Shaping The Future:

The path of least resistance in choosing an IT strategy is bottom-up, application-centered planning. In terms of future corporate business needs, it is not hard to make the case for such an approach. Senior executives neither can nor should be involved in the technical details of the firm's IT strategy. A set of policy-level requirements can be identified that provides a base for assessing a firm's existing IT capabilities and deciding whether there is a compelling business need for an IT platform. These policies can also help a firm's IT planners form a clear business framework and political mandate for designing a technical architecture for the platform built on a set of standards that imply criteria for selecting strategic vendors. The decision-making sequence for positioning the platform thus moves from business vision to policy to architecture to implementation.

The policy criteria that follow are essential to avoid being put at a competitive disadvantage. They incorporate the reasons an informed CEO is likely to insist on a platform that pushes as far along the reach/range combination as practical, while balancing technical risk with business risk and cost. No vendor has yet achieved either total integration or total openness. Hence there is inevitably some risk in choosing one vendor or standard as the strategic base for the IT platform. The business risk is the cost of loss of freedom, especially not being able to move with or ahead of the competition in areas in which IT is a necessary driver or supporter.

Policy 1: Business Practicality

Our IT base must never block a practical and important business initiative.

Most firms are paying more attention to the link between business needs and IT capacity and capability than they did when computers were confined to clerical operations and telecommunications meant telephones. In many instances, the IT links needed to enable or support a business initiative can be obtained from elsewhere. For example, a company that wants to establish links to its main suppliers for data interchange can use a value-added network service (VAN) supplied by companies such as GEIS or INFONET, or develop its own software but rent telecommunications facilities from, say, TELENET or MCI.

So long as a practical and important idea can be built on its own independent IT base, the specific choice of technology is irrelevant to the wider context of the firm as a whole. However, the inefficiencies and operational complexities of maintaining, say, four to forty different technical systems, with different skill needs, operating requirements, equipment, and software is in itself a problem; when they need to be made interdependent, the existing separate IT bases can be a blockage to business development.

Many likely business initiatives now and throughout the decade will involve linking firms' previously separate IT resources, such as manufacturing and financial systems, with their own and other firms' payments, materials management, stock forecasting, and EDI services. This requires additional reach. At the same time, multivendor systems must be able to interlink and connect with the systems of other firms, and fundamentally different technologies, such as image-processing systems and data base management systems, must be able to share information. This requires additional range.

Today's systems cannot do this automatically. Better IT planners will have anticipated these needs and made them major considerations in their choice of vendors and of standards. In doing so, they will have tried to balance short-term operations and cost pressures with provision for longer-term integration. Vendors, systems, and standards best suited to range are almost certain to limit reach and vice versa. Given the context, architecture planning today is at best an art form that involves complex trade-offs and judgments.

The problem is not as much reach as range. Reach can almost always be provided, if not always efficiently, through established telecommunications standards and gateways. Range is almost impossible to ensure without a firm commitment to a platform, and even then is a major challenge. The details of information sharing across technology bases are extremely complex. They relate to the basic design of the underlying computer systems, operating systems software, codes, technical data structures, message formats, and so forth. It is currently impossible to guarantee that an application built on vendor X's technology base can share information directly with an application on vendor Y's. Indeed, it often happens that such information cannot even be shared with a vendor X system that uses a different component.

There is another approach: avoid any business initiatives that involve the interdependencies that create the problem. This makes the technology base simpler to build and operate and almost certainly less expensive, and allows the CEO to leave major IT decisions to the business units. But this is a sensible option only if EDI, POS, customer-supplier linkages, computer-integrated manufacturing, and cross-linking of individual products and services are not likely to be in the mainstream of the firm's activities within the next five years. It is hard to find an industry in which this is likely to be the case.

Policy 2: Competitive Lockout

If our competition uses IT as the base for a successful initiative, we must not be automatically locked out of countering or imitating it.

Leading firms in many industries, even firms with strong IT capability, have been unable to respond to IT-based initiatives by competitors during the past decade. Delta, Johnson & Johnson, and Bank of America, to name just a few, were left at a competitive disadvantage by American Airlines, American Hospital Supply, and Citibank, respectively. The ability to respond to the moves of competitors depends both on having IT skills in place and on the technical bases of the competitor and the company looking to respond. An initiative based on an off-the-shelf software package can be countered easily by buying the same package. A space-age application that many other firms in the industry can copy is less likely to yield sustained competitive advantage than a platform-dependent application that few or none can match.

There is an immense difference between a competitive IT application and a competitive platform. A major weakness of the extensive academic and popular literature on the links between IT and competitive advantage is that it ignores the actual technology. Technology must be put back into discussions of IT and competitive advantage. Senior managers are aware that information technology is at least potentially linked to competitive positioning, that it is no longer an overhead function. That awareness opens up new possibilities for action. Without it, managers are blind to both possibilities and necessities.

