The Information Systems Outsourcing Bandwagon
During the 1980s, executives learned of the strategic role that information systems (IS) could provide to organizations. Through IS, companies could squelch competition, secure suppliers, obtain customer loyalty, and reduce the threat of new entrants. Executives read about the IS victory stories of American Airlines, American Hospital Supply, and Merrill Lynch as evidence of the success to be found in exploiting information systems. Many executives applied strategic IS models to their own organizations.
As we enter the 1990s, one would assume that the role of IS would escalate as the United States grows from a domestic, industrial society to a global information society. However, many practitioners, academics, and consultants now advise executives to outsource their IS services along with their cafeteria, mail delivery, and custodial services. By farming out many or all of their information services, executives are promised savings of 10 percent to 50 percent off their IS expenditures.1 Many companies are likely to heed this advice, as evidenced by a Yankee Group report that estimates that every Fortune “500” company will evaluate IS outsourcing and 20 percent will sign outsourcing deals by 1994.2 The outsourcing trend is not limited to American companies; Frost and Sullivan estimate that the European outsourcing market will grow from $1.5 billion in 1990 to $10 billion in 1996.3
So what has happened to the IS function in the past few years? Has IS suddenly become a commodity service that is best managed by a large supplier? Several practitioners believe so. Henry Pfendt, director of information technology at Eastman Kodak, claims outsourcing comes down to one question: “Do you want to manage commodities?”4 Elliot McNeil, Southland’s management of information systems (MIS) manager: “It’s like the electric company; you use less, you pay less.”5 Elizabeth Gardner, in a commentary on IS outsourcing in the health care industry: “Outsourcing takes over all information systems functions, much the way an outside company would manage food service or laundry.”6 Bernie Ward, editor of Sky magazine, described IS out-sourcing: “Like shoppers at a fruit stand, companies are picking an apple here or an orange there until they have selected a menu of outsourcing services.”7
Are electric utilities, fruit stands, and laundry services reasonable metaphors for IS? A strong argument for the IS as utility metaphor is that information, rather than information systems, provides a company with a competitive advantage. Max Hopper, senior vice president for American Airlines, notes that it is “dangerous still to believe that on its own, an information system can provide an enduring business advantage. The old models no longer apply.”8 Instead, Hopper and other executives contend that competitive firms will shift their attention from systems to information.9
Once IS is viewed as a utility, executives will shift their efforts from harnessing IS as a competitive weapon to providing IS services at the lowest possible cost. Under the theory of economies of scale, these executives would argue, mass production and specialized labor allow large vendors to provide utilities least expensively. By letting the technology vendors provide the IS utility, executives can focus energies on nurturing their company’s core competencies.
The notion that IS is a utility and therefore most efficiently provided through outsourcing has been reinforced by many trade reports. Computerworld and InformationWeek, for example, track outsourcing arrangements in which customers report savings of 10 percent to 50 percent. Many companies, in their exuberance to duplicate savings advertised in these reports, jump on the IS outsourcing bandwagon. Press reports, however, portray an overly optimistic view of outsourcing because they are made during the honeymoon period when a customer and vendor first sign a contract. At this point, the reported savings are only projected, not actual, savings. In addition, outsourcing failures are underrepresented in the press. After all, no organization wishes to offer a failure for public scrutiny.
Thus the outsourcing phenomenon can be explained by at least two trends. First, many executives believe that IS is quickly evolving into a utility. As such, utility services are more efficiently acquired through specialized vendors that achieve economies of scale. Second, a significant bandwagon effect has been noted.10 Outsourcing successes (primarily at Kodak, but also at American Bankshares, Southeast, and Continental) have prompted many executives to outsource without due consideration of the potential consequences.
In this article, we question the widespread endorsement of outsourcing by exposing several myths generated by press reports. Under specific circumstances, such as when companies cannot control IS costs on their own or need to sell IS assets to generate cash, companies may still wish to outsource. To help such organizations realize their outsourcing expectations, we also present some contract negotiation strategies.
What Is Information Systems Outsourcing?
Outsourcing, in its most basic form, can be conceived of as the purchase of a good or service that was previously provided internally. Information systems out-sourcing options have existed since the dawn of data processing. As early as 1963, Electronic Data Systems (EDS) handled data processing services for Frito-Lay and Blue Cross & Blue Shield. Other outsourcing options, such as the use of contract programmers, timesharing, and purchase of packaged software, have been exercised for two decades. What renews interest in IS outsourcing and demands our attention today is the dramatic change in scope. Early forms of IS outsourcing typically dealt with single-system contracts comprising a small portion of the IS budget — payroll, insurance processing, credit cards, or mailing lists. Outsourcing has recently grown to span multiple systems; it now represents a significant transfer of assets, leases, and staff to a vendor that assumes profit and loss responsibility. Enron’s $750-million, ten-year contract with EDS is a typical example of this increased scope. Vendors receive the “keys to the kingdom,” actually operating, managing, and controlling IS functions.
A typical outsourcing arrangement of this type works like this: The vendor charges a fixed fee for a prespecified number of services, known as the “baseline.” The customer is guaranteed that its IS costs for this baseline will be fixed over the contract duration, typically five to ten years. During the contract period, services not included in the baseline may be purchased from the vendor for an excess fee. Deals are often sweetened with financial incentives, such as stock purchases, loans at low interest rates, and postponed payments. At the outset, these deals are extremely attractive, especially to an organization that suffers financially. But what happens in the long term?
The Study
Our arguments are based on an in-depth, multiple-case analysis of fourteen Fortune “500” service and manufacturing companies that faced outsourcing decisions. We interviewed individuals who were directly involved in the outsourcing decision on behalf of the organization or outsourcing vendor for one to three hours. Participants included chief financial officers, controllers, treasurers, IS directors, account managers, and consultants hired to assist customers. In addition, we interviewed members of the IS staff responsible for gathering technical and financial details.
