WISDOM FROM WHARTON

When investing in information technology, it is essential to know “when” and “how.”

November 1 1989 by Chief Executive


Information systems (IS) can be strategic and can affect the ways in which firms compete. In some instances, IS can confer competitive advantage. Even the most important and strategic information systems, however, do not always grant advantage. Often the intended benefits do not materialize, or cannot be sustained because of competitor actions.

This will not surprise an economist, of course, nor a chief executive. Benefits from an innovation are notoriously difficult to protect since imitation by competitors often leads to competition, and then to the passing of most benefits to customers.

Some applications will benefit an early innovator, and should be done quickly; some will not. Some systems have competitive implications; others do not. Which systems should be developed? When? How aggressively should development be pursued? Studies show how a CEO’s beliefs about the profitability of his firm before and after innovation given his beliefs about competitors’ actions-can lead to successful strategies for selecting among opportunities for innovative application of technology.

One of the industries that was most dramatically altered by information technology (IT) is drug wholesaling: Electronic order entry is used by almost 100 percent of some wholesalers’ customers, and these electronic orders are used to drive warehouse picking, vehicle loading and routing, and wholesaler reorders. During the period in which these systems were being developed-at enormous expense the number of whole-salers was reduced nationally by more than 50 percent. The benefits to customers-better service, reduced inventory and wholesale prices-were crucial to the survival of these customers as independents in their competition with chain stores.

Yet few if any of the remaining major distributors have gained advantage over one another since all have systems that are largely equivalent. The major player in the industry saw its market share remain flat at 20 percent during the period of rapid adoption of electronic order entry. Still, no one in the industry would deny the strategic impact these electronic order entry and distribution systems have had on customers or on the structure of the industry itself: Profitability of independent pharmacies and drug stores has improved. Although attrition in the industry continues, the rate of exit has slowed. And of course, the continued existence of independents-their principal customers-is of fundamental importance to drug wholesalers.

IS may result in significant changes in the structure of an industry, as in drug distribution, and may be essential to remaining competitive within an industry, as is true with automatic teller machines (ATMs), but these systems may simply not have competitive impact. The decision to invest in innovative applications of IT is among the most difficult and most important decisions facing U.S. firms; it remains poorly understood by CEOs and chief information systems officers alike. It is difficult for many executives to determine which applications offer competitive impact. Not all good ideas can be developed at once, so it is necessary to set priorities. The competitive impact is influenced by competitors’ actions, which complicate the setting of priorities and timing.

 

UNIQUE CHARACTERISTICS OF IS

IS efforts are often characterized by enormous fixed costs and low or negligible variable costs, creating the possibility of important scale advantages for the first player to capture significant market share. Additionally, the fixed costs are often sunk costs, which implies significant cost of failure.

It has been extraordinarily difficult to protect IS innovations through patents, trade secrets or proprietary technology. This is evidenced by the widespread duplication by brokerage houses of customer asset management accounts, despite Merrill Lynch’s patent on its innovative Cash Management Account (CMA), by essentially equivalent systems for customer electronic order entry in drug wholesaling, and by largely equivalent ATM services in retail banking. The difficulty of protecting innovations significantly increases risks for the first developer, and perhaps increases the attractiveness of a strategy based on imitation. The difficulty of protecting innovation also limits the licensing fees that can be charged to imitators, as Merrill Lynch discovered with its CMAs.

Scale advantages would argue for first mover effects and early innovation. The difficulty of protecting innovation from imitators and the cost of failure would argue against innovation, early or late, except as a defensive move. However, the benefits that an IS innovation can present to the entire industry will sometimes argue for innovation and widespread adoption.

 

SIX OPTIONS

A company developing an IS infrastructure and strategy can choose to be a technology leader or follower. Additionally, and separately, the company can choose to lead or follow as an independent player or as a member of a development consortium (see figure). As the special nature of IS’s infrastructure innovations becomes clear, more IS applications will appear best undertaken as regional or industrywide consortia, rather than as individual competitive efforts. Benefits of consortium development include reduced costs for suppliers and increased service quality for customers; reduced system development costs; and increased benefits resulting from larger networks, as in the ATM network. For example, all members of an ATM network benefit from the number of ATMs available. Economists call such benefits “network externalities.”

The beliefs of the CEO or decision maker will be critical in selecting a strategy. Not surprisingly, the decision must include comparison of profits. But remember that the status quo may vanish rapidly if competitors choose to pursue the innovation; analysis must therefore consider implications for profits if the CEO chooses not to innovate, and then discovers that competitors have innovated and successfully preempted the market.

Similarly, a firm about to innovate should evaluate how important timing will be: a decision maker who believes that profits will increase if he innovates should attempt to ascertain if profits will still increase if he follows with later development. More importantly, perhaps, the firm should consider potential competitive impact. If benefits can be defended somehow, the innovation may prove competitive. But if the innovation is likely to be available to all players and to benefit innovator and imitator alike, consortium development may be most reasonable.

Profits for the firm and its competitors cannot be known with precision, and it may be impossible to determine if or when a competitor intends to innovate. However, the CEO ultimately must make a decision. The following analyses outline what the decision maker should believe-in subjective, imprecise terms-to justify a given decision. If he doesn’t agree with the assumptions used to justify a strategy, he may find that his decision ultimately was incorrect.

