computing services, data-base management systems were opened to proliferating personal computers, and low-cost, do-it-yourself software flooded the market.

Office technology was the first step in the organizational perestroika of IT. Business units purchased word processors from a wide range of vendors; personal computers appeared on the desks of many professionals and some executives; and minicomputers and microcomputers introduced an alternative for departmental computers beyond what came to be called "mainframes." Not a real option in the 1970s at a price of $5 million, departmental machines were practical at $100,000 in the 1980s and $25,000 or even less in the early 1990s. Time-sharing on mainframes was expensive in the 1970s and early 1980s, costing anywhere from $20 to $200 an hour.

The microprocessor-based personal computer, as much an economic as technical innovation, boasted an operating cost of pennies per hour. With this technology, the analyst who wanted to run what if ? analyses of budgets for ten hours a week no longer had to justify more than $20,000 annually for time-sharing. Initially, MIS units mostly ignored, blocked, or tried to control such innovations. The skill base of MIS rested on traditional programming languages, transaction processing systems, and formal project management methods, all of which remained important and will remain important in the future. Inertia in MIS actually stimulated the shift toward business unit autonomy, reinforced and even hastened by the personal computer. Business managers now had choices outside the old MIS monopoly, and they exercised those choices.

They also looked at computers through a new lens. Just about every major application in the DP and MIS era looked inward, at the company's own operations. The IIS era looked outward and searched for sources of competitive advantage from IT. Indeed, IT-and-competitive-advantage almost became one word in the early 1980s. Consultants and academics chased after examples of firms that had gained-or that seemed to have gained-a sustained edge over their competitors by using IT.

Some of these examples were over-hyped; the advantage turned out to be transitory, or where it was real, the firm got in trouble because of something totally unrelated to IT, thus showing that IT could never substitute for business vision, organization, and management. That said, the new focus on competition, business, service, and customers helped transform the mindset of IS and its vocabulary.

Part of sweeping away the cobwebs came from telecommunications. By the mid-1980s, the technology of telecommunications had begun a transformation that continues to be even more radical and spectacular than that of computers. The telephone lines of the 1970s, designed to carry voice traffic, were slow, unreliable, and very expensive for moving computer data. Local area networks, fiber optics, satellite technology, and a wide range of new equipment that exploited microprocessors increased the speeds and reduced the costs of telecommunications by factors of hundreds in the 1980s, a pace that looks slow compared with what is happening in the 1990s.

Deregulation of long-distance services in the United States stimulated fierce competition, imaginative new products, technological innovation, and new sources of supply. Large firms were no longer confined to POTS-plain old telephone systems-but could design their own telecommunications infrastructures that could deliver a wide range of electronic services. In addition, local area networks, which inexpensively linked personal computers across short distances within a building, provided a base for explosive growth in departmental computing and data management facilities. The resulting IT resources were more powerful, yet far more cost-effective and easy to manage, than the DP shops of the 1970s.

Together, telecommunications and personal computers liberated the use of IT in organizations. Without efficient and low-cost data communications, decentralized departmental computers could not share information among themselves. To share information across business units and functions, computers had to be centralized. The growing availability of telecommunications led to what is termed "distributed" systems, a combination of workstations linked to midsized departmental computers or central mainframes or both. Computers could now be located anywhere and linked among companies as well as departments.

The combination of frustration with the rigidities of traditional centralized MIS, the ready availability of personal computers and packaged software, and local and wide area telecommunications that enabled them to connect with one another and with larger computers, shifted the focus of IT from automation to innovation. This search for innovation has been explicit and encompasses all areas of IT. The role of MIS shifted to trying to position IT as a new source of competitive opportunity and advantage, in part by supporting the now-institutionalized, distributed, and effectively autonomous users of personal computers.

Individuals exploited electronic spreadsheets, laptop computers, word processing, and desktop publishing; groups and departments exploited electronic mail, local area networks, and shared data resources; and business units transformed customer service. Companies embraced IT as a way to just-in-time everything: online ordering, computer-integrated manufacturing, JIT inventory; the list grows.

The mentality of the new Information Services units that grew out of MIS shifted from control to coordination and support. The Information Services' view of distributed systems built from the mainframe outward, first linking personal computers to central services and information stores, and then redesigning transaction processing, communications, and data-base management systems to exploit the strengths of each element. The implementation of these strategies is variously termed the "client/server model" and "cooperative processing."

The old MIS career path of programmer to project leader to manager was complemented, and in many instances supplanted, by a career path emphasizing a level of business understanding that was virtually nonexistent before. Growth of knowledge and bridging the attitude gaps between business and IS professionals became priorities. Management "awareness" education programs for senior executives, "help desks" for personal computer users, and IS steering committees were typical initiatives. IS and business units encouraged the development of a new form of hybrid staff, skilled in business and knowledgeable about IT or vice versa. Leading IS managers, increasingly called chief information officers or CIOs, opened the MIS fortress and emphasized business support as the priority for IS. (The renaming of DP to MIS and MIS to IS was not a gimmick but a real effort to signal shifts in the role of the IT unit.)

