Computing Applications Technology strategy and management

Outsourcing Versus Shared Services

Choosing between outsourcing and shared services has significant implications for long-term corporate strategy.
  1. Introduction
  2. Outsourcing as Corporate Restructuring
  3. Outsourcing as Relational Contracting: Trust and Incentive
  4. Conclusion
  5. References
  6. Author
  7. Footnotes
pipes and pointing fingers illustration

Two diametrically opposed perspectives continue to coexist in IT and other business service functions. One camp argues in favor of shared services, wherein the IT organization becomes the internal service provider to the rest of the company. The other camp promotes outsourcing: the delivery of IT services all done under one roof but with that roof located somewhere other than at the company. Proponents do not agree on which is better. This column examines the background that led to the adoption of these practices and the reasons for this disagreement.

Most readers know about IT outsourcing, and many readers are aware of H. Ross Perot, the Texan who left IBM to found Electronic Data Systems (EDS) in 1962. His idea was to offer technology as a service by not merely equipping, but also operating, customer data centers. Following his success, others joined to extend outsourcing from IT services to a wide array of business services in finance and accounting, human resources, procurement, and logistics. Technology—a combination of computers, software, and networks—underpins these support operations.

The 1990s saw the birth of mega-deals, such as the seven-year $600 million human resource outsourcing (HRO) deal between the oil giant BP and a start-up provider, Exult.1,5 But is it really only third-party providers that can offer superior service at lower cost? Is it not possible for a company to perform equally well in its internal shared services by reducing duplication of processes and facilities and by reorganizing and sharing assets? I argue that a company can achieve similarly good performance levels in its internal shared services operation, under certain circumstances.

Companies outsource for several reasons. Cost saving may be the ultimate reason. But the means by which this is achieved vary, from introducing new technology, improving service quality, transforming fixed investments into variable costs, to freeing management time to focus on core competencies. Outsourcing is about the make-or-buy decision, and the term applies to a broad range of procurement activities in manufacturing (for example, automobile companies purchasing transmission components) or services (for example, a retailer sourcing TV advertising). But when applied to business (including IT) services, two essential features are highlighted.

Back to Top

Outsourcing as Corporate Restructuring

One feature is that outsourcing of business services combines two decisions. One is the make-or-buy decision concerning the corporate boundary; the other is the restructuring of the internal corporate hierarchy. Corporations’ organization structure can be a source of competitive advantage. As recounted by the business historian Alfred Chandler, modern corporations have been restructuring constantly to align structure to strategy.3 In reality, it is easier said than done for a global corporation to design and implement an appropriate multidimensional matrix structure to meet the competing demands of different products, corporate functions, customers, and countries.4 The creation of shared services and outsourcing are both part of this search for an appropriate organizational design, giving primacy to corporate functions over other considerations, often accompanying mergers and acquisitions (M&A). M&A create duplicated functions previously belonging to two separate corporate entities. Attempts at eliminating the waste of duplication trigger the creation of shared services. More likely than not, such streamlining requires some central direction from the corporate headquarters. Without such centralized control, the intended standardization and efficiency gains may not be realized.

Existing corporate structure affects the firm’s choice between outsourcing and shared services.

The contrasting experiences at Procter & Gamble (P&G) and Unilever illustrate this point.a Under A.G. Lafley’s leadership, P&G created an internal global shared services unit in 1999 as part of the ‘Organization 2005’ restructuring initiative. It gave itself five years to pull all essential corporate functions—finance and accounting, human resources, and later IT—away from regional and divisional companies into a single Global Business Services (GBS) operation. A statement by the GBS head, Mike Power, that "everyone has done something, but no one has done everything" illustrates the radical nature of this consolidation. Central direction from P&G headquarters in Cincinnati was essential in deploying SAP-based ERP systems throughout the company before such reorganization took place. By the time P&G’s shared services were outsourced, their operations were drastically transformed and streamlined. This experience led Filippo Passerini, Chief Information Officer, to remark: "we believe that there is a sequence in this process. You outsource only when you are internally optimized."

In short, centralized firms are better placed to move quickly to efficient shared services. Conversely, divisional autonomy is likely to get in the way of implementing standardized processes. So, what if a corporate headquarters does not have central control? This is where outsourcing comes in as a first port of call. Unilever, the Anglo-Dutch firm, has lived with characterization as a loose federation of national companies with strong country managers who had little interest in global shared services. The human resource (HR) function did, however, consider outsourcing at the global level, and regarded the fragmentation of IT infrastructure as a hindrance in implementing it. With such cultural and technical barriers to creating in-house shared services, global HR outsourcing to Accenture was used as a trigger to transform HR processes, in a "throw it over the fence" or "lift and shift" approach, with an expectation of rapid cost reductions through scale economies, labor arbitrage, and increasing return on assets. Unilever could not have transformed without outsourcing, and outsourcing was an integral concomitant of transforming the organization. Outsourcing is indeed a corporate turnaround trigger. In such applications of outsourcing, both the risks and potential rewards are high.

Thus, existing corporate structure affects the firm’s choice between outsourcing and shared services. Moreover, the creation of internal shared services first before outsourcing leads to greater retention of capabilities in-house; by contrast, a path to outsourcing, without an interim step of internal shared services, engenders greater reliance on suppliers’ capabilities. In fact, the first path is about "selling" shared services assets and capabilities to providers, while the second path is about "buying in" such capabilities from providers. Some argue that in immature markets without competent providers, firms have no choice but to create internal shared services. There may well be a timing effect, with pioneers opting for the introduction of shared services and followers opting for outsourcing. But can shared services be an end point, without proceeding to outsourcing?