Prior independent leader. A CEO choosing to develop early and independently, if his decision is rational, should believe the following: (1) The firm’s profits will increase due to the innovation; (2) the firm has the capability to develop the innovation independently; and (3) the firm will be able to protect the benefits of early introduction of the innovation, and thus will earn greater returns than a follower imitating the development. The innovator should also believe that the innovation has the potential to provide competitive advantage.

Experience to date shows that competitive advantage through IS, like competitive advantage anywhere, is difficult. It requires that the offering be attractive, or it will be ignored in the marketplace and will fail. For those offerings that prove attractive, it also requires significant barriers to competitors’ copying, or early and successful preemption of the market. Successful examples include Merrill Lynch’s CMAs, and the travel agent reservation systems of United and American Airlines. From 1984 until the present, these reservation systems have been operated as businesses, earning hundreds of millions of dollars, with profit margins of a magnitude higher than generally experienced in the primary airline business; moreover, the large majority of the revenues from these systems now come not from travel agents but from competitor airlines, which are charged for fare listings, reservations made, and tickets written.

Less successful examples of systems intended to provide competitive advantage, and the reasons for their limited competitive impact include: (1) commercial bank ing Treasury Work Stations-workstations failed to buy market share or provide customer loyalty; too many offerings, chasing too few customers made offerings unprofitable; (2) retail banking ATMs-slow retail customer adoption in most markets, combined with rapid competitive deployment of nearly identical offerings, limited competitive impact in most cases; ATMs are now widely offered and widely used, but in only a few instances did they confer lasting advantage; and (3) retail electronic home banking-these systems provide little value to most customers at present, and are generally poorly received in the marketplace.

Analysis becomes more complex when the CEO or decision maker holds some, but not all, of the beliefs that motivate the independent leader strategy. For example, a CEO who believes that a potential innovation offers competitive advantage, but who also knows that the firm lacks the resources needed to implement it successfully, may consider establishing a joint venture to complete development. Unfortunately, these often increase the probability of innovation details leaking to competitors, or of a partner in the joint venture emerging as a competitor, and thus may decrease the competitive lifetime of the innovation.

Alternatively, when joint ventures are inappropriate, and the competitive application cannot be attempted with the firm’s current resources, acquisition by or sale of equity to a large company can provide the necessary additional resources. For example, most London securities houses believed that technology was going to be an essential element to their survival after Big Bang (the rapid deregulation of the London Stock Exchange in October 1986) and that trading systems could offer competitive advantage; few firms had the financial resources needed to invest in these trading support systems. This is widely believed to be one of the factors leading to linkages between London market makers and British, U.S., and European banks.

Prior consortium leader. A CEO or decision maker choosing to develop early, but as a member of a common development consortium, should believe the following: (1) profits will be better for his firm and for players in the industry that develop the innovation; (2) there will be little or no opportunity to benefit against competitors by developing early, reducing the attractiveness of independent development; and (3) lower development costs, reduced costs of service delivery due to shared scale advantages, network externalities, or other benefits will accrue to consortium members.

Regulators are permitting early consortium strategies. The recently announced European common credit card and EFTPOS (electronic funds transfer point of sale) network, to be shared by all banks, is an example. It is intended to offer more than 100 million cardholders-in the 17 countries represented by the European Council for Payment Systems -common access to cash services at 15,000 branches and at ATMs, and to allow travelers cashless payment at shops and petrol stations. The Council expects competition among hanks to he as intense as ever.

Prior independent follower. Given that the decision maker has chosen to be a late developer and that the firm still elects to develop independently, the decision maker should believe the following: (1) although the firm chooses not to innovate first, it still expects its profits to increase, which is a reason for innovating eventually; (2) costs decrease over time, which is a reason for delaying development; (3) reduction in development cost due to waiting will exceed decrease in profits due to first development; (4) there will be limited benefits resulting from membership in a consortium; or (5) even though the firm chooses not to be the earliest developer, it believes that it is possible to gain advantage over other non-innovators or late developers.

For systems that will ultimately be commodities, waiting until necessary software is available from third-party vendors may be an attractive option for many firms. Additionally, there is evidence of late competitors, operating solo, gaining stronger negotiating position when joining a consortium.

For example, Provident National Bank developed an ATM network larger than any bank in the Money Access Center (MAC), the regional ATM consortium operated by Philadelphia National Bank. Provident used the strength of its network to negotiate better terms than those available to any other MAC participant at that time. However, firms selecting in advance to be followers, and still to develop independently, are following a risky strategy. It makes sense for nonstrategic applications, and for applications where the needs of the firm are idiosyncratic; otherwise participation in a development consortium appears more plausible.

Prior consortium follower. A CEO choosing to be a follower and to develop an innovation as part of a consortium should believe the following: (1) profits of the firm and those of other players adopting the innovation will increase; (2) costs will decrease over time; (3) all players are fairly equal with regard to exploiting the innovation, and it will be difficult for one late adopter to gain competitive advantage over another by imitating it, reducing arguments against cooperation; (4) there are benefits, like network externalities, that accrue to all players in the consortium; thus consortium members will enjoy benefits not available to nonmembers, greatly strengthening arguments in favor of cooperation; (5) the initial developer cannot preempt the marketplace and there will be limited benefits for developing early.

