Long-term sustained management of a strategic alliance is turning out to be the dominant challenge of effective IT outsourcing. From a relatively unusual entrepreneurial activity, IT outsourcing has recently exploded across the global corporate landscape.1 Xerox, Delta Airlines, AMP Insurance (Australia), British Aerospace, and the Inland Revenue Service are the latest of these mega-alliances. Several years ago, Shell Oil out-sourced its Brazilian IT activities. Like marriage, however, these arrangements are much easier to enter than to sustain or dissolve. The special economic technology issues surrounding outsourcing agreements necessarily make them more complex and fluid than an ordinary contract. Both parties need to make special efforts for outsourcing to be successful. In addition to clear successes, we have identified troublesome relationships and several that had to be terminated.

Our purpose in this article is to provide a concrete framework to help senior managers think about IT out-sourcing and focus on how to manage the alliance to ensure its success.

Why Outsourcing Alliances Are Difficult

Outsourcing contracts are structured for very long periods of time in a world of fast-moving technical and business change. Ten years is the normal length of a contract in an environment in which computer chip performance is shifting by 20 percent to 30 percent per year. (This standard contract length has emerged to deal with switching cost issues and to make the economics work for the outsourcer.) Consequently, a rigid deal that made sense at the beginning may make less economic sense three years later and require adjustments to function effectively.

Exacerbating the situation is the timing of benefits. For the customer, the first-year benefits are clear; usually the customer receives a one-time capital payment. Next, the customer feels relieved to shift its problems and issues to another organization. Finally, the tangible payments in the first year occur in an environment in which the outputs most closely resemble those anticipated in the contract. In each subsequent year, the contract payment stream becomes less and less tied to the initial set of planned outputs (as the world changes) and, thus, more subject to negotiation and misunderstanding.

The situation from the outsourcer’s perspective is just the reverse. During the first year, there is a heavy capital payment followed by the extraordinary costs for switching responsibility and executing the appropriate cost-reduction initiatives. All this is done in anticipation of a back-loaded profit flow. At precisely the time the outsourcer is finally moving into its earnings stream, the customer, which may feel the need for new services, is chafing under monthly charges and anxious to move to new IT architectures. If the customer has not had experience in partnering activities before, extraordinary tensions in the relationship can develop.

A further complication is the fact that only a few outsourcers have the critical mass and access to capital markets to undertake large contracts. Electronic Data Systems (EDS), Computer Sciences Corporation (CSC), IBM, and AT&T make up the largest part of the current market. A much larger group of firms specializes in certain niches in the outsourcing market, fulfilling either small contracts or specific subfunctions such as network operations. If an alliance is not working out, a company has limited options for resolving the situation, particularly because outsourcing is relatively easy, but insourcing again is very difficult. One international oil company transferred the relationship to another outsourcer.

Finally, the evolution of technologies often changes the strategic relevance of IT service to a firm. From the customer’s viewpoint, assigning a commodity service to an outsider is very attractive if the price is right. Delegating a firm’s service differentiator, however, is another matter. The customer that made the original decision based on efficiency will judge it differently if using effectiveness criteria later.

Outsourcing in Retrospect

IT outsourcing has been around for a long time. In the mid-1960s, for example, computer services bureaus ran a variety of programs. These applications were focused heavily in the financial and operations support areas (general ledger, payroll, inventory control, and so on). The programs were both customized and general purpose, in which the individual firm had to accommodate its operations to the standard options in the package. The customers of the service bureaus were mostly small and medium-size firms, although some large firms used them for specialized needs or highly confidential items like executive payroll.

A good example of a provider in this industry then and now is ADP. ADP, which began as a small punch-card payroll company in 1949, grew to a $3 billion organization by specializing in large-volume, standard transaction-type activities, such as payroll and handling proxy solicitations (almost 100 percent of the industry). Software contracting companies like Andersen Consulting in the private sector and CSC in the public sector developed large turnkey applications for firms that required either specialized staff or a large number of staff people, which the firm found was inconvenient, imprudent, or impossible to retain. EDS, in the state and local government sector, provided full outsourcing for organizations whose cultures and salary scales made it impossible to attract people with the necessary skills in a competitive job market. These were the exceptions, however, to the general trend of developing IT in-house. Until 1990, the major drivers for outsourcing were primarily:

  • Cost-effective access to specialized or occasionally needed computing power or systems development skills.
  • Avoidance of building in-house IT skills and skill sets, primarily an issue for small and very low-technology organizations.
  • Access to special functional capabilities.

Outsourcing during this period was important but, in retrospect, largely peripheral to the main IT activities that took place in midsize and large organizations.

Kodak’s decision in 1989 to outsource IT was the first real wake-up call for complacent IS (information systems) managers, now more commonly called CIOs (chief information officers).2 Kodak’s then CIO, who had been a general manager rather than a computer professional, took an aggressive position in outsourcing mainframes, telecommunications, and PCs. Until then, outsourcing for medium-size to large companies had been mostly a sideshow, and outsourcing was generally something reserved for small and medium-size companies with problematic, grossly mismanaged IS departments. After Kodak’s decision, there was a flurry of oversubscribed IS conferences on outsourcing at which the Kodak CIO was often the featured speaker. We both attended a number of these conferences and independently witnessed the hostility from many CIO participants (who perceived outsourcing as a terrifying threat to their status quo) toward the Kodak CIO as she explained her rationale.3 Even today, many of those very same CIOs who attended the conferences quickly point out with some relief that only one of the three original Kodak outsourcing contracts was totally problem free — albeit the far largest one (although all three contracts are still in place).