During the interviews, we first asked participants to tell us their outsourcing story. This way, participants could discuss what they thought was the essence of the outsourcing decision. Participants took between twenty and forty-five minutes to relate their stories. Then we asked a series of semistructured questions about the financial, historical, and political details that may have been absent from their previous recollections. In addition, the participants filled out a structured questionnaire on demographic and other vital information, and we gathered relevant documentation, including requests for proposals, outsourcing bids, outsourcing contracts, efficiency assessments, annual reports, and organizational charts. We used the documentation to corroborate statements made during the interviews and to analyze contract specifics.
Of these fourteen companies, six limited the out-sourcing domain to data center operations or telecommunications or both (see Table 1). Seven considered outsourcing the entire IS department — data processing, applications support, applications development, end-user computing, training, and so forth. Five companies decided against outsourcing (cases one through five) and nine decided to outsource (cases six through fourteen). Two of the latter companies suffered such severe service degradation that they prematurely terminated their contracts, erected new data centers, and populated new IS departments (cases thirteen and fourteen).
The current study tempers the media’s optimistic claims by examining (1) some outsourcing arrangements three to seven years into the contract, (2) several out-sourcing failures as well as successes, and (3) organizations that decided they could duplicate vendor savings without outsourcing. Although the current research is not a random sample, it does provide some alternative views absent from the trade press. In particular, the look at organizations that decided not to outsource provides a control group absent from other studies. These cases provide support for the argument that “there is nothing that vendors can do that you can’t do for yourself.”11
Information Systems Outsourcing Myths
Although we expected considerable variation in the nature, style, and outcome of the outsourcing decisions, a surprising number of common themes were apparent across the companies. In particular, three outsourcing myths surfaced.
Myth: Outsourcing vendors are strategic partners. Fact: Outsourcing vendors are not partners because the profit motive is not shared.
The idea that outsourcing vendors are “strategic partners” may be attributed to Eastman Kodak. In 1987, Eastman Kodak outsourced almost its entire IS operations to IBM, Businessland, and Digital Equipment Corporation. Vaughn Hovey, director of data center services for Kodak, told an audience of practitioners, “We think of our strategic alliances as ‘partnerships’ because of their cooperative and long-term qualities.”12 Kodak sealed these partnerships with little more than a gentleman’s agreement. According to Hovey, Kodak rarely refers to their “six- or seven-page contract.” The importance of Kodak’s success cannot be overstated; statistical analysis suggests that the outsourcing trend can be explained as “imitative behavior” of Kodak’s decision.13 Kodak made it acceptable for large organizations to out-source and popularized the notion that outsourcing vendors are strategic partners. Therefore, loose agreements could serve as acceptable contracts. What works for Kodak, however, may not work for others.14 Because the Kodak arrangement received so much press coverage, vendors are unlikely to behave opportunistically.15
The term “strategic partner” is unsuitable to characterize the relationship between an outsourcing vendor and its customer because the profit motive is not shared.16 Account managers at outsourcing providers are rewarded for maximizing profits, primarily by charging customers additional fees for services that extend beyond the contract (in the outsourcing parlance, these are called “excess fees”). When a customer’s costs increase, so do the vendor’s profits. How then can an outsourcing vendor be conceived of as a partner? The term “customer” is more appropriate.
The danger in viewing the outsourcing vendor as a partner is that the customer may sign a loose agreement. After the agreement goes into effect, the vendor may not provide the level of service the customer expects. Instead, the vendor may refer to the written contract as the only source of obligation. Customers may subsequently be charged excess fees for services that they assumed were in the contract. One petroleum company was charged almost $500,000 in excess fees the first month into the contract. These were for services its managers assumed were covered in the agreement. The vendor rightfully retorts that services not documented in the contract are above baseline and subject to excess fees. This does not imply that outsourcing vendors are opportunistic; more likely, the outsourcing customers and vendor representatives miscommunicated the full scope of the agreement. Customers in this situation have little recourse as the vendor possesses a significant amount of power over the relationship. What other information alternatives are available to the dissatisfied customer? It sold its assets, transferred leases, and transferred employees in order to enter into the arrangement.
Myth: Outsourcing vendors are inherently more efficient than internal IS departments. Fact: Internal IS departments can be cost competitive.
Proponents of outsourcing, such as Gary Biddle of American Standard, claim that outsourcing vendors are more efficient due to economies of scale.17 A review of the theory of economic efficiency, however, exposes several myths about its applicability to IS outsourcing.
The theoretical basis of economies of scale is that large companies have lower average costs than small companies due to mass production and labor specialization efficiencies. Mass production is presumed to reduce average costs by allocating fixed costs over more units of output and by receiving volume discounts on inputs. Labor specialization is presumed to reduce costs by allowing workers to focus on tasks at which they are most adept.
Applying the economic theory of efficiency to information systems outsourcing is a messy proposition. For one, the inherent problem of measuring efficiency surfaces. How do we know a vendor is inherently more efficient than the internal IS department? The only area of efficiency that has been adequately measured is data processing, in particular, cost per millions of instructions per second (MIPS).18 Real Decisions, for example, has documented that costs per central processing unit (CPU) minute are lowest for data centers in the 135 to 200 MIP range. Its analysis is based on data gathered and normalized from hundreds of its customers. Most of the efficiency arguments, however, are more esoteric. For example, vendors are assumed to have lower average hardware and software costs due to volume discounts and the sharing of software lease costs over multiple customers. On the labor specialization side, outsourcing vendors are assumed to be information systems specialists. They can allegedly design, develop, and maintain systems more efficiently because of their expertise. These arguments are, however, largely fallacious.
Even Small Companies Can Achieve Average Hardware Costs Comparable to Those of an Outsourcing Vendor.
Smaller companies can often negotiate dirt-cheap hardware leases by using older technology. The IS manager at a petroleum company, for example, leases an IBM 3081K for only $4,000 a month. He feels that smaller IS shops can contain hardware costs by renegotiating leases on dated equipment: “As long as we stay on the trailing edge of technology — and I’ve been pushing this concept to senior management — we have an opportunity to capitalize on cheaper computing costs.”
