This chapter recommends methods and procedures for acquiring information technology and related services. It is organised into three main parts:

• Linking technology acquisitions to state goals and objectives
• Establishing the internal rules for technology acquisition
• Standardising interactions with technology vendors

Linking Technology Acquisitions to State Goals and Objectives

    The first step in acquiring an untested technology is to assess the value of the new technology in the specific context in which it will be used. Technology Assessment, with roots at the intersection of the appropriate technology and decision science movements, grew from the realisation that not every technology or innovation represents progress for a society. First, the range of available technologies increases exponentially, so there are now simply more choices to be considered for adoption. Second, the adoption and use of new technology brings with it not only the expected benefits but various costs, some of which may be beyond the social, economic, environmental, or legal tolerance limits. Also, when the resulting costs and benefits are not equitably distributed, the political costs tend to rise. These negative outcomes are actually symptoms of the inappropriate selection and use of technology.

    In response to this problem, governments developed techniques to identify and evaluate the most important consequences of the use of new or modified technologies, often to minimise the undesirable effects of technological development on the society in general, and more specifically, to reduce the negative effects, both direct and indirect, of the use of technology on specific domains such as the natural environment. In recent years the USA, the Federal Republic of Germany, Japan, and the Republic of Singapore have used this technique, in one form or another.

    In many advanced countries, technology assessment techniques have diffused to private industry, where they are used to analyse the effects of changes in products, production technologies, production sites, and major investments or radical changes in business policy. The health care and pharmaceutical industries, which were required to perform such analyses, were the pioneers in private sector technology assessment. The industries that have experimented with technology assessment have indicated its usefulness to explore "questions that cannot be answered by traditional economic or market analysis." [Compton, 1978]

    Technology capability assessment represents a method for identifying and comparing the likely impacts of both locally developed and imported technologies on the social, economic, and natural environment, and is especially useful for small or developing countries. The technique provides a means for targeting technologies for public investment according to the values held by the society, its development needs, and its capacity for absorption. Further, it allows identification and examination of the risks inherent in the use of new technology.

    Because environmental, economic, cultural, social, and political factors vary from country to country, the consequences of the use of a technology in one country may vary greatly from its use in another. Also, technology transfer, which involves social, cultural, and environmental change, is inherently disruptive. Therefore, the short-term impacts of technology transfer are often negative, which creates resistance, which may be overcome only if decision-makers are aware of the long-term benefits.

    The several competing technology assessment methodologies have certain elements in common, and tend to follow a generic cycle beginning with the definition of the technologies of interest to the stakeholders, and proceeding to an analysis of the comparative impacts of alternative technologies on these stakeholders prior to linking these impacts to the decision at hand, often some form of public or private investment.
 

Technology Definition and Projection

    The first major element is to define, describe, and project the development of a technology (or group of competing technologies) from the present to a specified future time. This requires acquiring data, bounding the assessment domain, projecting the patterns of development and use along alternative paths, and providing outputs which can be used in later steps to compare the impacts, risks, and other likely consequences of the adoption and use of each technology under consideration. This methodology was used to develop the material presented in Chapter 2 of this report.
 

Impact Analysis

    The impact analysis element of the TA cycle identifies and evaluates the consequences of interest flowing from the adoption and use of specific technologies. The first task is to select criteria that capture the relevant impacts of technology, followed by prediction and assessment of these impacts, then comparing these across technology choices.
 

Selecting impact criteria

    This is a critical activity, and shapes the entire assessment effort. These criteria form a list of items to be used to compare and evaluate the consequences flowing from the use of alternative technologies. While there are many methods for identifying and organising impact criteria (by scientific discipline, stakeholder, or functional activity), the common tasks are to (1) list all possible direct and indirect impacts, (2) organise them coherently, (3) select those which appear to be significant to the stakeholders, and (4) convert these into criteria. As there are no methods that guarantee completeness, it is wise to subject them to a thorough internal review. The recommended impact criteria for the state government were derived using the balanced scorecard approach (see Table 5).
 

Predicting and assessing impacts

    The domains for impact assessment include the economic, social, and physical environments. The task is to design and develop mechanisms that will reliably predict the relationships between use of the technology and the intensity of its impacts. Here, there are many competing and complementary techniques, including expert opinion, conceptual, quantitative and analogue modelling, trend analysis, cross-impact analysis, and scenario analysis.