Today innovation is everywhere, much of it uncoordinated. Many "islands" of IT have been created, partly because business units were able to go their own way rather than depend on MIS, but mostly because the entire technology base, large scale and small scale, has for so long been dominated by "incompatibility." Every major vendor's equipment-hardware, software, telecommunications, and data management-has been "proprietary." IBM and Apple Macintosh personal computers, Ethernet and Token Ring local area networks, Digital Equipment Corporation's VMS and IBM's MVS operating systems, and most telecommunications networks were incompatible with one another. There are technical, marketing, and political explanations for this, but the reasons matter far less than the impact.

The most farsighted IS managers viewed with growing concern the fragmentation and incompatibilities among systems across the firm. As long as units remained self-contained islands, lack of integration was not a problem and could even be an advantage in that it enabled them to choose technology solely on the basis of what best suited their needs. But more and more elements of business were becoming interdependent: departments needed to share information with other locations and functional areas, and business units increasingly needed to combine data to create new products and services. Meanwhile, the cost of supporting multiple networks and software escalated. The absurdities of incompatibility meant that personal computers within a department could not even read files that had been word processed on another personal computer, use the same printer, or share a local area network, and efforts to link major transaction processes and rationalize telecommunications turned these simple local systems into organization-wide messes.

Whereas business units naturally placed priority on autonomy, quick installation, and low-cost operation, IS focused on the need for integration, immediately or later. "Architecture" became a key term in IS but not in most users' vocabularies. The dismal track record of old-line MIS led many users to view "architecture," "integration," and "standards" as rearguard efforts to reestablish control. The major barrier to central coordination was the very same tool that provided the most cost-effective IT base for user departments-the local area network. This area of IT innovation was easily the one most beset by incompatibility. The calculus of the 1980s favored innovation at the risk of loss of discipline. But local innovation turned out to have its own costs and limitations, and the pendulum began to swing back again near the end of the 1980s.

The Business Integration and Restructuring (BIR) Era

Today the pendulum is swinging toward an innovative discipline that balances central coordination of key IT infrastructures and the use of IT within these infrastructures with decentralized autonomy in decisions about applications. As the costs of incompatibility became clearer to IT users and providers, the telecommunications community mounted a major effort to create "open systems." Committees defined "standards" that would connect equipment manufactured by different vendors in order to provide various types of services. The user community played an increasingly active role in the standard-setting process, creating a number of de facto standards by its choices of vendors and systems. The IBM personal computer is a ready example of how a "proprietary" system, by virtue of gaining substantial market share, became an effectively "open" one. The success of the IBM PC led other vendors to create "IBM-compatible" products. Legally, they could not fully replicate IBM's hardware, so they built products that ran under the IBM PC's software operating system, called MS-DOS. Today, DOS machines are everywhere; users created this situation through their own purchasing decisions.

Standards, integration, and open systems moved to the top of the IS management agenda in the early 1990s. A new industry of "systems integrators" had emerged that grew at a rate of 20 percent a year, largely because there was so much systems disintegration to repair. Standard-setting committees focused on ending the Tower of Babel of incompatibility. For the first time, vendors anticipated the need to make sure their systems were compatible instead of trying to lock customers in through proprietary systems in the way that IBM and Apple had.

There are many ongoing efforts by vendors, government agencies, standard-setting committees, and business users of IT to agree on open standards or to ensure that proprietary ones can work with each other and with open standards. For instance, the Open Software Foundation defined a comprehensive Distributed Computing Environment that is a step toward the equivalent of the open systems used in construction-a components approach to designing a house that requires builders to use specific standard materials.

It is highly uncertain whether the many efforts to make IT components this open will succeed, because defining standards takes many years, and implementing standards in real products can take longer-if, in fact, the standard can be defined in the first place, the implementations can be made complete and consistent, the innovations in the technology do not make the standard obsolete, and if the vendors do not add their own special "nonstandard" features. That said, open systems are now the enthusiastic or reluctant target of every major vendor and user.

The question today is how to achieve these open systems. Because the standard-setting process cannot keep pace with the change in technology, there are many gaps in standards and controversies about the implementation potential of many that have been defined.

But the major policy issues in IT, from a business perspective, concern investments in infrastructures. The Information Innovation and Support era largely emphasized applications-new purchasing systems, local area networks, workstations for computer-aided design and manufacturing, funds transfer networks, and so forth. Formal policies on standards, because they were viewed as bureaucratic, tended to be avoided.

Today, with business integration driving technology integration, standards are seen as essential. Insurance firms increasingly rely on shared cross-product customer data resources for cross selling and targeted marketing; more and more customers and suppliers are linking their IT bases through electronic data interchange to eliminate such paper documents as purchase orders and invoices; transnational firms are connecting previously separate international systems in order to coordinate operations across twenty-four time zones.

Leading IT thinkers and practitioners have recently begun to question the ethos of automation that underlies the origins of IT. Terms such as "business process reengineering" and "informate instead of automate" have become the new watchwords of the IT profession in what is becoming part of a general rethinking of principles of organizational design in large enterprises; "teams," "communication," the "networked organization," and "collaboration" are replacing the traditional emphasis on planning, hierarchy, decision making, and control. The restructuring of organizational processes and structures is emerging as a business priority, with IT as a powerful potential contributor, especially through telecommunications, which removes barriers of location and time on coordination, service, and collaboration.