There is limited data with firms or industries selecting this strategy. Most late consortium development seems to be in response to moves by a single dominant competitor.

Ex post independent follower. Sometimes a corporation will be blindsided. A competitor will introduce an unanticipated product or service, and the innovation will require a response. A corporate decision maker who may counter through a late independent development effort should share many of the beliefs of a decision maker who chooses in advance to pursue the independent follower strategy, without waiting for a competitor’s innovation. Additionally, the decision maker should believe the following: (1) a competitor choosing to imitate must believe that the firm’s profits will be greater after duplication than before. Thus, adoption may be forced on the firm as a defensive measure by a competitor’s moves; (2) a firm choosing to forgo shared consortium development as a competitive response should believe there is little benefit to be obtained by reducing costs or development time through shared development, and that competition through the innovation will be advantageous to the firm.

In the late 1970s, United Airlines matched American Airlines’ Sabre reservation system with their Apollo system. This action indicates there may have been room for a second, early competitor when the market was not yet preempted. ‘TWA’s more recent experience with PARS and Texas Air’s experience with marketing System One indicate the difficulty of independent follower responses to established competitors, especially when mechanisms are in place to protect market share by imposing significant switching costs on customers defecting to competitors.

In most externally focused applications aimed at controlling a distribution channel, and in most applications with large-scale advantage, there appears to be little benefit to the independent follower strategy. Earlier players will benefit from progress down the learning curve and the early capture of market share. If the market can be preempted, it will already have been preempted; if it cannot be preempted, there is probably little benefit to the solo development strategy. An exception to this is late acquisition of standard or commodity applications by purchase from a systems vendor.

Ex post consortium follower. The CEO should share many of the beliefs of a decision maker who chooses the consortium follower strategy in advance. In particular, he should believe that (1) there is a clear benefit to duplicating the competitors’ innovation; (2) there is little or no benefit in attempting to develop a competitive response independently, and there will be little or no differential competitive impact fot players adopting after the initial innovator; and (3) there are benefits, like network externalities, that will accrue to all players in the consortium.

This appears to be a rational response to an attempt by a strong competitor to preempt and dominate the market. The development of the New York Cash Exchange (NYCE) ATM network in New York can be seen as a response to Citibank’s ATM network. (NYCE is a shared ATM network owned jointly by all major New York metropolitan area banks except Citibank.) Similarly, Shawmut and Bank of New England’s ATM consortium (Yankee) can be seen as a response to Bay Banks’ success with its proprietary ATM network in Boston.

 

COMPETITION OR COOPERATION?

Despite cultural norms that strongly favor competitive development, an increasing awareness of the difficulty of competing with IS innovation and of the various benefits available through shared development should make consortium development more common. This requires a shift in management attitude. It may also require regulatory or public policy changes in some industries.

The independent leader strategy appears most conducive to competitive use of technology. But given the high level of resources frequently required for technology programs, some attractive opportunities may not readily be pursued; moreover, this strategy is often quite risky.

How can a decision maker obtain the benefits of consortium development-reduced resource requirements and lower risk-without forgoing the competitive possibility of independent development? Several approaches can be employed: a pseudo-consortium, or almost a cartel, can be formed. Banks from different regions of the U.S. which cannot compete under current banking regulations may be able to get together to share consortium development.

A joint venture can he formed, often with a player outside the industry seeking to employ the innovation. Unfortunately, one player will often have incentives to leak proprietary aspects of the development; this reduces the period of time in which a developer can expect to obtain advantage.

Linkages, mergers, acquisitions and partial acquisitions all have proved valuable approaches in obtaining the benefits of consortium development. Mergers may give a player the scale needed to pursue or benefit from an innovation. Acquisitions and partial acquisitions offer the benefits of joint ventures; often the exchange of equity is sufficient to reduce significantly the risk of leakage of proprietary information.

Finally, it is interesting to note how closely tied the choice of competition or cooperation is to national and cultural norms. Japanese faculty colleagues say we place far too much emphasis on competition. British consultants, on the other hand, say we underestimate the importance of the independent leader strategy, and the importance their banking clients place on competition and technology leadership.

Stereotypes aside, it is clear that in today’s brutally competitive economic environment, gratuitous attempts to gain advantage, and redundant development efforts, are luxuries that U.S. business can no longer easily afford. Despite our preference for competitive actions, it may be advisable to increase our willingness to engage in cooperative development.

Sometimes there are enormous benefits to be gained by winning a technology development race; sometimes the innovator cannot appropriate the benefits of getting an IS product or. service to market first. An understanding of this fact, and of the conditions necessary to capture benefits, is essential to setting an IS strategy. It is not always essential to lead, or to win. It is essential to know when to lead, when to follow, and when to go your own way.


Eric K. Clemons is Associate Professor of Decision Sciences at The Wharton School of the University of Pennsylvania.