Risk Analysis

    Although it may be performed as a part of the modelling activity, risk analysis involves a set of techniques quite different to the other tasks, and is therefore treated separately here.

    Its underlying techniques integrate the use of logic and mathematical models to estimate the probability that various features (e.g., benefits, costs, effectiveness, or equity) of the technological systems under review
 

Comparing the Impacts of Alternative Technology Choices

    Once these steps have been taken, there would be sufficient data to compare the likely outcomes.
 

Capability Assessment

    Capability assessment (in the private sector, often called absorption capacity) indicates the state’s ability to acquire and use technology.
 

Policy Analysis

    The objective of the policy analysis element of TA is to present decision-makers within the applicable policy community a comparative analysis of the currently feasible policy options, and their probable impacts or consequences. [Armstrong and Harman, 1980] The organisation and presentation of the findings is critical to the success of the technology assessment. The ITMP provides a draft policy, which must be discussed and adopted, then gradually amended as the context evolves over time.
 

Requirements Definition and Specification

    Once the role of technology has been defined, the state is in a position to define its requirements for technology and convert these into specifications through which products and services can be selected and acquired.
 

Requirements specification and validation

    This section is a brief generic overview of the staff work which must be completed before requirements are understood well enough to proceed with acquisition.
 

Requirement levels

    Requirements exist at three levels: strategic, operational, and technical, mirroring the structure of benefits.

Strategic business process requirements

    Needs for technology which (1) originate at the policy level and (2) affect the performance of the organisation are the primary and most influential sources for defining systems needs. For a public safety unit, an online dispatch support system may be a critical requirement, while for the State Treasurer, a point-of-sale system which helps control cash receipts may be essential. At this strategic level, only a broad non-technical statement of what is desired is necessary, as more detailed requirements will emerge from the next stage, while actual requirements come much later in the process.

Operational requirements

    Operational requirements emerge from investigations into the actual information flows required to meet strategic requirements, and reflect both the technological and organisational changes necessary to meet these targets. For example, it may be necessary to reschedule personnel to meet the rising demand for a specific service resulting from the improved service levels which are expected to follow the implementation of a new purchasing system, and this may in turn require other changes like new job descriptions or administrative regulations. These issues must be clarified before moving forward.

Technical requirements

    Technical requirements identify the direct resources (terminals or workstations, data storage, input devices, local printers, etc.) required to carry out the intended function, plus the technical and organisational infrastructure necessary to support these resources. This is normally an iterative process, in which the resource implications are fed back to the project sponsor, who may choose to modify the requirements rather than justify added resources.

Interface requirements

    It is mandatory to establish clear standards regarding data formats for systems that either feed or acquire data from other systems, and desirable to establish standards for screen formats, keyboard layouts, and other man-machine interactions.

Requirements definition

    While acquiring simple technology (a router, workstation, or printer) involves a fairly simple statement of specification geared to industry-standard products, the acquisition of a complex technology package (a turnkey POS network, for example) involves much more preparation and a far more detailed specification.

    Mandatory requirements are those that cannot be omitted from the final system, while value-added requirements are those which are not absolutely necessary, but which would provide benefits if they were available. For example, a purchaser of a new PBX switch might demand a capacity of at least 600 lines, with direct inward dialling, but would find call forwarding and call-waiting facilities desirable if the marginal cost for these enhancements is low. To find out about these issues, it may be necessary to communicate optional requirements to vendors.
 

Procurement Planning

    Ideally, the acquisition process is a disciplined series of risk-avoiding steps, each with clear accountability. The critical success factors in procuring major IT systems (those which consume large amounts of scarce financial, human, and technical resources) are to focus on structural simplicity, standardisation, and inheritance. If these goals can be achieved, the resulting system is far more likely to meet state functional requirements while providing the reliability, affordability, and supportability needed to ensure a long useful life.

    The critical success factors for procuring smaller systems and components are responsiveness, compatibility with installed technology, and life cycle cost. For both major and smaller systems, uncomplicated designs using standard technologies cost less, enhance reliability, and reduce requirements for training and maintenance. Also, to the extent that it is possible to define standards in advance of major procurements, such "homework" shortens the acquisition life cycle and reduces uncertainty on the part of both the vendors and the state procurement team.
 