A Vendor’s Volume Discounts Are Often Negligible.
Economic theory purports that large companies have lower average costs than small companies partly because they receive volume discounts on inputs. The volume discount theory presumes that large companies buy in bulk and therefore receive quantity discounts. This argument extends to the outsourcing arena by assuming that outsourcing vendors buy hardware for less.19 Many internally managed companies, however, receive discounts similar to outsourcing vendors. One prominent outsourcing consultant offers the following example:
Can an outsourcing vendor buy a machine for less than Enron? Sure it can. I’ll give you numbers that aren’t exactly accurate, but a large IBM 3090 probably is in the $15-million-dollar list range. An out-sourcing vendor could probably get it for $11 million. Enron or First City could probably get it for about $11.5 or $12 million. Now, over a five-year period, there is not a whole lot of money being saved.
This finding is corroborated by a recent Datamation survey that reports that the major hardware vendors typically sell equipment at 9 percent to 17 percent below list price.20
Recent Changes in Software Licensing AgreementsHave Greatly Reduced a Vendor’s Advantage.
The assumption that average costs decrease when fixed costs are spread over more units of output has been applied to software costs.21 The argument is that outsourcing vendors spread software licensing fees over multiple customers. This argument is no longer valid for two reasons. First, software companies have changed the structure of their software licensing fees in response to outsourcing. Second, software companies charge customers excessive fees to transfer licenses to an outsourcing vendor.
Before outsourcing came into vogue, software houses issued site licenses to customers. This meant that a company paid a fixed fee for one copy of a software package used at a single data center site. Software houses, however, have changed the structure of site licensing fees in reaction to outsourcing. Now they charge a fee based on the size of the hardware. With these new licenses, called group licenses, customers with bigger machines get charged more money.
The change to group licenses has seriously curtailed the outsourcing vendor’s profitability. An account manager for a mining customer explains the impact on an outsourcing vendor’s costs:
Computer Associates is thinking, “Every time these deals are signed, the outsourcer [vendor] can channel everything into one box, use one copy of the software. Therefore, we are going to lose money.” And from a business perspective, I suppose that makes sense. So what the software vendors have done is gone to group pricing. And of course, the outsourcer operates a larger box than [the mining company], so it pays more fees. . . . This has dire consequences for outsourcing deals.
In addition to the structural change, software houses may charge transfer fees to their customers that out-source. For example, a software house sued a large commercial bank for $500,000 for transferring the software license to an outsourcing vendor. From the software house’s perspective, outsourcing reduces its revenue; the transfer fee is its way of seeking compensation.
Access to Technical Talent Is Limited Because Customers Are Often Supported by Their Previous Staff.
Many companies outsource to access the vendor’s pool of technical talent, only to find (1) they are supported by their previous IS staff and (2) additional vendor expertise is expensive.
One industrial equipment manufacturer outsourced because managers felt their IS staff did not have the technical skills to implement a new computer architecture. The conversion (not to mention the whole out-sourcing arrangement) failed because the same people (now vendor employees) performed the installation. The company’s IS director notes, “[The vendor] took over all the people as they usually do, so what happens in that environment is [that] your unqualified IS people become unqualified [vendor] people.”
One mining company outsourced largely to acquire vendor expertise. In this case, only half of the company’s IS department transferred. The company informed employees of the outsourcing decision early in the process, and many found positions elsewhere. The vendor replaced the missing staff, but the new talent was expensive. A purchasing manager from the company notes, “None of it is cheap. There is a perception that once you have [a vendor] locked in, you have a conduit to all this expertise. But you pay.”
Access to Business Talent Is Questionable.
Many participants who outsourced felt a real loss of business expertise. The problem is that vendors often siphon talented employees to woo other accounts in the industry. In addition, customers claim that outsourcing vendors prefer specific directions and rarely initiate business strategies.
An IS manager at an industrial chemicals manufacturer believes that outsourcing reduced the business expertise of his IS staff. He claims talented businesspeople are often transferred to new accounts: “You pay for them to learn your business, then they move those people to court other companies in your industry. They transfer skills to get new business; now the learning curve is yours to pay for again.”
A vice president of liquid fuels at a diversified services company concurs. He claims that the vendor not only transferred the best employees to other accounts, it retrained the remaining staff to be more technical:
Of the managers we had, the three best ones were transferred. Has the caliber improved or changed radically? The nature of the people has changed. They are more technical, less user oriented; they are more technically oriented and less functionally oriented, less industry oriented. So if you were to grade them overall, I lost.
Low Bids Do Not Necessarily Indicate Greater Efficiency.
If a vendor submits a bid that undercuts current IS costs, the potential customer should determine how the vendor has managed to reduce costs. The vendor may be inherently more efficient, but this should not be accepted without scrutiny. Remember that a natural cost advantage must be significant to cover a vendor’s profit margin. The IS manager at an equipment manufacturer warns: “Let’s just assume that they have a certain amount of cost [that is lower]. But those guys look for a gross margin of 50 percent to 60 percent. So how can they do it cheaper?”
He stated that the average cost of a CPU minute is $6. Vendors, however, often charge their customers $10 or more per CPU minute. Also, customers should investigate what costs are included in the calculation. If cost per MIP is low, vendors may charge excess for other resources such as tape mounts or disk space.
If the vendor does not have inherent efficiencies of scale, the potential customer may decide to reduce costs on its own.
Myth: Savings of 10 percent to 50 percent can be achieved only through outsourcing. Fact: Internal IS departments can achieve similar results without vendor assistance.
If outsourcing vendors do not possess inherent cost efficiencies, how then do they propose to reduce IS costs? Three answers are possible. First, the vendor may offer a purely financial package that is independent of IS management issues. Second, vendors may exploit price-performance improvements. Third, the vendor may initiate cost-saving strategies that the customer could implement on its own.