Ironically, raising senior managers' awareness about IT can decrease their understanding of the importance of information technology. The main reasons for this are that:

• almost none of the cautionary tales and examples describe the technology base used to obtain the competitive advantage, implying that the choice of technology is irrelevant;⁶
  
  
• few identify precisely what establishes the advantage, hiding the difference between a competitive application and a competitive platform and between an advantage and a sustainable advantage; and
  
  
• awareness is passive, removing roadblocks to changing how IT is used but not providing a guiding vision for IT planning.

If we ask, "What difference does the choice of technology make?" many of the old stories take on new meaning. Merrill Lynch's Cash Management Account, for example, relied on linking its systems and those of Banc One of Ohio. The cornerstone of CMA was a Banc One Visa card and a transaction processing system that included a "zero balance" account that cleared all CMA transactions daily. CMA could have been a major fiasco had Banc One not been able to adjust smoothly and quickly to the totally unexpected level of demand. Merrill Lynch had forecast about 40,000 customers for CMA. At one point, it was signing up 20,000 accounts per month. Eventually more than one million accounts were established without exhausting capacity.

An example of a comparable business initiative by another frequently cited exemplar, Citibank, shows how a convenient, cost-effective, short-term technical choice can be a long-term mistake. The special-purpose hardware Citibank used to implement its international COSMOS system permitted rapid development of small-scale applications and low-cost operation but could not handle large volumes, necessitating a change of vendor at great expense and loss of momentum. Additionally, Citibank had to run two versions of COSMOS, one under each of the new vendor's different operating systems for mid-sized and large computers.

The AAdvantage frequent flyer program is widely cited as an IT innovation that changed an industry. Sabre, AA's reservation system, and United Airlines' Apollo are both viewed as making IT a central element in airline marketing and distribution to a degree that computerized reservation systems (CRSs) have become the central battleground domestically and internationally.

AAdvantage succeeded mainly because it was based on the Sabre platform, which enabled American to cross-link passenger data from Sabre to AAdvantage to its yield control systems and to its hub planning systems. Its competitors could not do that. They had the same applications as American but not the platform that integrated them and that allowed-and still allows-American to add application after application, with cross-application linkages an automatic by-product. A manager in one of American's leading competitors describes its equivalent frequent flyer program-a separate application with the same degree of reach as its CRS but without the range to cross-link with it-as an "albatross," all cost and no benefit. Sable not only created a distinct competitive advantage for American; it also imposed a competitive lockout on most of the airline industry.

Policy 3: Electronic Alliances

We will match the competition in being able to make alliances, create value-added partnerships, or enter consortia.

Increasingly, non-technical standards for business interchange are becoming as important as the technical standards and architectures upon which IT platforms are built. Many of these industry-specific standards, which define formats for transaction documents and procedures for electronic processing, are defined by industry associations. They include the U.S. grocery industry's uniform communication standard (UCS), the international banking community's SWIFT standard, and comparable standards in transportation, warehousing, automobiles, and aerospace. The international EDIFACT and closely related U.S. X.12 standards are the basis for many of these business standards. It is no exaggeration to say that EDIFACT is as important as SWIFT in creating a base for entirely new electronic markets and "meta-businesses."

These interchange standards are critical requirements for inter-company uses of IT such as point of sale, EDI, customer/supplier links, and the creation of new cross-industry services such as hotel and car rental bookings provided with airline reservations. As more businesses move their core operations on line and create electronic links to other firms, they will choose electronic partners and be chosen by them on the basis of IT reach and range and interchange standards. One of the largest U.S. banks, for example, lost a $400 million client because it could not link its payment systems to the customer's as part of an electronic supply system that used General Motors' Manufacturing Automation Protocol (MAP). A firm whose IT base meets only its internal needs will be at an obvious disadvantage.

In the 1990s, on-line operations will naturally, rapidly, and inevitably be extended across organizational boundaries, with IT as the enabler. This poses very new challenges for both IT managers and business executives. What range of business interchange capabilities must they plan for? Must a bank's platform be designed to ensure transactions and information sharing with petrochemical companies and auto makers for point-of-sale and electronic payments? Should an insurance company prepare for electronic data interchange for all employee benefits paperwork and payments involved in servicing corporate clients? Should a parts supplier assume that its engineering department must be able to interchange technical specifications with vendors and customers? Should a bank adopt the MAP standard that General Motors requires its suppliers to use on the premise that banks (and GM's insurance providers) will have to provide the same electronic linkages? Since there are no widely established standards for graphic design documents, which ones should a manufacturer monitor and be ready to adopt at the appropriate time? Should pharmaceutical firms agree on common standards for image and electronic data interchange so that they can propose them to the FDA and speed up the complex certification process for new drugs?

The answer to all these questions is surely "yes." They do not necessarily relate to specific applications; they are quite literally a platform on which to build applications. They are part of long-range corporate planning.