Build or Buy?

    Systems development is the one activity most likely to generate user concern and dissatisfaction, while the development process is typically plagued by missed deadlines, cost overruns, and delivered products with brief useful lives. These inherent risks make purchasing packaged applications attractive in many cases. Even when a system must be developed from scratch, it must be decided whether to develop the systems using in-house resources, or to acquire resources from an external source. The factors that must be considered are:

• Internal capability to do the project, which includes not only technical skills, but also the top management support required to structure the project and see it through
• Financial capability to find cash to pay for external development inputs
• Project management abilities, which will be required for either alternative
• Flexibility in the specifications, allowing for the purchase of standard products
• Operational arrangements, which might involve an external facilities manager, in which case it usually makes little sense to develop a system internally

Managing Vendor Relationships

    While vendors are a valuable supplementary resource, their proper management also consumes resources. The correct relationship with a vendor provides more scarce resources than it consumes. To achieve this, the relationships between the state government and IT vendors should be channelled through one of the three proposed competence centres: SCSD, SSL, or KKIP-C. The latter firm should position itself to develop strong vendor relationships and act in partnership with those vendors who can help it build the specific competencies needed to reach the high level of network management and information packaging competencies required to carry out its mandate.

Communications with vendors

    Formal communications paths should be defined between the state and its IT vendors, which will increase transparency and help each centre of competence acquire the stable support needed to move forward. Generally, day-to-day communications focus on supporting the technology in place, so vendor support staff should be able to contact users directly. However, identifying new requirements is a marketing function, for which one of the three centres of technology must be in the loop. The STU should play a support role for all technology acquisitions involving the direct expenditure of public funds from the State Treasury, and should monitor communications with vendors throughout the acquisition cycle.

Preferred vendor policies

    In the Southeast Asian context, it is often difficult to find an IT vendor with the capacity and willingness to provide the desired level of responsiveness and support. In other cases, a specific technology is available only from one source. In such cases, a preferred vendor policy is one expressing management expectations of lower overall costs, improved service, or other benefits through a policy which focuses on an alliance with specific vendors.

    Of course, technology changes constantly, so a good way to manage the dynamic aspect of a preferred vendor policy is to automatically evaluate any new technology for inclusion into the policy. This keeps the policy open, thus ensuring that the state is not committed to obsolete technology. If a preferred vendor policy is selected by the state, it should be formalised, and must reflect the following elements:

• Policy objectives, in terms of expected discounts, support levels, etc.
• Identification of preferred vendors, perhaps expressed as "The state will source its (specific technologies) primarily from vendors HAL and ABC, but will reconsider this policy prior to each budget cycle," or alternatively "The state will source its (specific types of technologies) primarily from vendors HAL and ABC, but will choose the lowest cost compatible products for (specific technologies)"
• The explicit process by which a vendor is considered eligible for preferred status
• Criteria for build or buy decisions
• The standard criteria used for proposal evaluation in tender processes
• The organisational unit (STU) responsible for monitoring vendor relationships

    Expertise is required to search for and qualify vendors, specify needs for technology, and negotiate, monitor, and enforce agreements. This implies that the state should minimise the number of parties that conduct transactions for the more complex elements in IT systems.
 
 

Establishing the Internal Rules for Technology Acquisition

    Vendor selection and management involves identifying and comparing the ability of competing vendors to meet specific organisational needs.
 

Developing a Vendor Inventory

    The first step is to develop an inventory of available vendors who offer the technologies required by the state. An effort should be made to qualify vendors, based on publicly available information, which should focus on the following elements:

Technology offered

• Product / service list
• Assessment of technology
• Delivery capability
• Installation and customer support capabilities
• Documentation and training

• Knowledge of state government
• Organisation and staffing
• Quality control
• Management and personnel turnover
• Personnel technical skills

• Annual reports (publicly listed companies)
• Credit history

• Market share
• Competitive position: technology, management, earnings

• Client assessments
• State history: previous purchases, tenders, etc.