Some Outsourcing Vendors Underbid Current IS Costs by Offering an Attractive Financial Package.
The vendor may offer a financial package whose net present value is extremely attractive to the prospective customer. Cash infusions for information assets, postponement of payments until the end of the contract, and even purchases of the customer’s stock may render outsourcing desirable. These outsourcing arrangements are not based on sound IS management decisions but are solely a monetary arrangement.
The First City and EDS outsourcing arrangement provides an example of the cash an outsourcer can bring to a faltering organization. EDS provided a much needed cash infusion by purchasing First City’s information systems assets. EDS also hired First City’s IS staff. In addition, EDS purchased $20 million in First City’s preferred stock. A Ross Perot biographer commented, “Both companies insist the transactions are unrelated, but the facts are that EDS landed a $550 million account, and the bank completed a badly needed recapitalization.”22 An EDS manager on the First City account stated, “Currently, the whole banking industry is scaling back, and systems falls low on the list of the bank’s priorities.”
In essence, outsourcing is being used to salvage a losing enterprise. Is it sound business practice to liquidate the IS department to rescue a firm? Many shareholders believe so — stock prices systematically rise just after an outsourcing announcement.23
Vendors Exploit Price-Performance Improvements.
Outsourcing vendors can often achieve savings simply by taking advantage of improvements in price-performance ratios. The improvement in price-performance is significant; a unit of today’s computing resources will cost 20 percent to 30 percent less next year.24 For example, one mainframe MIP cost $1 million in 1965 and now costs less than $30,000. Outsourcers know that cost per resource unit falls significantly over time. They can take over a data center, offer 10 percent savings over a ten-year period, and enjoy the profits generated by riding the price-performance curve. It’s primarily a waiting game.
Proposed Vendor Savings May Be Achieved Internally.
Many participants felt that cost savings achieved by vendors could be produced internally. Companies may be able to consolidate data centers, optimize current resource use, and implement more controls.
One way vendors may offer lower costs is by running the customer’s information systems through one data center. Prior to outsourcing, companies may run multiple data centers, each having excess capacity. Because the overhead costs can be excessive, vendors consolidate data centers to reduce costs. The question remains, however, whether companies could consolidate their data centers themselves. The corporate culture may resist consolidation without considerable upper management support. One large petroleum company tried and failed to consolidate because the IS department could not convince the powerful operating divisions to consolidate the six data centers into one. Not until senior management threatened divisions with outsourcing were they willing to consolidate. This company subsequently reduced costs by $22 million — without vendor assistance.
Companies may also be able to reduce costs by optimizing current resource use. One outsourcing consultant said that IS managers should occasionally “spring clean” operations:
Across the board, if I were to look at hardware, there is probably some old hardware that is around that I am paying maintenance on that we inherited from someone else, some other acquisition that we did, that we can turn around and say, “Get that out of here.” Software that nobody is using anymore. DASD [direct access storage device] space that is sitting out with files on it that haven’t been used in a year and a half.
Typically, however, a formal efficiency audit is needed to supplement spring cleaning. A number of reputable consultants are available to help customers “insource” —that is, use the current IS department to achieve savings similar to those of a vendor. These consultants will recommend improvements to resource usage, controls, and hardware and software purchasing practices.
IS departments can decrease costs by implementing cost controls such as chargeback systems, monthly software releases, or user request prioritization. Surprisingly, many participants did not employ these rudimentary controls. Without these controls, users simply view IS resources as “free.” With no demand restrictions, users request many services that simply are not cost justified.
In a recent article, Richard Huber of Continental Bank argued that the outsourcing culture in his organization gained momentum partly because outsourcing ties real dollar costs to support services. For example, before Continental Bank outsourced legal services, users routinely sent standard documents to the legal department. He noted:
Perhaps half of Continental’s problems with in-house services stemmed from overuse. For instance, the most routine documents were always sent to the legal department for review. “Better safe than sorry,” people would say, while thinking, “and besides, it’s just an internal cost, not real dollars.”25
With the outsourcing of legal, peripheral, and information services, Continental Bank was able to reduce costs. The question remains, however, whether Continental Bank could have implemented chargeback systems and other control methods without vendor assistance. One outsourcing consultant who participated in our study stated that customers could achieve the same savings as vendors by implementing chargeback systems:
A chargeback system is typically the best run-time improvement there is. With a chargeback system you get a bill that shows you “here’s everything that you ran for that month.” And if you were wasting resources, and the bill jumps as a result of that, you’d be amazed how much people reduce their costs the minute a chargeback system is implemented.
At another large petroleum company users would run production systems two or three times due to sloppy procedures. With a chargeback system implemented by the vendor, they postpone execution until input data is validated.
Another cost-saving measure is to implement monthly releases. At a large international bank, users told analysts to change production code daily. Once the bank out-sourced, the vendor tightened controls so that changes would be migrated to production only every two months. The bank’s outsourcing consultant explains:
[The bank] is an example where people made changes to production at least once a week. And they make changes to production every time the user calls and says, “I need this fixed, I want the screen changed.” There is a tremendous amount of money to be burned up in doing that.
Another cost-saving measure is to have users prioritize work requests. In many companies, the “squeaky wheel get greased.” By requiring user managers to prioritize requests for their departments, IS will only address the most pertinent issues. In addition, users are less likely to request cosmetic or nonessential changes if they know their managers will scrutinize work requests.
The contract is the only mechanism that establishes a balance of power in the outsourcing relationship.
Returning to Continental Bank, after it outsourced, Huber organized a technical oversight group (TOG) composed of user representatives from the business units:
The TOG . . . balances the technological requirements of individual units with the bank as a whole and decides which projects we will contract out and which must wait. Individual fiefdoms used to stake out priorities. Now they must submit IT proposals to the TOG for evaluation, which ranks them on a bankwide scale of priorities.”26
In summary, the cost-saving strategies mentioned here may all be achievable by an internal IS department. If the vendor does not enjoy an inherent efficiency advantage, “insourcing” may be more prudent because an internal IS department does not have to turn a profit.