Outsourcing in the 1990s

We have conducted more than four years of case research on the Kodak, General Dynamics, and over a dozen other outsourcing situations and have concluded that IT outsourcing is not a “flash in the pan” management fad.4 IT outsourcing is a harbinger of the transformation of traditional IT departments and provides a glimpse at the emerging organizational structures of the information economy. Our research indicates that more than half of midsize to large firms have outsourced or are considering some type of outsourcing of their IT activities. This phenomenon has not occurred only in the United States; for example, in 1993, AMP Insurance Company (the largest insurance company in Australia), British Aerospace, and the U.K. Inland Revenue Service all outsourced substantial parts of their IT activities.

Two factors have affected the growth of IT outsourcing — the recognition of strategic alliances and the changes in the technological environment.

Acceptance of Strategic Alliances.

The value of strategic alliances has been widely recognized.5 Interrelated forces underlie the creation of alliances. On one level, finding a strong partner to complement an area of weakness gives an organization an island of stability in a turbulent world. It is difficult to fight simultaneously on all fronts, and alliances allow a company to simplify its management agenda safely. Alternatively, alliances allow a firm to leverage a key part of the value chain by bringing in a strong partner that complements its skills.6 Such a partner may create an opportunity to innovate synergistically, in which the whole is greater than the sum of the parts. Also, successful early experiences with alliances increase a firm’s confidence in undertaking new alliances in other parts of the value chain as a profitable way to do business. This experience gives the firm insight into how to increase the likelihood of a successful alliance.

It is worth noting that, for an alliance to be successful and endure for the long term, both firms must believe that they are winners. Because of the synergistic potential of the relationships and the opportunity to specialize, both firms should legitimately feel that they are benefiting from the alliance. This is not a zero-sum game.

IT’s Changing Environment.

The stages theory of computer growth describes how companies have assimilated computers over extended organizational learning periods (fifteen to twenty years).7 Each period has a target market focus in the firm and a particular model for application (see Table 1). The DP era (1960–1980) focused on transaction processing systems by automating the existing manual systems. The micro era (1980–1995) focused on leveraging professional workers (e.g., engineers, financial analysts, and managers) by using the computer to do analytic computations, access data, and print various renditions of drawings or graphics. During both of these eras, the use of the computer was primarily within the company.

Overlapping the micro era is an emerging external market where the computer is embedded into products (e.g., automobiles) and services (smart cards). By and large, outsourcing is not having an impact on this market; when General Dynamics outsourced its IT activities, it retained the 800 systems people doing this type of work.

The network era (1990–?), currently overlapping the micro era, has emerged from the fusion of computers and telecommunications technologies.8 In this era, firms are integrating both internal and external computers so they can change their structure to more efficient forms for competing flexibly in the global marketplace. We call this the IT-enabled network firm. This integration is putting extraordinary pressures on firms that are trying to keep the old services running while developing the interconnections and services demanded by the new environment. Thus outsourcing has become a viable alternative for those firms to get access to appropriate skills and to speed up the transition reliably and cost effectively.

We should note that, as shown in Table 1, the development of most of the code that companies now use has already been outsourced. A distinct minority of the code in operating systems, e-mail systems, word processing packages, and spreadsheet software was actually developed within the firm (with a much smaller percentage expected in the future). This trend, which occurred for the obvious reasons of economies of scale and scarcity of competent staff, will only continue. Currently, Computer Associates, Lotus, IBM, Borland, and Microsoft are the de facto software providers to most companies. The internal IT organization is already a selector of code rather than a developer.

At the same time, many firms have a residue of fifteen- to thirty-year-old systems primarily written in COBOL and PL/1. Although this problem is particularly acute in the financial services industry and manufacturing, it is not confined to it. The cost-effective transformation of these systems to the client/server model (a key technology of the network era that separates the management of files and their integrity onto one machine, the server, from devices accessing these files, the clients) is an enormous challenge. On the one hand, firms are looking for low-cost maintenance of the old systems to ensure they operate reliably, while gaining access to the new skills to permit their transformation to the new model. This shift is as significant today as the move from tabulating equipment thirty-five years ago in terms of providing new capabilities to the firm. A number of organizations see outsourcing as a way of bringing the appropriate specialized skills to this task.

What Drives Outsourcing?

The mix of factors that raises the possibility of outsourcing varies widely from one company to another.9 In our research, we have uncovered a series of themes that, in aggregate, explain most of the pressures to outsource.

  • General Managers’ Concerns about Costs and Quality. The same questions about IT costs and response times came up repeatedly when we talked to managers: Can we get our existing services for a reduced price at acceptable quality standards? Can we get new systems developed faster? We have uncovered the following ways an outsourcer can save money for a customer:
  • Tighter overhead cost control of fringe benefits. On balance, the outsourcers run much leaner overhead structures than many of their customers.
  • More aggressive use of low-cost labor pools by creatively using geography. Frequently, the outsourcer moves data centers to low-cost areas (modern telecommunications make this possible).
  • Tough world-class standards applied to the company’s existing staff, all of whom have to requalify for appointment at the time of outsourcing. Frequently, employees may have become lazy or are unskilled in leading-edge IT management practices.
  • More effective bulk purchasing and leasing arrangements for all aspects of the hardware/software configuration, through discounts and better use of capacity.
  • Better management of excess hardware capacity. The outsourcer can sell or utilize underused hardware that would otherwise be idle by combining many firms’ work in the same operations center. One small firm’s on-line operations (a $27 million ten-year contract) was transferred to a larger data center at no extra cost to the outsourcer. Capacity was simply better used.
  • Better control over software licenses, through both negotiation and realistic examination.
  • More aggressive management of service and response time to meet, but not wildly exceed, corporate standards. Tighter control over inventories of paper and other supplies.
  • Simply hustling. Outsourcers are professionals; this is their only business, and their success is measured by satisfied customers who recommend them to others, by bottom-line profitability, and by stock market performance.
  • The ability to run with a leaner management structure because of increased competence and critical mass volumes of work.
  • Creative and more realistic structuring of leases.