Procurement Process Options

    The procurement process includes all activity necessary to acquire technology from a supplier, following the decision to source rather than develop internally. While such process may either be competitive or non-competitive, it must conform to national and local laws governing expenditures of public funds.
 

Non-competitive options

Sole-source procurements

    Sole-source procurements are common when the technology to be acquired is a proprietary component of a larger system, such as new terminals for a mainframe computer, or when the technology is part of a complex package of hardware, software, and services, as in an outsourcing contract.

Unsolicited proposals

    Unsolicited proposals reflect the natural desire of vendors to find new places to place their goods, and are useful for the customer because they provide information which help clarify requirements. A vendor may formalise this interaction by preparing a proposal that attempts to bypass a competitive bidding situation. In some cases, this may provide real benefits to the buyer, such as lowering the cost of acquisition, or provide access to technology early in its life cycle. Normally, such proposals are very difficult to fit into the restrictive public sector purchasing culture.
 

Competitive options

Request for information

    The request for information (RFI) is a formal notification that the state intends to acquire technology, and alerts the vendor of your serious intent, while providing for the systematic gathering of information that can be used to develop specifications. Normally, the RFI will screen out vendors that are not really interested in getting your business, and are not binding on either the state or the vendors.

Request for quotations

    The request for quotation (RFQ) is a formal request for a firm price for a standard, off-the-shelf product that requires little or no adaptation prior to installation and use. The RFQ content normally corresponds closely to industry-standard specifications (type of microprocessor, bus, or clock speed) for the products desired. In certain cases, the scope of an RFQ can be expanded to include specific services that are closely related to the product, as when a microcomputer system is purchased as a package with its operating systems, standard applications, and a maintenance contract.

    However, RFQs cannot be used for high-technology products which require extensive customisation, because such acquisitions typically require a far more extensive exchange of information, and are thus more suitable for either a tender (RFP) or negotiated acquisition. From the legal perspective, a quotation normally commits the vendor to provide at the price quoted without committing the purchaser to buy.

    An important variation on the RFQ model is the benchmark pricing approach, which establishes the price of the good to be provided based on relevant indexes, such as the cost of labour or the wholesale price of a standard memory chip, etc. This approach is suitable for establishing standard prices for IT products whose inputs are highly subjected to a combination of shifting market forces and rapid technological change.

Public tender (request for proposal)

    Usually, public tendering is intended to provide transparency while creating at least an appearance of efficient and orderly expenditure of public resources. In fact, tendering has been shown to be efficient only under certain narrow circumstances, which are not always present in the acquisition of information technology and the related support services. While tendering is one alternative among the competing approaches to acquisition, it is not entirely suitable for every acquisition. The theoretical advantage of the tender process is its ability to simultaneously create a competitive environment among vendors, while providing for future co-operation between the winning vendor and the buyer. However, because both the process costs and the delay between the initial request and delivery are potentially high, this approach is not suitable for fast-track projects.
 

Fitting Process to Requirement Type

    The type of technology to be purchased, and the degree of certainty regarding the specification can categorise requirements.
 

Software

    Software purchases can be categorised along a broad continuum. Shrink-wrap systems (e.g., Windows) and applications (e.g., Quicken) software is essentially a freely traded commodity, and is best purchased either through the quotation or contract methods. However, as software features change frequently, the market price for an obsolete but still useful version may fall dramatically with the introduction of the latest release. Tailored software is a highly differentiated product, which must be acquired through negotiations, which may be either competitive or sole-source.
 

Desktop machines and servers

    Personal computers are dynamic commodities as features and prices are subject to frequent change, which makes it impractical to use a purchase contract unless the price can be tied to some reliable benchmark. Because the microprocessor, memory, and hard drive make up a major part of the cost of a PC, and these components are widely traded, they can be used as the basis for a composite benchmark (which must still be adjusted for performance level and brand). This can be the basis for an index, through which the price can be adjusted. For example, if the current wholesale price of a 1-gigabyte hard drive is RM300, while a 200 MHz MMX-enabled Intel P55C processor sells for RM700, the price for a multimedia desktop machine might be negotiated at MR4500. A contract can be established to link the price at a given time with the current wholesale price of the major components, which are posted at various sites on the Internet: if they fall by 20 percent from the base period, the price paid to the vendor for this type of product could be adjusted proportionally.
 