Some companies, however, may still wish to out-source for a number of reasons: the political climate of the organization’s culture may prevent the IS department from initiating cost savings, the company may prefer to focus energies on more strategic issues than information systems, or the company may need the cash generated by the sale of information assets. If companies decide to outsource for one or more of these reasons, they need to protect their interests. The next section addresses fourteen negotiation strategies designed to balance the power between the outsourcing customer and the outsourcing vendor.
Negotiation Strategies
Once potential outsourcing customers realize that vendors are running businesses and are therefore motivated to maximize profits, they can ensure that outsourcing expectations are realized by signing a tight contract. The contract is the only mechanism that establishes a balance of power in the outsourcing relationship. Every person we interviewed who decided to outsource stated that the contract is the number one key to a successful outsourcing relationship. In this study, the companies most dissatisfied with outsourcing all signed contracts that dramatically favored the vendor. These contracts merely stipulated that the vendor would provide the same level of service that the company received prior to outsourcing. However, the service levels in the base year were poorly documented, so customers were subjected to costly excess fees for services they presumed were included. In contrast, participants most pleased with their outsourcing arrangements found that rigorous contracts reduced the threat of opportunism. When service levels, cost structures, and penalties for nonperformance are specified in the contract, the vendor becomes legally obligated to accommodate.
The following lessons provide advice on how customers should negotiate contracts with outsourcing vendors. Although the lessons seem to favor the customer over the vendor, the true motive is to establish a balance of power that benefits both parties. The initial position favors the vendor; it is typically the expert at negotiation and has experience in outsourcing, unlike the customer. The customer thus needs to leverage its position by attending to the major lessons presented in Table 2.
Discard the Vendor’s Standard Contract.
Vendors will likely parade their standard contract in front of prospective customers. The vice president and director of IS for a commercial bank that recently outsourced notes that this contract should be immediately discarded. She, as well as other participants, feels that the key to successful outsourcing arrangements is building a site-specific contract: “One thing for sure: you cannot use the vendor’s contract. It is too one-sided. The vendor gave us a generic contract, but we didn’t use it. The problem is that all deals are so different.”
The vendor’s standard contract typically obligates the vendor to perform the same level of service that the company’s internal IS department provides during a baseline period. These contracts do not, however, set performance standards or include penalty clauses if the vendor fails to meet requirements.
The payment schedules in these standard contracts may also favor the vendor. For example, an outsourcing vendor for a metals company proposed its standard contract. This contract required the metals company to pay the bill on the first day of the month, prior to service delivery. The IS consultant who assisted the metals company with the contract negotiations explains the impact of the vendor’s proposed payment schedule: “[The outsourcing vendor] wanted it day one net fifteen. We got it to in arrears net forty-five. There was a difference in a ten-year contract of $8 million. So they play those games.”
Do Not Sign Incomplete Contracts.
As both parties are often anxious for the relationship to begin, the temptation to close negotiations swiftly is strong. The out-sourcing vendor, in particular, may try to convince its customers to sign the contract before items are clearly specified. It assures customers, “We’ll take care of the details later.” But since the vendor is not legally bound to alter the contract later, it may never agree to supplement the original contract.
At a diversified service company, the CEO signed an incomplete outsourcing contract in January 1989. The vendor promised to define services, service level measures, and service level reports within the first six months. As of June 1992, these items remained incomplete, primarily because of the discrepancies over the contents of the baseline bundle. Managers at the company argue that certain services were understood to be covered in the contract. The vendor argues that if they are not already in the contract, then those services are subject to excess fees. This problem could have been avoided if the commencement date was postponed in order to complete the contract.
Hire Outsourcing Experts.
During negotiations, the vendor uses a host of technical and legal experts to represent its interests. These experts thoroughly understand the way to measure information services and how to protect their interests. In order to counterbalance the vendor’s advantage, the customer should also hire experts to represent its interests. Participants in our study noted that experts are a critical success factor in negotiating an equitable contract. Although participants admit that outsourcing experts are expensive, they believe experts will help prevent excessive above-baseline charges.
We recommend two types of outsourcing experts: a technical expert and a legal expert. Technical experts are particularly helpful when measuring baseline services. Not only do they create technical measures of the customer’s information resources, but they are also able to convert these measures to the technical idiosyncracies of the vendor’s environment. In simpler terms, technical experts convert the customer’s apples to the vendor’s oranges — a crucial skill that many customers do not possess. In addition, customers may feel wary about using their in-house technical staff to assist in baseline measures, as many of these people may be affected by the outsourcing contract.
Legal experts, who typically work in conjunction with the customer’s internal legal department, ensure that the customer’s wishes are adequately documented in the contract. Together, the legal expert and internal lawyer can pose a formidable legal team.
Customers will typically wish to hire a legal expert at the final stages of negotiation. A technical expert, however, is typically needed much sooner, particularly during the measurement of baseline services.
Measure Everything during the Baseline Period.
The customer’s current information services are documented during the baseline period, which becomes the yardstick that determines what services the vendor is obligated to provide to the customer. The outsourcing vendor will charge a fixed fee for delivering this bundle of services and an excess fee for services above and beyond the baseline. Therefore, customers must measure every service during the baseline period to ensure that these services will be included under the fixed-fee obligation.
Assuming that the outsourcing arrangement encompasses the entire IS department, customers should measure data processing, telecommunications, applications development, applications support, and residual services. Residual services include any services that are not captured in the other categories, such as user consultation, training, report distribution, and office moves. Each of these areas is briefly discussed below.