While the cumulative impact of these items can be very significant, we issue a few cautionary notes. Until several knowledgeable bidders have closely analyzed your existing operation to propose an alliance, you don’t have a true picture. An IT efficiency study funded by the IT department and done by a consulting company hoping to get future business is simply inadequate. And equally important is assessing whether the outsourcer can rapidly mobilize its staff for the quick-response development jobs needed to get products and services to market much faster.

Breakdown in IT Performance.

Failure to meet service standards can force general management to find other ways of achieving reliability. As we reflect on the past thirty years of computer growth in most companies, it is not atypical to find a company in which cumulative IT management neglect eventually culminated in an out-of-control situation from which the current IT department could not recover. For example, Massachusetts Blue Cross and Blue Shield’s decision to outsource to EDS was triggered by the failure of three major systems development projects (and losses in the tens of millions of dollars). It saw outsourcing as a way to fix a broken department. Similarly, a midsize bank’s interest in out-sourcing came after a one-day total collapse of its ATM network, caused by faulty, internally designed software patches.

An additional driving factor is the need to rapidly retool a backward IT structure to maintain its competitiveness. In one firm, general managers thought (correctly in our judgment) the internal IT culture was both frozen and backward; it needed to leap forward in performance. The general managers, who lacked both the time and the inclination to personally undertake the task, found outsourcing a good choice for making the transition from the DP era to the network era.

Intense Supplier Pressures.

Kodak’s decision to out-source its data center and telecommunications to IBM, DEC, and Businessland was a flash point. Suddenly, all general managers saw outsourcing as a highly visible, if often misunderstood, alternative. At the same time, IBM and DEC were looking for new value-added services to reach their customer bases and compensate for declining hardware margins and sales. They moved aggressively into the field with expanded and highly energetic sales-forces. EDS, the largest firm in the field, used its General Motors operations center to demonstrate its expertise. CSC, strong in the federal sector, built a bridge to the commercial sector with its General Dynamics contract. The visibility of these and other arrangements, combined with the suppliers’ aggressive salesforces, enabled them to approach general managers with compelling reasons to outsource.

Simplified General Management Agenda.

A firm under intense cost or competitive pressures, which does not see IT as its core competence, may find that outsourcing is a way to delegate time-consuming, messy problems so it can focus scarce management time and energy on other differentiators. If the managers perceive the outsourcer as competent and are able to transfer a noncore function to reliable hands, they will not hesitate to choose outsourcing. These activities must be done respectably, but long-term upside competitive differentiation does not come from executing them in an outstanding fashion.

Financial Factors.

Several financial issues can make outsourcing appealing. One is the opportunity to liquify the firm’s intangible IT asset, thus strengthening the firm’s balance sheet and avoiding a stream of sporadic capital investments in the future. An important part of many of the arrangements struck in the past two years has been the significant upfront capital paid for both the real value of the hardware/software assets and the intangible value of the IT systems. General Dynamics, for example, received $200 million for its IT asset. Publicly held outsourcers that have access to the capital markets have pushed this; partnerships like Andersen Consulting do not have such access.

Outsourcing can turn a largely fixed-cost business into one with variable costs; this is particularly important for firms whose activities vary widely in volume from one year to another, or which face significant downsizing. The outsourcer can make this change much less painfully than the firm, broker the slack more effectively, and potentially provide greater employment stability for the company’s IT employees who are there because of their ability to handle multiple operations. In fact, outsourcing has been very positively received by the staffs at several of the firms we studied. They saw themselves leaving a cost-constrained environment with limited potential for promotion and entering a growth company where IT was the firm’s only business. In variable cost arrangements, price deescalation clauses should be negotiated in the sections of the contract that deal with IT hardware costs, rather than inflation protection clauses, because of the dramatic downward changes in the technology costs.

Finally, a third-party relationship brings an entirely different set of dynamics to a firm’s view of IT expenditures. It is now dealing with a hard-dollar expenditure that all users must take seriously (it is no longer soft-dollar allocation). There is a sense of discipline and tough-mindedness that an arm’s-length, fully-charged-out internal cost center has trouble achieving. Further, firms that do not see IT as a high leverage function may perceive outside professionals as adding special value and, hence, as quite influential.

For a firm considering divestiture or outright sale of one or more of its divisions, outsourcing has special advantages. It liquifies and gets value for an asset that is unlikely to be recognized in the divestiture. It gives the acquirer fewer problems to deal with in assimilating the firm. And the outsourcing contract may provide the acquirer a very nice dowry, particularly if the firm is small in relation to the acquirer. The contract can be phased out neatly, and the IT transaction volume can be added to the firm’s internal IT activities with little or no additional expense. In several midsize banks we studied, this was the guiding rationale for outsourcing. It gave them access to reliable IT support, while making their eventual sale (which they saw as inevitable) more attractive from the acquirers’ viewpoint.