Systems

    Because mini-computer and mainframe systems are far more complex than desktop machines, the purchasing approach will depend on the specific situation. For example, if the state wishes to upgrade an existing proprietary system, a sole-source negotiation may be suitable, but if the intent is to replace entirely obsolete systems with a new generation of technology, a tender process will be more appropriate.
 

Standardising Interactions with Technology Vendors

    The best results are found when the interactions with vendors follow a fairly predictable pattern, which requires a well-developed set of processes to qualify the vendors, communicate with them, and evaluate their quotations and proposals,.
 

Vendor qualification

    A qualified vendor is one that meets the minimum requirements in technology, support services, management, and financial viability established by the state. These may change over time.
 

Outsourcing

    In 1989 Kodak startled the business community by announcing it would farm out a large part of its information systems (IS) activities to three companies: IBM would operate the data centres and local data communications network; Digital Equipment Corporation would manage world-wide telecommunications; and Business Land would provide support for personal computing. Since then, the company has begun to look for alliance partners to develop and maintain some of its applications as well. The impact of the Kodak outsourcing contracts was such that one conference on the subject divided IT outsourcing into two eras — one pre-Kodak and the other post-Kodak.

    While the popular press focuses on the wholesale take-over of a organisation’s entire IS function, the evidence shows a more finely tuned approach is required for good results. For example, it may make sense to outsource development while retaining operational systems development in house. Or it might be sensible to arrange for an outsider to own and operate the network, while retaining control over decisions regarding which technologies will be used. Different third party relationships will be most appropriate, depending upon the impact that a specific aspect of IT has on the firm, and the competence the firm displays in executing that aspect.

    While experts disagree on exactly how many firms have chosen to outsource part or all of their IT function, few would dispute that the phenomenon is growing rapidly. Although the projected annual growth rates of 15-20 percent will push the global total over US$100 billion by the year 2000, capacity in Asia is severely limited.

    Outsourcing includes not only facilities management, but also time sharing, network management, telecommunications, service bureaus (e.g., payroll processing), technology planning, IS management consulting, systems integration, application development, etc. This wide range of services implies three types of relationships between the company and the provider of IT services.

Occasional, short term contracts that cover a well-defined task over a short period of time. Examples include the purchasing of a software package or the hiring of consultants to develop an application, as a way of supplementing or replacing in-house IT professionals.

Major, long term contracts that cover a substantial part of a firm’s IT capability. For example, Continental Bank’s outsourcing of data centre operations and applications development to IBM and Ernst & Young falls into this category.

Various joint venture and partnership arrangements in which the two parties substantially share the cost and risk of various IT activities. In these arrangements, each party looks beyond financial justification as, for example, when a vendor tries to develop expertise in the company’s industry, while the company is seeking exposure to new technologies or management practices.

    Kodak executives emphasised that their primary goal was not to reduce costs, but to free up time and attention so that the company can focus on its strategic priorities. Similarly, after two years’ experience with outsourcing data centre operations and a nation-wide network, American Standard downplays cost savings (which exceed 25 percent) and emphasises the value of being able to focus on more important business issues. And by turning over the operation of its data centre to Citicorp Information Resources, Talman Home Federal Savings & Loan freed up its own staff to develop more strategic customer-oriented services.

    In addition to these examples of "back room" outsourcing where undifferentiated, commodity-like activities are turned over to a third party, an increasing number of firms are working with outside companies on projects of strategic importance. At a large mutual fund group, for example, management believed that critical shareholder services could be enhanced if a new client account system could be developed. Such a system is very large and complex, and the only financially feasible solution was to develop the system using the absolute cutting edge in software development tools. Because internal IS professionals had little experience with either the tool set or a project of this scope, the company created a joint venture with an outside firm to create the needed system. The outsiders provided the expertise that was needed, and the joint venture format allowed the company more control over this critical project. Interestingly, the company’s IS organisation was completely against using the outside company, but the CEO overrode their objections and proceeded anyway.

    In a similar situation, a large financial service provider commissioned a third party to develop an enormous database application to store and manage information on tens of millions of households in North America. Because the requirements of such a system were poorly understood, and because the required data base technology was unfamiliar to the company’s IS professionals, they decided to develop the system through an alliance with a software development vendor. The difference between this arrangement and a more traditional arms-length contract is largely a matter of degree, but an important feature is the lack of precision regarding the deliverables and the commensurate increase in the need for true co-operative effort.

