Getting outsourcing right: here are lessons from the biggest deals of what to do-and what not to do.

The gutting of the corporate back office continues at a rapid pace, as CEOs outsource large chunks of their operations. In 2002, a flurry of year-end activity--including a $4.5 billion deal between Bank of America and EDS, as well as a $5 billion agreement between JP Morgan Chase and IBM-brought the total value of large-scale outsourcing deals to $29 billion, compared with $16 billion in 2001.

But for every deal that works, it seems another comes up short, resulting in finger-pointing and bad publicity. More important, botched outsourcing deals fail to create value for shareholders.

What separates the successes from the failures? Our research and client work reveal that outsourcing deals are usually won or lost long before the vendor is chosen. Winning arrangements often depend on the ability of the client company to look beyond organizational boundaries to find combinations of related activities that can be merged into bundles, to be packaged and farmed out.

We at Marakon Associates studied 45 "mega-deals" completed between 1993 and 2002 (average size: $1.6 billion), tracking total shareholder return data and market indices for each company from two months before the announcement of the deal to two months afterward. To find the return attributable to the deal, we looked at the difference between the total shareholder returns for each company and the market index.

Our first finding was that outsourcing still pays. Over the 10-year period, companies in the sample outperformed their market indices by an average of 3.5 percent in total returns. Twenty-seven of the 45 companies showed positive returns relative to the market.

Our second finding was that selectivity matters. Nineteen of the 27 companies that created value selectively bundled their outsourced activities. Only eight of the 18 companies that declined or broke even in value had bundled parts of their operations.

Selective bundling can be the key to a successful outsourcing deal, but only if it's approached in the right way. One of the most common mistakes companies make is to farm out unrelated business activities that happen to be managed together in the organization under a common umbrella. Doing it that way appears to solve an internal headache, but outsourcing disparate activities can be like mixing apples and oranges, and can backfire. Activities are sometimes so unrelated in nature that it is difficult for a supplier to manage their costs effectively. Furthermore, if different "apples" and "oranges" in other parts of the organization stay in-house, the company may lose out on potential cost savings.

The point is that CEOs should not be limited by organizational boundaries when thinking about what and how to outsource, and should look for ways to outsource oranges with oranges. For example, the wholesale banking unit at ABN Amro saw that it had related technology service and applications development in Amsterdam and London, and that both were operating below capacity. Top management cut across organizational boundaries, bundled these activities together and, in November 2002, announced a $1.5 billion contract to turn them over to EDS. By doing this, ABN Amro created efficiencies of scale it could not have achieved otherwise.

Another financial services giant discovered similar benefits in early 2001. It analyzed cost savings from outsourcing under three different scenarios. The first was farming out large, end-to-end activities such as the management of the company's IT operations and networks; this would have yielded 15 to 25 percent savings on average. The second was outsourcing smaller, more discreet activities such as data processing or payment processing, which would have yielded from zero to 25 percent in savings. Between these two extremes, the company spotted a third path: creating bundled activities such as the provision and support of IT processing and distributed computer services, which would achieve as much as 30 to 35 percent in savings. By doing the analysis up front, the company was able to see clearly the benefits of selective bundling.

Despite potential gains, many companies have difficulty relying on an outside company to do things they have done themselves for years. They may even maintain a long list of "untouchable" activities that will never be farmed out, because doing so might threaten the power base of key managers. To deal with this inertia, we suggest that CEOs start with a blank slate and assess each and every activity their companies perform. They should begin by assuming that the company will provide no internal support, and then ask which groups of related activities the company will be able to perform at a high level of quality and in a cost-effective manner. American Express did that in February 2002 when it signed a $4 billion service contract with IBM to host its Web site, network servers, data storage and help-desk support. This agreement was the culmination of a 14-month analysis, during which top management thought hard about giving up control of part of their IT operations in exchange for staying competitive on cost an d quality. It worked. As a result of this focus on fewer, better services, American Express outperformed the market.

After deciding which bundles of services should be outsourced, the next step is to allow the outsourcing vendor to make a profit. As simple as this sounds, many well-intentioned companies play potential suppliers off each other to gain concessions. Then they draw up contracts that demand strict service-quality agreements and operating procedures. But this approach can backfire as the company's needs change and cannot be met under the terms of the contract. Instead of taking an adversarial approach, the company can help its partners do what they do best. This means creating a "win-win" agenda.

In short, it's worth challenging the conventional wisdom that outsourcing is a good thing in and of itself. Outsourcing can deliver big gains in shareholder returns if done for the right reasons and with discipline. But it can also destroy significant value. To improve the odds 0f success, CEOs should take advantage of lessons revealed by the best and biggest deals to date.

RELATED ARTICLE: Key Questions to Consider:

* What is the right group of activities to bundle together?

* Can these bundles be independently managed or sourced?

* What should we keep in house and do ourselves?

* Which activity bundles offer rewards that exceed other investment opportunities in the company?

* How can we help our partners win?

* Are we writing a contract that enables flexibility and accountability?

* Do we have an agreed way of resolving problems?

* How can we manage the transition in a way that ensures success?

* Do we have a clear "benefits recapture plan"?

Craig Albright is a senior manager in the New York office of Marakon Associates, a management consulting firm.