  • Corporate Culture. The company’s values can make it very hard for managers to do the right thing. For one firm we studied that had several internal data centers, there were obvious and compelling advantages to consolidating them. The internal IT department, however, simply lacked the clout to pull off the centralized strategy in what was a highly decentralized firm, built up over the years by acquisitions. The firm saw the decentralized culture as a major strength, not subject to reconsideration. Outsourcing, driven by very senior management, provided the fulcrum for overcoming this impasse since it was not directly associated with any division or corporate staff. Similarly, an internal IT organization may fall behind the state of the art without being immediately attacked, while an outsourcer is forced to keep up with the latest technology to be successful.
  • Elimination of an Internal Irritant. No matter how competent and adaptive existing IT management and staff are (and usually they are very good), there is usually tension between the end users of the resources and the IT staff. Often this is exacerbated by the different language IT professionals use, lack of career paths for users and IT staff across the organization, perceived high IT costs (often inaccurate), perceived unresponsiveness to urgent requests (but that sometimes are not really urgent), and perceived technical obsolescence (some recent college graduate is always discovering a new software package that will apparently solve all problems like magic). In this context, the notion of a remote, efficient, experienced outsourcer is particularly compelling.

Other Factors.

We found a variety of other drivers for outsourcing in specific situations. Some companies with a low-technology culture appeared to have trouble attracting and retaining high-technology IT staff. Out-sourcing offered a way to gain these skills without getting involved in the complex management issues they were not skilled in and did not want to manage.

A midsize high-tech firm needed to develop and run a series of critically important applications. Outsourcing gave it access to skills it could not attract to its organization. Managers felt that outsourcing had substantially reduced their corporate risk while giving them needed access to specialized knowledge.

One large organization felt it was getting a level of commitment and energy that would have been difficult to gain from an in-house unit. In its outsourcing arrangement, it was an external reference for the outsourcer, which was growing rapidly. Good performance on the contract was critical for the outsourcer to get the kind of market growth it wanted.

Still another firm was frustrated by its inability to get its products to market faster. Its in-house resources, limited in size and training, were simply not moving fast enough. Outsourcing gave it a boost of adrenalin to build the IT infrastructure for a two-thirds improvement in time to market.

When to Outsource IT

When do the benefits of outsourcing outweigh the risks? Our research suggests that there are five factors that tip the scale one way or the other. Each factor is fundamentally linked to the basic research models in the IT field.10

Position on the Strategic Grid

The strategic grid is a framework that shows IT’s strategic relevance to a company at a particular time.11 It assesses two dimensions: (1) the company’s current dependence on computers or networks, and (2) the future importance of computer applications under development. In Table 2, we have generalized the two dimensions. First is the degree of the company’s current dependence on information — that is, the extent to which business operations are dependent on reliable, real-time information processing. Second is the future importance of innovative information resource management — that is, the extent that future competitive position is dependent on known initiatives requiring innovative computer applications.

Outsourcing operational activities is generically attractive, particularly as the budget grows and the contract becomes more important to the outsourcer. The more the firm is operationally dependent on IT, the more sense outsourcing makes. The bigger the firm’s IT budget, however, the higher in the customer organization the decisions will be made, and thus the more careful the analysis must become. At the super-large scale, the burden falls on the outsourcer to show it can bring more intellectual firepower to the task.

When the application’s development portfolio is filled with maintenance work or projects, which are valuable but not vitally important to the firm, transferring these tasks to a partner holds few strategic risks. However, as the new systems and processes increasingly come to deliver potentially significant differentiation and/or massive cost reduction, the outsourcing decision comes under more and more scrutiny, particularly when the firm possesses a large, technically innovative, well-run IT organization. The potential loss of control, loss of flexibility, and inherent delays in dealing with a project management structure that cuts across two organizations become much more binding and of greater concern. There are examples, like General Dynamics, where outsourcing was successful, but it was more the exception than the rule.

As shown in Table 2, for companies in the support quadrant, the outsourcing presumption is yes, particularly for the large firms. For companies in the factory quadrant, the presumption is yes, unless they are huge and perceived as exceptionally well managed. For firms in the turnaround quadrant, the presumption is no; it represents an unnecessary, unacceptable delegation of competitiveness. For companies in the strategic and turnaround quadrants, the presumption is no; not facing a crisis of IT competence, companies in the strategic quadrant will find it hard to justify outsourcing under most circumstances. Subcritical mass in potentially core differentiating skills for the firm, however, are one important driver that might move such a company to consider outsourcing.

For larger multidivision firms, this analysis suggests that various divisions and clusters of application systems can legitimately be treated differently (i.e., strategic differentiated outsourcing). For example, an international oil company outsourced its operationally troublesome Brazilian subsidiary’s IT activities while keeping the other items in-house. Similarly, because of the dynamic nature of the grid, firms under profit pressures after a period of sustained strategic innovation (in either the transforming or strategic quadrants) are good candidates for outsourcing as a means to clean up their shop and procedures. This was true for one large high-technology organization that saved over $100 million by outsourcing.

Development Portfolio

The higher the percentage of the systems development portfolio in maintenance or high-structured projects, the more the portfolio is a candidate for outsourcing.12 (High-structured projects are those in which the end outputs are clearly defined, there is little opportunity to redefine them, and little or no organizational change is involved in implementing them.) Outsourcers that have access to high-quality, cheap labor pools (in, for example, Russia, India, or Ireland) and good project management skills can consistently outperform, on both cost and quality, a local unit that is caught in a high-cost geographic area and lacks the contacts, skills, and confidence to manage extended relationships. The growth of global fiber-optic networks has made all conventional thinking on where work should be done obsolete. For example, Citibank does much of its processing work in South Dakota, and more than 50,000 programmers are working in India on software development for U.S. and European firms.

High-technology, highly structured work (i.e., building a vehicle tracking system) is also a strong candidate for outsourcing, because the customer needs staff people with specialized, leading-edge technical skills. These technical skills are widely available in countries such as Ireland, India, and the Philippines.