    Not all outsourcing works out favourably, of course, as several examples illustrate. In one case, a medium-sized bank decided to hire a large and well respected vendor to take over all of the IS activities formerly handled in-house. The key reason for this decision was to cut total IS expenditures by 15 percent. The vendor argued that these savings could be achieved because they enjoyed economies of scale which allowed them to pass substantial savings on to their customers and still remain profitable. Unfortunately, the bank found it much more difficult to specify service levels then cost savings, and most users of IS grumbled that the lower costs were achieved at the expense of quick response to their needs. Eventually the bank brought some of the more critical functions back in-house.

    In another example, Bank of America moved away from using a service bureau for securities settlement processing and set up its own system in-house. The service bureau could not provide the kind of broker/dealer capabilities that the bank wanted, and as one of many customers of the service bureau, the bank could not dictate the development of new functions. Interestingly, the bank did not attempt to develop a new system in-house, but instead purchased hardware, software, and services from outside vendors. In one sense, this example might be seen as an outsourcing relationship gone awry, but in fact it might better be explained as a move to a more "appropriate" outsourcing relationship for the bank’s needs.

    For NHP Inc., however, outsourcing was a decidedly negative experience, and one that led to lawsuits on both sides. When NHP finally returned their IT function from an outsource vendor to full in-house operation, they realised a 52 percent reduction in IT costs — the reverse of the traditional wisdom about economies of scale and lower cost. More importantly, they regained control over IT staff and operations, adding people who understood their business and who cared about the company’s performance.
 

Back to Basics — Focus, Quality, Productivity, and Flexibility

    In many ways, outsourcing is just a new term for the old "make versus buy" decision. But three forces are at work which make this decision more critical, and thus of greater concern to senior officials, than ever before. These forces are:

• a renewed strategic focus on that which is essential to the firm’s success, and the discarding of anything that is peripheral
• the extension of quality and productivity principles to all aspects of the firm’s operation
• the need for flexibility rather than absolute lowest cost, and thus the trend away from deep vertical integration

    If the 1960s and 1970s were the age of diversification, many companies spent the 1980s restructuring to rid themselves of extraneous operations and concentrate on what they do best. Kodak, for example, has recently divested more than US$1 billion in peripheral businesses. In the same spirit, companies have realised that vertical integration is not as necessary as they once thought. With advances in information and communications technologies, they can manage the workings of the supply chain without owning every link. Inspired partly by the Japanese example, firms have learned how to work more closely with their suppliers and customers.

    In effect, these companies have decided to concentrate on the essence of their business and to transfer nonessential activities to others. Reebok provides a good example. The essence of its business is design and marketing, not manufacturing — a function that Reebok outsources. Unlike traditional shoe manufacturers, its reliance on others enables the company to flexibly respond to major market shifts. Reebok can turn quickly from one sneaker maker to another if demand for canvas shoes burgeons or inflatable sneakers collapse — they are not stuck with plant and equipment that has been "optimised" for one sort of production and which must be completely retooled for another.

    Through a process of disaggregation and electronic re-integration, companies are reassigning production and distribution processes to multiple parties and still maintaining (or even improving) control. With each player specialising in the activities it does best, the highest possible level of quality is ensured.

    Outsourcing IT is part of this pattern of putting tasks in the hands that can perform them best. Many firms have difficulty attracting and retaining IT professionals of the very first rank. Like other specialists who are in great demand, IT wizards are more likely to gravitate to a professional environment in which their expertise constitutes the essence of the business. In most organisations, they often play a supporting role. Where IT is a core competence, as for example with an IT vendor, however, they will be the stars, not the drones. Under these circumstances, it would not be surprising to find the very best IT experts gravitating into specialist professional firms. Top managers recognise the same pattern at work in other fields. When they need the very best legal advice, they do not ask their in-house staff to go it alone. Instead, they call in the specialists from a major law partnership.