Policy 4: Reorganization and Acquisitions

If we reorganize, make acquisitions or divestments, or relocate operations, our IT systems will adapt quickly and routinely.

The first three policy requirements for the IT platform relate to ensuring firms' competitive health. This policy relates to firms' organizational health. Every CEO of a large firm assumes that mergers, acquisitions, and divestments are likely to occur. Reorganization, relocation, and related organizational changes are even more likely. Ideally, the IT platform should provide an electronic organization structure that can shift as the organization shifts. A company that acquires another firm is very likely to have IT applications that differ significantly in terms of hardware, software, and (most constraining on integration) information formats and structures. The ease with which the acquiring firm can mesh the acquired firm's systems into its own technology base may be a major determinant in its decision whether to make a deal.

Two U.K. financial service companies, for example, agreed to merge and then canceled the decision by mutual agreement when they found that their IT systems could not be combined and interlinked. Instead of creating a company with advantages of scale and joint operations, they had two separate companies within a company. Their operations and information systems were so tightly interdependent that without a common platform for linking the IT base they could not link the operations base.

The almost certain future importance of IT in acquisitions decisions strongly suggests that firms keep their IT architectures as close to the mainstream as possible. Yet very few IS and business managers have even a clue about customers' and competitors' architectures.

Policy 5: Third-Party Intrusions

No firm in our industry, or third parties outside it, will be able to intrude on our areas of strength or into the mainstream of our marketplace.

One of the distinctive features of IT is the degree to which it erodes and even eliminates boundaries between industries. The most obvious example is electronic funds transfer at point of sale (EFTPOS). Is this retailing or banking? Publix, a Florida supermarket, used its own POS base to become a leading supplier of banking services at the expense of the major state banks; it franchised its platform out to smaller banks and savings and loans and other financial institutions.

Airline reservation systems similarly now include non-airline services, and customer/supplier distribution and EDI systems incorporate electronic payments. A firm that has a first-rate platform with extensive reach will naturally look for opportunities to add traffic. If that platform combines extensive reach and range, the firm may be more effective in some areas outside its industry than firms in those areas. For example, international airlines' CRS capabilities are outstanding, but their systems for handling payments for bookings between airlines and travel agents are primitive. Banks in several countries have proposed replacing the airline bank settlement plan with their own electronic payments systems. This would not entail new facilities, only an extension of the banks' processing and communications base to the airlines.

McKesson was an early exploiter of this opportunity. Its initial innovation was to link its customers, small pharmacies, to its purchasing, distribution, and inventory management systems. It has since added insurance claims processing. McKesson had the reach into the pharmacies that the insurance providers did not and extended its range to add claims processing as a by-product to its mainstream transactions. It is now the third largest claims processor in the United States. All that business once belonged to insurance companies.

Firms must assume that, just as their partners may come from outside their industry, so too may their competitors. The IT platform must ensure defensive capability against third parties.

Policy 6: Vendor Staying Power

We will not be dependent on "brochureware" nor on vendors that will not be able to stay the course.

"Brochureware" refers to the claim that "we have it all: OSI, integration-you name it." Many previously strong IT vendors lost ground badly during the past decade because they either failed to provide extensive reach and range in their products or lacked the R&D base and funds to move with the mainstream of the industry toward integration and implementation of key standards. Wang, Data General, and Prime are examples of vendors that have lost their position as strategic suppliers for these reasons.

Early 1980s leaders in data base management systems have similarly been unable to broaden their products to accommodate large firms' current needs for reach and range. Software providers in the personal computer field have experienced increasing problems and delays in upgrading their products to handle a plethora of telecommunications and multi-vendor systems. Cullinet's fall from leadership to acquisition and dismemberment is an early signal that product features are not enough; proven ability to move with the mainstream of standards and integration is now essential. This is beyond the capabilities of all but a few vendors, but it is in the brochureware of many. Selecting standards that will help a firm evolve a truly integrated platform means selecting vendors that can stay the course.

Policy 7: Comparable International Capability

The preceding policy requirements will be applicable in an international context.

It is vital that firms with an international strategy for competing in a global marketplace have an international IT strategy and an IT platform that extends worldwide. Very few firms have such a strategy, and very few vendors can fully support it yet. International telecommunications is apolitical, regulatory, and technical minefield. In the 1990s no major firm will be able to operate efficiently in a global environment unless it has a global IT platform that ensures international reach, which means adopting international telecommunications standards, and sufficient range to support information-sharing across all functions and locations.

Reuters stands out in having created a worldwide market through its platform. The Economist summarizes the firm's accomplishments.

Reuters has effectively become the global market for currencies, carrying price-bending news as well as banks' quotations.... [After 1981] some Reuters terminals allowed dealers to execute transactions with other subscribers-a kind of souped-up telex.
The new system, named "Dealing 2000" . . . will allow traders around the world to enter orders into the system, which will then match them automatically.⁷

Multinationals will increasingly need to create their own equivalents of Reuters' global information and communications delivery base.