Conversely, large, low-structured projects pose very difficult coordination problems for outsourcing. (Low-structured projects are those in which the end outputs and processes are susceptible to significant evolution as the project unfolds.) Design is iterative, as users discover what they really want by trial and error. As noted earlier, this work requires that the design team be physically much closer to the consumers, thus eliminating significant additional savings. The work can, of course, be outsourced, but it requires more coordination to be effective than the projects we just described. One firm outsourced a large section of this work to a very standards-oriented outsourcer as a way of bringing discipline to an undisciplined organization.13

Organizational Learning

The sophistication of a firm’s organizational learning substantially facilitates its ability to manage an outsourcing arrangement in the systems development area effectively. A significant component of many firms’ applications development portfolios are projects related to business process reengineering or organization transformation.14 Process reengineering seeks to install very different procedures for handling transactions and doing the firm’s work. Organization transformation tries to redesign where decisions in the firm are made and the type of controls used. The success of both types of projects depends on getting internal staff people to radically change the way they work and often involves significant downsizing as well. While much of this restructuring relies on new IT capabilities, at the heart it is an exercise in applied human psychology where 70 percent of the work falls disappointingly short of target.

Responsibility for this type of development work (low structure by its very nature) is the hardest to outsource, although there is an extraordinary consulting industry that is assisting its facilitation. A firm with substantial experience in restructuring will have less difficulty in defining the dividing line between the outsourcer and the company in terms of responsibility for success. Firms that have not yet worked on these projects will find that outsourcing significantly complicates an already difficult task. The more experience the firm has had in implementing these projects, the easier the outsourcing will be.

A Firm’s Position in the Market

The farther a company is from the network era in its internal use of IT, the more useful outsourcing is as a way to close the gap. The DP era and early micro era firms often do not have the IT leadership, staff skills, or architecture to quickly move ahead. The outsourcer cannot just keep its old systems running but must drive forward with contemporary practice and technology. The advanced micro era firms are more likely to have internal staff skills and perspectives to leap to the next stage by themselves. The world of client/server architecture, the networked organization, and process redesign is so different from the large COBOL systems and stand-alone PCs of 1985 that it is often prohibitively challenging for a firm to easily bridge the ten-year gap by itself. For firms in this situation, it is not worth dwelling on how the firm got where it is but, rather, how to extricate itself.

Current IT Organization

The more IT development and operations are already segregated in the organization and in accounting, the easier it is to negotiate an enduring outsourcing contract. A stand-alone differentiated IT unit has already developed the integrating organizational and control mechanisms that are the foundation for an outsourcing contract. Separate functions and their ways of integrating with the rest of the organization already exist. Cost accounting processes have been hammered out. While both sets of protocols may require significant modification, there is a framework in place to deal with them.

When there are no protocols, developing an enduring contract is much more complex, because the firm must establish both the framework for resolving the issue and the specific technical approaches. This structure facilitated the General Dynamics implementation effort tremendously; the lack of this structure in another high-technology firm extended the resulting outsourcing process over a period of years, with a diminution of savings and a complex conversion.

Structuring the Alliance

Establishing the parameters of the outsourcing arrangement at the beginning is crucial. The right structure is not a guarantee of success, but the wrong structure will make the governance process almost impossible. Several factors are vital to a successful alliance.

Contract Flexibility

From the customer’s viewpoint, a ten-year contract simply cannot be written in an iron-clad, inflexible way.15 The arrangements we have examined have changed over time, often radically. Evolving technology, changing business economic conditions, and emergence of new competitive services make this inevitable. The necessary evolutionary features of the contract make the alliance’s cognitive and strategic fit absolutely crucial. If there is mutual interest in the relationship and if there are shared approaches to problem solving, the alliance is more likely to be successful. If these do not exist, a troublesome relationship may emerge.

No matter how much detail and thought goes into drafting the contract, the resulting clauses will provide imperfect protection if things go wrong. Indeed, the process of contract drafting (which often takes six to eight months) is likely to be more important than the contract. At this phase, one side gains insights into the other’s values and the ability to redirect emphasis as the world changes. Kodak has already altered its outsourcing arrangement as both business circumstances and technologies changed (see the sidebar). General Dynamics had eight contracts to provide for different divisions evolving in separate ways.

Standards and Control

One concern for customers is that they are handing control over an important part of the firm’s operations to a third party. This is particularly true if IT innovation is vital to the firm’s success or if the firm is very dependent on IT for smooth daily operations. A company must carefully address such concerns in the outsourcing agreement.

Control in some ways is just a state of mind. Most organizations accustom themselves to loss of control in various settings, as long as the arrangement is working out well and the supplier is fully accountable. However, it is worth noting that vital parts of a firm’s day-to-day operations have always been controlled by others. Electricity, telephone, and water are normally provided by third parties; interruption in their support can severely cripple any organization in a very short time. Providing sustained internal backup is often impractical or impossible. In the case of a hotel, for example, failure of either electricity or water for more than twenty-four hours is enough to guarantee its closing until the situation is rectified. The managers at a major chemical company who were particularly concerned about loss of control were brought up short by one of us who asked to see its power-generating facilities and water wells (of course, they had neither).

Disruption of operations support has immediate and dramatic implications for many firms. It is also a short-term area where firms are likely to feel comfortable about their protection if carefully structured. Conversely, putting innovation and responsibility for new services and products in the hands of a third party is correctly seen as a risky, high-stakes game. As we discuss later, in outsourcing, these issues are much more capable of being resolved for the firms in the factory and support quadrants (see Table 2), where innovation is much less important, than for firms in the turnaround and strategic quadrants (unless they are midsize and looking for access to critical innovation skills that they cannot command on their own). A company must carefully develop detailed performance standards for systems response time, availability of service, responsiveness to systems requests, and so on. Only with these standards in place can the company discuss the quality of support and new trends.