    Peter Drucker echoed this view several years ago, when he commented:

    In the past, outsourcing has often been motivated by a sense that IT costs have become too high, or that some aspect of IT is broken and needs to be fixed. The real value of IT outsourcing, however, is that it enables the organisation to obtain the highest-quality IT services available, and to focus on the essence of the organisation’s business. In fact, the most powerful IT outsourcing message may be that strategic focus, quality and flexibility are the critical competitive objectives, and that we should not always assume we can "do it better ourselves."

    Because IT activities are increasingly vital to service delivery, outsourcing is sometimes perceived to pose unacceptably large risks. Three specific objections may be raised to outsourcing proposals, which we will consider in turn:

• outsourcing IT activities leads to the loss of control of an important function
• if the outsourcing vendor fails to deliver, it will be difficult to reverse the decision
• outsourcing vendors are outsiders, and may not have sufficient knowledge of agency business processes and success factors

Control

    Critics of IT outsourcing argue that no outside vendor can match the responsiveness and service levels offered by an in-house IT function, largely because the outsider is not subject to the same management direction and control as employees. In addition, there are often concerns about the confidentiality of data and strategic applications, and about provisions for disaster recovery with an outside vendor. These anxieties about control might be summed up under the rubric "nobody else cares about our business the way we do."

    From a purely legal perspective, adequate control might be achieved only if the contract specifies in detail how an IT outsourcing vendor must perform, and what penalties can be involved if performance is sub-par. However, two major difficulties present themselves.

• Measurement — It is actually very difficult to precisely specify and measure the performance of a vendor of complex services. What does "good service" mean, and how will it be measured?
• Uncertainty — For many complex IT activities the requirements of "good performance" change over time. Because the traditional control on vendor performance has been largely contractual, it means that all of these uncertainties must be captured in the contract language. With employees, it is much easier to redirect them if changes in the business warrant a new set of activities.

    Many managers feel more comfortable outsourcing other back office functions because performance controls are easy to specify, and a failure to meet the standards set in the contract is not disastrous. It is fairly easy to define performance criteria for a mailroom (collection and delivery time, and so on), and these standards are stable over time. A failure to achieve contracted performance levels is not disastrous, and is easily corrected.

    Contracts alone provide an inadequate control mechanism for outsourcing complex, technology-dependent, and potentially critical activities such as IT. A relationship must be created so it is in the interest of both parties to seek solutions to unanticipated problems. In many cases, a vendor’s need to maintain its reputation is a reliable incentive for superior performance.
 

Reversibility

    Internal IT professionals may argue that outsourcing makes users hostage to the vendor; once IT functions have been turned over to an outsider, it will be extremely difficult and costly to bring them back inside again. The organisation often will give up (or fail to develop) certain technical skills, transfer hardware and software to the vendor, and may become locked in to the vendor’s proprietary technology.

    One way to mitigate the vendor’s leverage in this regard is to explicitly accommodate reversibility (or transfer to another vendor) as part of the original agreement. For example, an agreement for outsourcing network operations might specify that the vendor will dedicate certain identified personnel to work on this account exclusively, and might explicitly entitle the state to hire these people back if the agreement is cancelled. A vendor who contracts to develop and maintain applications might agree to train members of the client’s staff to do maintenance and enhancement work.

    Critics of IT outsourcing cite the high cost of safeguards that ensure reversibility of the decision as a reason to keep IT in-house miss an important point. If the function is critical, then cost savings should be a minor reason for considering outsourcing. If strategic focus, quality, and flexibility are improved, then costs need only be comparable, not lower.
 

Business Knowledge

    Successful application development hinges on good user and developer communication — a problematic issue even when the players are all members of the same organisation. Some say that outside vendors cannot be expected to do as good a job as insiders, who have become familiar with the industry as well as the idiosyncrasies of the agency.

    However, it is important to distinguish those IT activities that require "inside" knowledge. Local knowledge has little relevance to the design of a cashiering system, for example. One point-of-sale process is much like another, and, in general, idiosyncrasies should be avoided. Moreover, even where local knowledge is important, the vendor will learn the specifics of their business over the course of a long-term relationship. In fact, because agency IT professionals are often recruited from outside, rather than developed internally, and because their turnover is high, they are likely to have less tenure in the organisation than other managers. In many cases, they move readily between agencies, and frequently between industries. As a result, the insiders’ advantage may not be as great as it might seem.
 

I.T. Costs and Outsourcing

    The reason most frequently given for choosing IT outsourcing is cost reduction — outsourcing substantially reduces the agency’s direct investment in IT assets and/or the annual expense of providing IT services. Yet, except in the case of small companies that cannot operate at efficient scale themselves, direct cost savings alone seldom provide a good rationale for outsourcing IT activities. Because hardware, software, and other IT capabilities can be scaled to suit virtually any requirement, a vendor is unlikely to have significant scale or scope advantages over a large agency that is operating efficiently. For example, software packages offer an extreme example of scale economies where the fixed cost of development are spread over hundreds or even thousands of units.
 

Control without Ownership

    Arm’s-length contracting — the traditional mode of IT outsourcing — is only one option among many. The challenge for top management is to understand which (if any) IT activities should be performed internally, which should be contracted out, and which should be handled through some form of joint venture or partnership. These decisions hinge on the strategic impact of a particular activity and the extent to which the agency has significant competence to undertake the activity successfully.
 

Strategic Impact

    Strategic impact — how central the activity is to achieving the goals of the business — is the key to determining whether the long term objective should be to perform the function inside or outside the agency. In general, the agency should strive to perform in-house those activities that are important and outsource those that are peripheral.

    As conditions change and as new technologies are developed, agencies inevitably find that they are not sufficiently competent in all of the areas that are critical to success — they identify new core competencies and related functions that aren’t now available within the agency. Developing strategic competence can be accomplished in one of three ways: through independent effort, through a temporary partnership or strategic alliance with a third party, or through a joint venture.

    The ultimate goal of a temporary partnership is to bring the competence offered by the vendor in-house, by developing the skills and experience among its employees that will enable it to internalise the capability in the future. Of course, partners need to be chosen very carefully, with an eye to their business stability, mix of skills, and compatibility of interests.

    A handshake, a contract, or a formal joint venture agreement may govern such partnerships, but unlike standard contractual relationships, the agency’s goals extend beyond obtaining a satisfactory product or service. Contractual provisions cannot guarantee the success of a partnership alone; rather, the partners must have a genuine shared interest in the common undertaking. It is important that both partners explicitly recognise their motivations at the outset. Otherwise the partnership may metamorphose into a form of imprisonment in which the party with greater market power calls all the shots (e.g., a vendor who does not care about continuing the relationship may just keep raising prices).

    When both risk and strategic impact are high, the state may want tighter control than is typical of a loosely formed alliance. Some form of joint venture is a promising alternative in situations where the revenues will support a new enterprise.

    As with partnership arrangements, it will be important to think through the terms of a joint venture carefully, to ensure that both parties’ needs are met, and that neither takes advantage of the other. In particular, some thought must be given at the outset to the question of dissolving the joint venture when it has served its purpose. After all, after an agency has internalised the required skills, its motivations change. If the activity remains strategically important, the state will wish to perform it internally and will view interference from the joint venture partners as unnecessary. On the other hand, the vendor/partner will begin to view the agency’s ownership position in the joint venture as a "tax" to be eliminated.
 

Contract Competence

    A common approach is to outsource non- essential activities that the organisation is not especially good at performing. Sometimes low quality and high costs are the motivation to outsource these IT functions. However, current performance may be adequate, yet management wants to refocus on more important topics.

    IT activities in which a large outsourcing vendor can achieve significant economics of scale often fall into this category. Examples include such bread-and-butter IT capabilities as general accounting and payroll systems, data centre operations and PC acquisition and service. The technology is well understood and the application requirements are well defined. Although these capabilities are necessary to agency well being, they are rarely central to the agency’s strategy. There is little reason to design or operate your own payroll system, since a specialist can do it at least as well, and probably at lower cost. More to the point, doing it yourself represents a diversion of resources from more essential business activities.

    One objective of outsourcing is the improvement in overall quality at reduced cost, yet with minimal risk. The payroll system, for example, may not be central to the agency’s strategy, but if people do not get paid reliably, management attention will certainly be distracted. On the other hand, to spend excessively to develop internal payroll services diverts resources from the important to the trivial. The objective of this type of outsourcing, then, is to refocus attention on more important problems.