1. Data processing and telecommunication services. Of all the IS service areas, participants felt that data processing and telecommunications were the easiest to measure. Most participants used the system’s monitoring facilities (SMF) to capture resource usage during the baseline period. In IBM environments, for example, SMF data were typically used to assess baseline data processing activity. Monitoring facilities track the number of jobs submitted, the resources used for each job (tape mounts, data storage, computer minutes), turnaround time for jobs, on-line response time, and system availability. Similar reports are generated by network management systems for telecommunications.
During the baseline period, participants felt confident that their monitoring systems adequately captured their current level of processing. Resource requirements, however, vary based on the machine. Thus a company’s systems may perform differently on a vendor’s machine. Care must be taken to convert the company’s baseline activity to a comparable load on the vendor’s machine. Vendors will typically develop a conversion model based on a sample set of test transactions run on their test machine. One technical expert we interviewed advises customers to discard the vendor’s conversion model because a vendor’s test environment may vary significantly from its operating environment. The customer runs the risk of being charged excess fees if the vendor underestimates the resources required to run the customer’s systems. To avoid this risk, the customer may stipulate in the contract that the conversion model be updated after the customer’s systems are run at the vendor site.
2. Applications development and support. These service areas are difficult to measure because the activities are labor intensive. Participants in the study typically decided to use head count as the baseline measure. Thus if 200 analysts and programmers currently work in applications, the customer becomes entitled to 200 full-time equivalents (FTEs). Two hundred FTEs typically equate to 8,000 hours worth of work a week (200 people times a 40-hour workweek).
Participants, however, cited four problems with this measure. First, the vendor may eliminate people and make the remaining staff work excessive hours. The IS manager at an industrial chemicals manufacturer complained about the vendor reducing staff: “They [the remaining staff] pick up the slack, so I still get my 1,000 hours. Contractually I can’t do anything about it. But the programmers have to pick up more hours; they are tired, sick. They make mistakes.” Second, several participants complained that the quality of the analysts and programmers diminished. Vendors siphoned their best employees from the account to attract other customers. Third, included in the FTE hours are nonproductive hours such as vendor group meetings and analyst training. Fourth, the FTE does not provide a measure of productivity, merely hours worked. Because the vendor charges for hours that exceed the FTE, customers fear that they have no way to detect if vendors exaggerate project estimates.
These service problems, however, cannot be blamed on vendors. They are free to hire, fire, and assign staff any way they see fit. If the customer wants control over head counts, hours worked, project estimates, and staff quality, perhaps it should not outsource applications.27
3. Residual services. In full-blown outsourcing arrangements, companies often neglect to measure many services because (1) the services are not reflected in current IS budgets, (2) they are not currently monitored or measured, and (3) customers assume they fall within other service areas. For example, users often ask in-house analysts to help them set up their printers, verify spreadsheets, recommend products, and so forth. The analysts usually respond without documenting or charging users for these favors. However, if the customer does not document and measure these services, the vendor will not include them in the baseline.
Residual services include such items as disaster recovery testing, environmental scanning for new hardware and software, microcomputer support (purchase decisions, installation, training, repair), office relocation services such as rewiring and node changes, storage management to balance cost and performance trade-offs, and teleconferencing support. Many residual services are difficult to measure; how do you measure advice, courteous service, and environmental scanning? In most cases, these services can be measured only during the baseline period by maintaining service logs. Because outsourcing decisions often degrade morale, IS employees may ignore or sabotage the measurement effort.Therefore, some companies assign the users to maintain the service logs during the base period, rather than the IS employees.
The consequences of not measuring services were readily apparent in this study. Participants that neglected this phase all suffered serious service problems and excess charges because they failed to measure all information services during the baseline period. Some participants assumed that their requests for proposals (RFPs) documented their service needs, but RFPs are only high-level descriptions of service requirements. Baseline measures must be monotonously detailed.
The best way to avoid having services fall through the cracks is to specify 100 percent service accountability.
The length of the baseline period is also an important consideration. As service volumes typically fluctuate with the tax season, seasonal business oscillations, end-of-year processing, and so forth, we recommend a baseline period of six months. Typically, measures are calculated once a month for each service. For example, during March, the customer may use X hours of a computer resource. During April, volumes may decrease or increase. At the end of the baseline period, six observations exist for each service. The customer and vendor must then establish an algorithm to determine the baseline number. Vendors often suggest averaging monthly measures, but this results in the customer exceeding baseline services 50 percent of the time. Perhaps a more equitable solution is to create a volume variance for each service level. The customer will not be charged an excess fee as long as volumes remain within specified ranges.
Develop Service Level Measures.
Some may question why service level measures need to be developed given that the baseline period allegedly covers this issue. The answer is simple: the customer or vendor may wish to add, combine, improve, or delete measures. Baseline measures merely provide a yardstick for the vendor’s obligations during the arrangement. For every service that the vendor is expected to provide, a service level measure should unequivocally express the level of required service.
During this phase of contract negotiations, participants warn that vendors will try to manipulate measures in their favor. At a diversified services company, for example, the vendor attempted to dilute measures in two ways. It tried to dodge accounting for 100 percent of services, and it tried to manipulate the laws of probability in its favor.
Measures typically require vendors to deliver a certain amount of work in a certain period of time. For example, vendors may agree to process 90 percent of all service requests within three days. The customer, however, may never know what happens to the remaining 10 percent of the service. This 10 percent may be serviced very late or not at all. Despite the fact that 10 percent of the work may never be accounted for, vendors technically meet their service level requirements.
The best way to avoid having services fall through the cracks is to specify 100 percent service accountability. For example, if the vendor agrees to process 90 percent of all service requests in three days, then the customer should require that the remaining 10 percent be completed within five days and that any exceptions be fully documented and reported.
Vendors may also dilute measures by exploiting some simple laws of probability. Returning to the outsourcing experiences of the diversified services company, the vendor finally agreed to deliver 95 percent of a particular service within the agreed-on time frame. The company agreed to the measure, as long as the service was delivered correctly. The vendor countered with a proposal to implement two measures: 95 percent of the service would be completed by the target date and 95 percent of the service would be accurate. These two measures would dilute the service level because the probability of the service being delivered on time and accurately is only 90 percent.