Outsourcing at Kodak »

Areas to Outsource

A company can oursource a wide selection of IT functions and activities. Data center operations, telecommunications, PC acquisition/maintenance, and systems development are all examples of pieces that can be outsourced individually. Continental Illinois outsourced everything, while Kodak kept systems development but outsourced, in separate contracts, data center operations, communication, and acquisition management of PCs. At its core, outsourcing is more an approach than a technique. As we noted earlier, significant portions of a firm’s IT software development activities have been routinely out-sourced for years. What is at stake here is a discontinuous major shift to move additional portions of a firm’s IT activities outside the firm. Between the current situation and total outsourcing lie a variety of different scenarios. When assessing partial outsourcing, managers frequently ask the following questions:

  • Can the proposed outsourced piece be separated easily from the rest of the firm, or will the complexities of disentanglement absorb most of the savings?
  • Does the piece require particular specialized competencies, which we either do not possess or lack the time and energy to build?
  • How central are the proposed outsourced pieces to our firm? Are they either more or less significant to the firm’s value chain than the other IT activities and, thus, deserve different treatment?

Total outsourcing is not necessary for attracting a supplier’s attention, but the portion to be outsourced must be meaningful enough so that the vendor will pay attention to it. Several organizations we studied had spun off bits and pieces of their activities to various organizations in a way that engendered enormous coordination costs among multiple organizations. These contracts were also very small in relation to the outsourcers’ other work, and we had significant concerns about their long-term viability (in one case, the firm has already insourced again).

Cost Savings

Some CIOs believe that the firm’s IT activities are so well managed or so unique that there is no way to achieve savings through outsourcing, or for the vendor to profit. This may be true. But two caveats are important. First, only if several outsiders study outsourcing with senior management sponsorship are you ensured an honest, realistic viewpoint. Having a study done objectively under the sponsorship of the local IT organization can be very difficult. The IT operation may see an outsourcing recommendation as so deeply disruptive that the study may be negatively biased from the start. Additionally, since consultants retained by the IT organization are often dependent on it for future billings, either consciously or unconsciously, they may skew the results. One firm’s internally initiated IT study, done by a consultant, purported to show that it was 40 percent more efficient than the average firm in its industry. Needless to say, IT’s control of the evaluation process led to general management’s skepticism; the study is being redone under different sponsorship.

Disinterested professionals can make a real contribution in evaluating cost savings. A major aerospace firm recently retained a consulting firm to help it design and manage the outsourcing review process. Doing an audit in this area is also difficult. Because situations change rapidly and new priorities emerge, it is usually impossible to determine what the results would have been if the alternative had been selected. Thus the IT organization may be tempted to anticipate internal efficiencies so that IT outsourcing does not appear to be viable.

Supplier Stability and Quality

How a supplier will perform over a decade is an unanswerable question, and one of the most critical a customer must ask. In ten years, technologies will change beyond recognition, and the supplier that does not have a culture that encourages relentless modernization and staff retraining will rapidly become a liability as a strategic partner. In addition, the stability of the outsourcer’s financial structure is critical. Cash crunches, subchapter 11, or worse are genuine nightmares for customers. This issue is complicated by the reality that once a firm outsources, it is very hard to insource again, once the firm’s technical and managerial competence have evaporated. While it is difficult to move quickly from one outsourcer to another (the only practical alternative), if a firm considers the possibility in advance, it can mitigate the risks. Some of the best bids come from newcomers trying to crack the market.

Problems are intensified if the way a firm uses technology becomes incompatible with the outsourcer’s skill base. For example, a firm in the factory quadrant that selects an operationally strong outsourcer may be in trouble if it suddenly moves toward the strategic quadrant and its partner lacks the necessary project management and innovation skills to operate there.

Finally, there is a potential, built-in conflict of interest between the firm and the outsourcer that must be carefully managed so it is does not become disabling. The outsourcer makes its money by lengthening leases, driving down operational costs, and charging premium prices for new value-added services. Conversely, the customer has no empathy for the joys of harvesting old technology benefits (one of the reasons it got out of the business in the first place) and also wants rapid access to cheap, high-quality project development skills on demand. Managing this tension is complex, imperfect, and very delicate and must be covered in the contract. Both firms must make a profit. The more the customer moves to the strategic quadrant, the harder it is to ensure a good fit with an outsourcer.

Management Fit

Putting together a ten-year, flexible, evolving relationship requires more than just technical skill and making the numbers work. A shared approach to problem solving, similar values, and good personal chemistry among key staff people are critical determinants of long-term success. Various outsourcers have very different management cultures and styles. It is worth giving up something on the initial price to ensure that you find a partner with whom you can work productively over a long term. The information gained in the tortuous six- to eight-month process of putting an alliance together is crucial for identifying the likelihood of a successful partnership. Of course, this chemistry is a necessary, but insufficient, condition for ultimate success. Realistically, it is corporate cultural fit that is most important, since, after several years, the key people in the initial relationship will have moved to other assignments.

Conversion Problems

The period of time for an outsourcing study and conversion is one of great stress for a company’s staff. Uncertainties about career trajectories and job security offer the possibility of things going awry. All the expertise a firm gains when acquiring another is vital during conversion. The sooner plans and processes for dealing with staff career issues, outplacement processes, and separation pay are dealt with, the more effective the results will be. Almost invariably, paralyzing fears of the unknown are worse than any reality.

Managing the Alliance

The ongoing management of an alliance is the single most important aspect of outsourcing’s success. We have identified four critical areas that require close attention.

The CIO Function

The customer must retain a strong, active CIO function. The heart of the CIO’s job is planning — ensuring IT resources are at the right level and appropriately distributed. This role has always been distinctly separate from the active line management of networks, data centers, and systems development, although it has not always been recognized as such. These line activities, as noted earlier, have been successfully outsourced in a variety of companies. In a fully outsourced firm, however, sustained internal CIO responsibility for certain critical areas must be maintained. (For an example, see the side-bar on the evolution at Kodak.)

  • Partnership/Contract Management. As we described earlier, outsourcing does not take place in a static environment. The nature of the technologies, external competitive situations, and so on are all in a state of evolution. An informed CIO who actively plans and deals with the broad issues is critical to ensuring that this input is part of the alliance so it can continuously adapt to change. The evolving Kodak contract gives ample evidence of this.
  • Architecture Planning. A CIO’s staff must visualize and coordinate the long-term approach to interconnectivity. Networks, standard hardware/software conventions, and database accessibility all need customer planning. The firm can delegate execution of these areas, but not its viewpoint of what it needs to support the firm in the long term. The wide range of management practice in firms that do not outsource complicates this. One insurance company we studied supported a network of 15,000 PCs, 18 e-mail systems, and literally hundreds of support staff to maintain its networks. Its CIO stated succinctly, “Even if technology costs ‘zero,’ we cannot continue in this way.”

Emerging Technologies.

A company must develop a clear understanding of emerging technologies and their potential applications. To understand new technology, managers must attend vendor briefings and peer group seminars and visit firms currently using the new technology. Assessing the hardware/software network alternatives and their capabilities requires an understanding of what is in the market and where it is going. This cannot be delegated to a third party or assessed by sitting in one’s office.

Similarly, identifying discontinuous applications and the opportunities and problems posed by them is critical. At one large pharmaceutical organization, the CIO’s staff was vindicated when it became clear that they had first spotted business process redesign as an emerging area, funded appropriate pilot projects (which were skillfully transferred to line management), and finally repositioned the firm’s entire IT effort. (Users and an outside systems house executed the project, with the CIO playing the crucial initiator role.) Clearly, an outsourcer has an incentive to suggest new ideas that lead to additional work, but delegating responsibility for IT-enabled innovation in strategic and turnaround firms is risky because it is such an important part of the value chain.

Continuous Learning.

A firm should create an internal IT learning environment to bring users up to speed so they are comfortable in a climate of continuous IT change. An aerospace firm felt this was so important that, when outsourcing, it kept this piece in-house.

An Outsourcer's Viewpoint »

Performance Measurements

Realistic measurement of success is generally very hard, so a company must make an effort to develop performance standards, measure results, and then interpret them continuously. Individual firms bring entirely different motivations and expectations to the table. In addition, many of the most important measures of success are intangible and play out over a long period of time. Hard, immediate cost savings, for example, may be measurable (at least in the short run), but simplification of the general management agenda is impossible to assess.

The most celebrated cases of outsourcing have evolved in interesting ways. Of Kodak’s three selected vendors, while the major one remains intact, another of the vendors has gone through several organizational transformations, triggered by financial distress. General Dynamics, in the first eighteen months, has spun off three of its divisions, along with their contracts. It is too early to determine the outcomes. EDS and GM took years to work out an acceptable agreement; ultimately several very senior EDS managers resigned.

A major power company recently postponed an out-sourcing study for a year. Its general managers believed that their internal IT staff and processes were so bloated that, while outsourcing IT would clearly produce major savings, they would still be leaving money on the table. Consequently, in 1993, they reduced their IT staff from 450 to 250 and reduced the total IT expenditure level by 30 percent. With the “easy” things now done, they are now entertaining several outsourcing proposals to examine more closely what additional savings and changes in their method of operation would be appropriate.

Mix and Coordination of Tasks

As we noted earlier, the larger the percentage of a firm’s systems development portfolio devoted to maintaining legacy systems, the lower the risk of outsourcing the portfolio. The question becomes, can we get these tasks done significantly faster and less expensively? The larger the percentage of large, low-structured projects in the systems development portfolio, the more difficult it becomes to execute a prudent outsourcing arrangement and the more intense the coordination work to be done. Large systems development projects that use advanced technology play directly to the outsourcers’ strengths. Conversely, issues that relate to structure (and thus close, sustained give and take by users) require so much extra coordination that many outsourcing benefits tend to evaporate.

On the one hand, the costing systems, implicit in the outsourcing contracts that use hard dollars, force users to be more precise in their systems specifications early on (albeit a bit resentfully) and thus cut costs. On the other hand, evolving a sensible final design requires trial and error and discussion. Both the contract and the various geographic locations of the outsourcer’s development staff can inhibit discussion and lead to additional costs if not carefully managed. Managing the dialogue across two organizations with very different financial structures and motivations is both challenging and, at the core, critical to the alliance’s success. Concerns in this area led Kodak to not outsource development. Other firms, such as British Aerospace, did, after careful analysis.

Customer-Outsourcer Interface

The importance of the sensitive interface between the company and the outsourcer cannot be overestimated. First, outsourcing can imply delegation of final responsibility to the outsourcer. The reality is that oversight simply is not delegatable, and, as we noted earlier, a CIO and supporting staff need to manage the agreement and relationships. Additionally, the interfaces between customer and outsourcer are very complex and should occur at multiple levels, as noted in both side-bars. At the most senior levels, there must be links to deal with major issues of policy and relationship restructuring, while, at lower levels, there must be mechanisms for identifying and handling more operational and tactical issues. For firms in the strategic quadrant, these policy discussions take place at the CEO level and occasionally involve the board of directors.

Both the customer and outsourcer need regular, full-time relationship managers and coordinating groups lower in the organization to deal with narrow operational issues and potential difficulties. These integrators are crucial for managing different economic motivations and friction. The smaller the firm in relationship to the outsourcer’s total business, the more important it is that these arrangements be specified in advance before they get lost in other priorities.

During the past five years, an entirely different way of gaining IT support for outsourcing has emerged. While outsourcing is not for everyone, some very large and sophisticated organizations have successfully made the transition. What determines success or failure is managing the relationship less as a contract and more as a strategic alliance.

References

1. R.L. Huber, “How Continental Bank Outsourced Its ‘Crown Jewels,’ ” Harvard Business Review, January–February 1993, pp. 121–129.

For a critical discussion of the outsourcing phenomenon, see:

R.A. Bettis, S.P. Bradley, and G. Hamel, “Outsourcing and Industrial Decline,” Academy of Management Executive, February 1992, pp. 7–22.

A library search on “outsourcing information technology” articles for the past year and a half identified over 700 articles, almost all describing IT outsourcing as reported in the various industry trade journals.

2. For a discussion of the impact of outsourcing, see:

L. Loh and N. Venkatraman, “Diffusion of Information Technology Outsourcing: Influence Sources and the Kodak Effect,” Information Systems Research 4 (1992): 334–358.

3. See “Kodak’s Outsourcing Deal Brings Risks — And Not Just for Kodak,” The Business Week Newsletter for Information Executives, August 1989; and

D. Norton, H. Pfendt, G. Biddle, and R. Connor, “A Panel Discussion on Outsourcing,” Stage by Stage, 25 January 1990, pp. 13–16.

4. “General Dynamics and Computer Sciences Corporation: Outsourcing the IS Function” (A), (B), (C), (D) (Boston: Harvard Business School, Publishing Division, 1993).

5. The theory of what transactions are best conducted within the firm and outside the firm has been comprehensively covered. See:

O.E. Williamson, Markets and Hierarchies (New York: Free Press, 1975).

For further contributions on the inefficiencies of hierarchies, see:

H. Leibenstein, Inside the Firm: Inefficiencies of Hierarchy (Cambridge, Massachusetts: Harvard University Press, 1987).

For the notion of a new form of relationship evolving among firms, see:

R. Johnson and P.R. Lawrence, “Beyond Vertical Integration: The Rise of the Value-Adding Partnership,” Harvard Business Review, July–August, 1988, pp. 94–104.

For further discussion of the nature of the evolving firm, see:

J.L. Badaracco, Jr., The Knowledge Link: How Firms Compete Through Strategic Alliances (Boston: Harvard Business School Press, 1991).

For a discussion of the process of managing strategic alliances, see:

M.Y. Yoshino, Managing Strategic Alliance (Boston: Harvard Business School Press, 1994).

Finally, for a discussion of competitive strategy and the role of strategic alliances, see:

M.E. Porter, Competitive Advantage (New York, Free Press, 1985).

6. Porter (1985).

7. See R.L. Nolan, “Managing the DP Crisis,” Harvard Business Review, March–April 1979, pp. 115–126; and

R.L. Nolan, D.C. Croson, and K. Seger, “Note on Stages Theory Today” (Boston: Harvard Business School, Publishing Division, 1993).

8. See S.P. Bradley, J.A. Hausman, and R.L. Nolan, eds., Globalization, Technology & Competition (Boston: Harvard Business School Press, 1993).

9. See L. Loh, “Determinants of Information Technology Outsourcing: A Cross-Selectional Analysis,” Journal of Management Information Systems 9 (1992): 7–24.

10. See K.P. Arnett and M.C. Jones, “Firms That Choose Outsourcing: A Profile,” Information Management 26 (1994): 179–188.

11. For a discussion of the strategic grid, see:

F.W. McFarlan, J.I. Cash, and J.L. McKenney, Corporate Information Systems Management (Homewood, Illinois: Richard D. Irwin, 1992). McFarlan developed the framework in 1982. See:

F.W. McFarlan, J.L. McKenney, and P. Pyburn, “The Information Archipelago — Plotting a Course,” Harvard Business Review, January–February 1983, pp. 145–160.

We have generalized the two dimensions of strategic importance. The operations dimension has been generalized from dependence of the company on the computer for processing transactions to overall dependence of the company on sustained real-time processing of overall business processes. We term this dimension “information intensity.” See:

M.E. Porter and V.E. Millar, “How Information Gives You Competitive Advantage, Harvard Business Review, July–August 1985, pp. 149–160.

The second dimension of the strategic grid that we generalized was the importance of new applications development. Here we generalize applications development to the importance of innovative information resource management. Indeed, in-house applications development is one aspect of this dimension, but modern development initiatives commonly involve a number of outside partners as well.

12. See McFarlan et al. (1992).

13. For additional issues in outsourcing development, see:

“The Managing of Partnering Development in IS Outsourcing,” Proceedings of the Twenty-Sixth Annual Hawaii International Conference on Systems Sciences 4 (1992): 518–527.

14. See M. Hammer and J. Champy, Reengineering the Corporation (New York: Harper Collins, 1993); and

T. Davenport, Process Innovation (Boston: Harvard Business School Press, 1993).

15. For a discussion of issues involved in an outsourcing contract, see: W.B. Richmond and A. Seidmann, “Software Development Outsourcing Contract: Structure and Business Value,” Journal of Management Information Systems 10 (1993): 57–72.