Back to Top

Outsourcing as Relational Contracting: Trust and Incentive

The answer to the question of shared services as an end point without proceeding to outsourcing depends on the second key feature of business service outsourcing, and that is the nature of "relational contracting." For example, in an M&A deal the signing of the contract closes the deal. By contrast, in outsourcing, such "closing" is just the start of a long-term collaborative relationship between two firms. In order for such a relationship to operate well, it relies on trust and incentives. In fact, incentive and trust may be structured better in outsourcing in some cases, but in internal shared services in other cases.

Organizational economists define a relational contract as a contract that is incomplete (due to difficulty of full specification) and informal (due to difficulty in third-party enforcement, for example in courts).2,6 A multiyear business service outsourcing deal is a perfect example of a relational contract. It is typically incomplete due to two reasons. First, future contingencies are difficult to specify in the face of unknown market conditions and new technology in several years’ time. Second, quality of business services to be delivered is sometimes difficult to describe and verify. Indeed, both parties may wish to retain a certain degree of post-contractual flexibility. If a contract is incomplete, then it is also difficult for a third party to enforce it. Parties to the contract must rely on alternative enforcement mechanisms. One of these mechanisms is to rely on trust as a social norm to work things out through discussion, with social sanctions in the event of untrust-worthy or unethical behavior.b This works well in stable business communities, and where parties are chosen for their "cultural fit."

Another enforcement mechanism is incentive alignment. Service Level Agreements (SLAs) are employed to secure high performance in outsourcing and shared services. With stable processes, performance is easily measured, and the bonus and penalty regime gives a good incentive for providers to perform well. But the credibility of the client firm to commit to paying a bonus is different for external and internal providers. Outsourcing faces "high-powered" incentives, with the client able to credibly threaten to terminate the contract when the provider underperforms; with an internal SLA, it is not easy to do anything if the internal shared services center does not perform. And what’s worse, the internal operation is often a cost center rather than a profit center. At the same time, whenever a provider is offering standardized processes that could be delivered to more than one client, SLA acts as a powerful incentive to perform well for a specific client offering the bonus. By contrast, with processes that are customized for a specific client, SLA does not create as powerful an incentive.

There may well be a timing effect, with pioneers opting for the introduction of shared services and followers opting for outsourcing.

To summarize, the following is the implication from a perspective based purely on incentives. Outsourcing works best to make an external provider truly accountable for performance, whenever processes are standardized and stable for easy SLA specification. By contrast, an internal shared service is a better option in cases where processes are either customized or being transformed. The incentive-based argument highlights the fact that parties must rely more on other mechanisms such as trust if outsourcing is applied to processes requiring customization or transformation. Thus, an optimal degree of contractual incompleteness—somewhere between a "blank check" and a "nail it all down" level of detail—depends on the task at hand (service delivery versus transformation) as well as the availability of incentive alignment mechanisms and trust.

Back to Top


The jury is still out on whether or not outsourcing or shared services is ultimately the best service delivery model. I have argued in this column that, amid all the management fads and fashion, there is more than one way to do things, and that each way has its merits and demerits, with associated risks and rewards.

Outsourcing and shared services are both part of organization redesign to give primacy to the efficiency of corporate functions. Compared to shared services, outsourcing is a combination of decisions about the firm’s external boundary and its internal structure. Outsourcing may take place at different stages in corporate activities, either as an initial trigger to bring about fundamental restructuring in a "big bang" mode or a next step after a period of internal process transformation. Relational contracts, if well designed, can service the maintenance of high-powered incentives to ensure the delivery of high-quality service. However, the firm may wish to retain internal shared services without outsourcing if it anticipates instituting further business changes in structure and scope of business services. The choice between outsourcing versus shared services is not simply a matter of timing (in mature versus immature markets). It is more crucially a matter of long-term corporate strategy.

Back to Top

Back to Top

Back to Top

    1. Adler, P.S. Making the HR outsourcing decision. MIT Sloan Management Review (2003), 53–60.

    2. Baker, G.R., Gibbons, R., and Murphy, K.J. Relational contracts and the theory of the firm. Quarterly Journal of Economics 117, 1 (Jan. 2002), 39–84.

    3. Chandler, A. Strategy and Structure. MIT Press, Cambridge, MA, 1962.

    4. Galbraith, J.R. Designing Matrix Organizations that Actually Work: How IBM, Procter & Gamble, and Others Design for Success, Jossey-Bass, San Francisco, CA, 2009.

    5. Lawler III, E.E. et al. Human Resources Business Process Outsourcing: Transforming How HR Gets Its Work Done. Jossey-Bass, San Francisco, CA, 2004.

    6. Macneil, I.R. Contracts: Adjustments of long-term economic relations under classical, neoclassical, and relational contract law. Northwestern University Law Review, 74 (1978), 854–906.

    a. See detailed comparisons in H. Gospel and M. Sako, "The Unbundling of Corporate Functions: The Evolution of Shared Services and Outsourcing in Human Resource Management." Industrial and Corporate Change, (Mar. 2010), 1–30.

    b. There is a large body of work on the topic of trust in business relations.

    DOI: http://doi.acm.org/10.1145/1785414.1785427

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More