Develop Service Level Reports.
During outsourcing negotiations, companies may spend a significant amount of time developing measures but fail to require the vendor to report on these measures. Vendors may tell their customers that their standard reports address the measures, but several participants found this assertion untrue.
One participant, for example, complained that the vendor’s standard reports indicated only the volume of service performed. For instance, the vendor’s security request report indicated “100 security requests were implemented this month.” This report did not specify how many security requests were submitted or the average turnaround time for the requests. According to the contract, the vendor was meeting service levels. Users, however, complained that security requests sometimes took more than two weeks to process. They felt this was too long to wait for a log-on identification or access to a needed data set.
Service level reports should document the agreed-on service level, the service performance for the current time period, exception reporting for missed measures, and a trend analysis of the performance from previous reporting periods. Customers are often charged for the creation of these reports. The investment is well worth it — how can service level measures be monitored without service level reports?
Specify Escalation Procedures.
Customers realize that IS is often a volatile business — there are bound to be occasional events that prevent the vendor from meeting a service level measure. In some instances, the customer may even be at fault. Thus, in addition to the service level reports, the customer and vendor must agree on problem escalation procedures.
Typically, the vendor will request that fault (customer or vendor) be determined for each missed measure. This protects the vendor’s interests; because it is contractually bound to meet measures, it should not be punished for customer errors. Perhaps a bipartisan committee will determine blame for missed measures. Granted, this task is repulsive to most; as professionals we want to fix problems, not fix blame. However, the reality is that dollars may be exchanged as a result of nonperformance.
Services may be divided into critical and noncritical categories to prevent micromanagement. For noncritical measures, such as analyst training hours, perhaps the vendor may miss this measure once or twice a year. For critical services, such as on-line availability, the customer may require immediate reporting, problem resolution within a specified period of time, and perhaps even a cash penalty.
Include Cash Penalties for Nonperformance.
In cases of severe service degradation, the customer may insist on cash compensation. A commercial bank, for example, has an agreement with its vendor that includes penalty charges for failure to meet end-user response time, system availability, and batch delivery deadlines for critical systems. Customers may also wish to escalate cash penalty amounts with frequency. For example, the first occurrence may result in a penalty of $25,000. Another occurrence for the same service within a specified time period may cost the vendor $50,000.
Participants that specified cash penalties in their contracts hope that they will never need to enact the penalty clause. The chief financial officer at a transportation services company notes that penalty clauses do not fully compensate the customer. Rather, the purpose of penalty clauses is to ensure that the vendor’s senior management will attend to service level problems:
You don’t get total reimbursement for your lost profit and your lost cash. But you do have a penalty that is significant to [the outsourcing vendor] and gets their attention. Our penalty is in the $100,000 range up to $1 million. So it’s not enough to compensate us for the downtime, but it will certainly get their attention. It will get somebody fired. You know it will get to the top layers within [the outsourcing vendor].
Determine Growth Rates.
Most outsourcing contracts include a growth rate whereby the customer gets a certain amount of growth for free. The reasoning is that the cost of a unit of processing decreases every year, and the customer deserves to share the benefits of price-performance improvements. The problem, however, is that the vendor managers understand growth rates much better than the senior executives with whom they typically negotiate. The vendor may underestimate growth so it can charge excess fees in the future.
For example, the vendor may state that the customer’s growth rate is 5 percent to 6 percent per year, based on IS budget increases. The vendor then offers to provide a 6 percent increase in resource requirements (MIPS, storage, tapes) free of charge. To a customer unfamiliar with the intricacies of IS, this deal seems appealing. However, in actuality, the resource requirements may be growing at 10 percent to 20 percent because the IS department is able to exploit price-performance improvements. During the outsourcing arrangement, either the customer will pay extra for resource growth above 6 percent or it will curtail growth.
Adjust Charges to Changes in Business.
Customers should also include a clause for severe volume fluctuations caused by acquisitions, mergers, or sales of business units. In the case of a business unit sale, the customer may specify a major reduction in the fixed-fee expense. The vendor, however, may insist on several months’ notice to allow ample time to redirect resources. The customer may also want the vendor to promptly accommodate mergers or acquisitions. The vendor may again insist on advance notice. In addition, the vendor may insist on charging the customer a transition fee for the volume adjustment.
Select Your Account Manager.
At a mining company, the controller insisted on one truly unique feature: he specified the name of the account manager in the contract. The controller had so much faith in this particular individual, one of the vendor’s account managers, that he required him to manage the arrangement. This account manager is indeed a special character; as a previous outsourcing customer, he was once the vendor’s most vocal critic, often complaining about services and fees. The vendor finally hired him to “shut him up.” The controller felt that the account manager’s background as a former outsourcing customer would prevent opportunism. His choice was wise; after one-and-a-half years with the vendor, the mining company has never contested the vendor’s bills or service. Although the contract seems weak in other areas (lack of service measures and penalty clauses), the controller’s utter trust in the account manager compensates for any legal loopholes.
Include a Termination Clause.
Most lawyers will insist that a termination clause be included in the contract. This clause protects both parties, as the desire to terminate by one party will severely affect the other. Either party may need to terminate because of bankruptcy or sale of the company. In addition, the customer may wish to terminate because of failure to provide services. Most contracts require either party to notify the other within a specified time period, such as three months. Failure to give adequate notice may result in a severe penalty charge.
The customer, however, will typically require more than three months to find an alternative to meet its information needs. Negotiation with another vendor may require six months; rebuilding an internal IS department may require a year or more. In addition, the customer needs the vendor’s assistance to transfer systems to an alternative site. Therefore, vendor assistance should be specified as a requirement for termination —regardless of the initiating party.
Watch Out for Change-of-Character Clauses.
Another weakness of several outsourcing contracts is the change-of-character clause. This provision states that the customer will be charged for any changes in functionality. This clause has triggered several disputes. At a diversified services company, the vendor wanted to charge for changing its word processing software. The vendor argued that this represented a change of character, as the new product was not supported during the baseline period. The customer argued that this was not a change of character; the function, word processing, had not changed, only the software.
Personal computers (PCs) are another point of contention. The same company’s contract states that the vendor will service all personal computers for a fixed price. However, the number of computers has doubled since 1989, and many PCs are now connected to local area networks (LANs). The vendor claims that LAN technology is a change in character, whereas the customer claims it is only a difference in technology. In addition, the outsourcing vendor wants to charge a particular dollar amount for each additional PC supported.The customer contends that “volumes do not equal costs.” In other words, doubling the number of PCs does not require double the cost to support them.
Take Care of Your People.
The discussion so far has concentrated on protecting the customer’s interests (we assume the vendor is capable of protecting its own interests). The collective term “customer,” however, excludes many of the organization’s people. In particular, the IS employees will be dramatically affected by the outsourcing decision. Companies have a social responsibility to treat these people fairly. That includes informing them of the final decision as soon as possible and helping them secure positions elsewhere if necessary.
In typical outsourcing arrangements, the vendor will hire the majority of IS employees for a one-year trial basis. This allows the employees to prove themselves to the vendor before the vendor commits to permanent employment. The vendor may also request performance ratings for each analyst from the customer. If the customer views outsourcing as an opportunity to eliminate low performers, it may be tempted to share this information. However, customers should examine the ethical implications carefully; a person’s career is at stake. Furthermore, the customer may be pleasantly surprised to find that low performers thrive in another culture.
Conclusion
Practitioners who have yet to face outsourcing decisions can learn from the successes and mistakes of the companies in our study. In particular, do not outsource merely because vendor bids are more desirable than internal IS bids. Ask yourself: Why is the vendor’s bid better? Could we achieve these results on our own? If not, how do we ensure that the vendor keeps its promises? We conclude with the following thought: whether or not a firm decides to outsource its IS function, the management of IS cannot be outsourced.
References
1. P. Krass, “The Dollars and Sense of Outsourcing,” Information Week, 26 February 1990, pp. 26–31;
J. Rochester and D. Douglas, eds., “Taking an Objective Look at Outsourcing,” I/S Analyzer 28 (1990): 1–16;
R. Hamilton, “Kendall Outsources IS Chief,” Computerworld, 13 November 1989, pp. 1, 4;
G. Anthes, “Perot Wins Ten-Year Outsourcing Deal,” Computerworld, 8 April 1991, p. 96;
C. Wilder, “Bend Me, Shape Me,” Computerworld, 24 December 1991, p. 14; and
S. Huff, “Outsourcing of Information Services,” Business Quarterly 55 (1991): 62–65.
2. Network World, 17 February 1992, pp. 1, 31–36.
3. J. Greenbaum, “Planning for Pan-European Outsourcing,” Information Week, 22 June 1992, pp. 40–48.
4. E. Kass and B. Caldwell, “Outsource Ins, Outs,” Information Week, 5 March 1990, p. 14.
5. J. Ambrosio, “Outsourcing at Southland: Best of Times, Worst of Times,” Computerworld, 25 March 1991, p. 61.
6. E. Gardner, “Going On-line with Outsiders,” Modern Healthcare, 15 July 1991, p. 35.
7. B. Ward, “Hiring Out: Outsourcing Is the New Buzzword in the Management of Information Systems,” Sky Magazine, August 1991, p. 40.
8. M. Hopper, “Rattling SABRE — New Ways to Compete on Information,” Harvard Business Review, May–June 1990, p. 119.
9. R. Huber, “How Continental Bank Outsourced Its ‘Crown Jewels’,” Harvard Business Review, January–February 1993, pp. 121–129.
10. L. Loh and N. Venkatraman, “Diffusion of Information Technology Outsourcing: Influence Sources and the Kodak Effect,”Information Systems Research 3 (1992): 334–358.
11. A. Radding, “The Ride Is No Bargain if You Can’t Steer,” Computerworld, 8 January 1990, p. 70.
12. V. Hovey, Presentation to the University of Houston’s Information Systems Research Center, 22 January 1991.
13. Loh and Venkatraman (1992).
14. Radding (1990), pp. 67, 70–72.
15. O. Williamson, Markets and Hierarchies (New York: Free Press, 1975).
16. For circumstances in which the term “partnership” is applicable, see:
J. Henderson, “Plugging into Strategic Partnerships: The Critical ISConnection,” Sloan Management Review, Spring 1990, pp. 7–18.
17. Krass (1990);
D. Livingston, “Take My System — Please!” Systems Integration 23 (1990): 40–44;
Radding (1990); and
Huff (1991).
18. T. Barron, “Some New Results in Testing for Economies of Scale in Computing,” Decision Support Systems 8 (1992): 405–429.
19. A. Friedberg and W. Yarberry, “Audit Rights in an Outsourcing Environment,” Internal Auditor, August 1991, pp. 53–59.
20. J. Moad, “Large-Scale Systems Survey,” Datamation, 15 May 1988, pp. 56–64.
21. Friedberg and Yarberry (1991).
22. T. Mason, Perot (Homewood, Illinois: Dow Jones-Irwin, 1990).
23. L. Loh and N. Venkatraman, “Stock Market Reaction to Information Technology Outsourcing: An Event Study” (Cambridge, Massachusetts: MIT Sloan School of Management, Working Paper No. 3499-92BPS, November 1992).
24. R. Benjamin and J. Blunt, “Critical IT Issues: The Next Ten Years,” Sloan Management Review, Summer 1992, pp. 7–19.
25. Huber (1993), pp. 122–123.
26. Huber (1993), p. 129.
27. W. Richard and A. Whinston, “Incomplete Contracting Issues in Information Systems Development Outsourcing,” Decision Support Systems 8 (1992): 459–477.