The view of IT as a major enabler of new forms of organization includes new inter-organizational relationships and processes, and a mechanism for streamlining and eliminating work rather than automating it. Electronic data interchange (EDI) is but one of many contemporary examples. As more and more large firms insist that suppliers link to them electronically or lose their supplier relationship, EDI becomes less an option and more an imperative.

In the retailing industry, EDI and "quick response" systems have enabled a few firms to reduce to a few days their logistics cycle for ordering and stocking; the gap between the leaders and the laggards is so pronounced that most experts expect that fully half of today's retailers will be out of business by 2001. The "very quick" response firms in the fabric industry take 10 days from getting an order from a store to delivering the goods, while the "average" response firms take 125 days. The same patterns and impacts are apparent in the logistical systems of insurance firms in terms of lead times to issue a policy and manufacturers in "time to market." IT is now about how firms carry out their business and how they relate to suppliers and customers. We have come along way from the DP era.

The most distinctive aspect of IT's Business Integration and Restructuring (BIR) era is that, relatively suddenly, change has accelerated in just about every area of both business and technology. Business is dominated by radical not incremental shifts; downsizing and reengineering are examples of deliberate efforts to transform basics. Most large companies are working actively to transform organizational forms and principles to free themselves and their employees from the constraints, waste, and inflexibility of the hierarchical structures that have dominated business and that are seen as a major element in the decline of long-successful industry leaders, such as IBM, General Motors, and Sears.

The technology of every element of IT is in a new state of flux, mainly because the old distinctions between computers and communications have been eroded. PC operating systems, for instance, are being designed to manage enterprise-wide telecommunications, making them far more complex than before. Similarly, software for PCs is designed to share information and coordinate work across locations.

The new force of radical change in the BIR era is IT industry change, especially in telecommunications. In the Data Processing era, IBM dominated computers and AT&T dominated telephones, with almost no commercial data communications available. In the MIS phase, the pattern continued, with limited data communications provided through AT&T, still the national monopoly. A number of new entrants became successful, mainly in providing minicomputer alternatives to IBM's machines; Digital Equipment rose to number two in the industry (but plummeted in the 1990s). The Information Innovation and Support era saw competition accelerate in personal computers, PC software, local area network communications, and long-distance communications. This era saw the rise of Microsoft, Lotus, Novell, Cisco, SynOptics, Oracle, Sybase, Dell, Compaq, and many others whose sales reached more than a billion dollars a year. It also was the era of the deregulation of long-distance telecommunications, with MCI becoming an increasingly effective competitor to AT&T.

During this period, long-distance telecommunications technology remained stable, with continued improvements in price and performance. That has changed dramatically in the mid 1990s. Cable television firms, cellular providers, long-distance companies, and the local phone company monopolies are suddenly all in competition, allying with one another, or converging on each others' markets. In both local and wide area telecommunications, breakthroughs in technology are moving faster than the microchip revolution of the 1980s, which put the personal computer on the desktop and in the briefcase. Cellular wireless communications have, after years of fragmented growth, entered a phase of rapid expansion and competition. "Information" technology now means any type of information and "technology" means any type of delivery base. This is creating a turmoil of new changes, leaving managers with no previous experience to draw on.

In the computer industry, too, turmoil is the norm. The computer manufacturers that dominated and in many ways controlled the entire pace of change during the first three eras are almost all either in trouble or peripheral to the new mainstream drivers of change. IBM, Digital Equipment, Wang, Data General, and Unisys are examples. Microsoft dominates the use of computers and the main directions of key parts of the PC hardware, software, and local communications markets, much like IBM dominated the IT industry during the 1970s and much of the 1980s. A series of fierce contentions are underway, many involving the well-armed Microsoft, to dominate as the supplier of desktop and network hardware and software products and services to large companies.

Business has changed radically in the mid-1990s, both domestically and internationally. The 1970s and 1980s pictured the United States, Japan, and West Germany as dominating global business. These days we see a limping Japan, a battered Germany, and a United States where no business can be sure of long-term survival.

Organizational forms and technology are changing radically. The IT industry has changed into an entirely new set of businesses and markets. The firms that balance and mesh their responses to these combinations of radical forces will not only be successful but will shape the new rules of change in the coming decade. It is unlikely that any firm that focuses on only some of these changes will succeed. Most firms will find it easier to address business and organizational change, and perhaps IT change in traditional areas. IT is the likeliest to be neglected or overlooked because this has been true for decades now, and very few managers are comfortable with and skilled in managing the technology.

In this context, business managers can still delegate the technical work, but not the planning that drives it. Because defining, funding, and implementing infrastructures, architecture, and integration involve long lead times and the crossing of functional boundaries, they demand top-level attention to policy and careful definition of business and organizational priorities to guide the choice of standards and pace and degree of integration.

This is the context in which we ask the question, "How little do managers need to know about IT to play an effective role in planning its use?" The answer is, "Enough to help move the firm forward to the era of business integration and restructuring."

Order from:

Order from: