You’ve heard the tales of outsourcing gloom and doom and read about the staggering percentages of outsourcing failure. Now consider these three CIOs’ experiences:
- Five years ago, a business unit at energy giant Cinergy Corp. outsourced database administration services, with no plan to extend the contract to any other part of the business. But when Cinergy centralized IT two years later, CIO Bennett Gaines called on the outsourcer to provide database administration services enterprisewide. Since then, the outsourcer has proved instrumental in a major technology shift—from data marts to an enterprise data warehouse.
- Four years ago, Summit Information Systems, a software developer for credit unions, outsourced disaster recovery services for its data center, located in central Florida. In 2004, as Florida faced the worst hurricane season in its history, “[the outsourcing vendor] was willing to do whatever it took to keep our systems up and running,” says Steve Steinbach, Summit’s vice president of data center operations.
- Three years ago, JM Family Enterprises outsourced all mainframe hardware, software and operations because mainframe usage at the $8.2 billion automotive holding company had leveled off. The outsourcing vendor immediately optimized operations so that critical month-end financial reports landed on the desks of JM Family executives first thing in the morning rather than late in the afternoon, as was the norm. “It was the same hardware. The same data. But they were able to gain efficiencies because they knew how to run a mainframe better than we were ever able to,” says Senior VP and CIO Ken Yerves.
What’s this? IT executives who are satisfied with their outsourcing arrangements—even praising their vendors? This might seem strange, as mass indictments of outsourcing have led to misperceptions. In fact, some slices of outsourcing are almost always successful.
Cinergy, Summit and JM Family achieved success by outsourcing well-defined processes that had clear business rules. Jeanne W. Ross, principal research scientist at MIT’s Center for Information Systems Research (CISR), calls such outsourcing arrangements “transaction relationships.” These are the most straightforward of outsourcing deals. The work is relatively easily defined, and the CIO wants to farm it out for clear reasons: to gain access to specific technology expertise, to deal with variable demand for certain IT services, or to free up internal staff for higher-value work. And in these relationships, vendor and customer needs are usually aligned; what the two parties want coincides more often than not.
In a recent study by CISR and CIO magazine of 90 outsourcing deals at 84 companies, CIOs showed much greater satisfaction with transaction relationships than with any other type of outsourcing. (See “Sustainable Value from Outsourcing,” this page, for a summary of the research findings and the three types of outsourcing arrangements identified.) Nine out of 10 IT executives in the study reported success with their transaction outsourcing. And CIOs who develop satisfactory transaction relationships can reap more rewards than simply saving money or freeing up staff. “Companies excelling at transaction relationships achieve agility because managers focus on processes that distinguish the company, rather than the silly things they must get right but don’t want to bother with,” Ross says.
That’s not to say all these stories have happy endings. After all, there is the 10 percent that doesn’t work out. And even successful transaction relationships can create significant problems if you’re not careful, such as application silos and poorly structured enterprise architectures. But CIOs who engage in well-defined and well-managed transaction relationships with outsourcers will, in all probability, gain the sustainable value they’re seeking.
How to decide what goes
Choosing which work to outsource and which vendor to send it to are the first decisions any CIO makes in a transaction relationship. But these first steps can be the most critical. The key to success with this kind of outsourcing, says Ross, is making sure that what you send to the vendor is something “extractable”—that is, easily definable, removable or considered noncore. Tasks such as desktop provisioning, business continuity and mainframe processes tend to fall into this category.
In some cases, the categorization of a slice of IT work as extractable is straightforward. Guy de Poerck, CIO of the International Finance Corp. (IFC), an arm of the World Bank, has been outsourcing help desk services to Affiliated Computer Services (ACS) for six years and hasn’t had to give it a second thought. “With help desk, the transaction is pretty simple,” says de Poerck. “I know it’s successful because I hear so little about it. I’ve never been confronted with a problem.”
But what is easily definable, removable or considered noncore can vary by company. JM Family has a well-defined annual process for categorizing its technology. The process helps it to prioritize IT spending and decide what should be done in-house versus what might be a candidate for outsourcing. Every August, CIO Yerves divides all technology into four categories: emerging, mainstream, contained and retirement. Emerging and mainstream technologies are where the IT department focuses its budget, training and staff. Contained technologies are systems—often legacy systems—that are not growing in usage but that must be kept up and running and are therefore candidates for outsourcing (although internal improvements in efficiency might suffice as well). And retirement systems are, well, on their way out.
Going through this process in August 2002, Yerves and his leadership team categorized mainframe operations in the contained category for the first time. One of the company’s three business units had plans to wean itself from the mainframe. The others continued to employ the services but with no increase in usage. In addition, managing mainframe operations was a fixed cost in the IT budget, even though the company’s usage was variable.
So Yerves began to examine potential outsourcing vendors in early 2003, narrowing the field to two contenders by April. He put together a valuation matrix that explored what the vendors could offer in variable capacity and cost. Variability was important because of the peaks and valleys in the business’s mainframe processing needs. Ultimately Yerves went with IBM.
But settling on the right outsourcer wasn’t the trickiest part for JM Family. As evidenced in the CISR-CIO study, protracted contract negotiations can ruin a transaction outsourcing arrangement. “The concept of a transaction relationship is, in some respects, an oxymoron. Ideally, there is only barely a relationship—much like we’d have at the grocery store with the sales clerk. I find what I want, the clerk rings it up, I pay,” Ross says. “If client and vendor engage in protracted negotiations about unique features or special pricing arrangements, they no longer have a simple transaction.”
And Yerves had heard horror stories from peers and analysts about disputes over outsourcing contracts. So he sought to circumvent contractual conflicts by streamlining the process. Yerves gathered leaders and lawyers from both the vendor and JM Family and sequestered them in a local hotel. “We gave them 72 hours and said if we can’t get to an agreement by then, we don’t have a deal,” he says. And it worked. Before the clock ran out, JM Family and IBM had inked a deal. “We were able to mitigate that early risk” of long or contentious negotiations, he says.
In fact, a mark of good transaction relationships is that contract negotiations are straightforward. The services to be outsourced are extractable, and companies and vendors have similar goals for the arrangement: Make it fast, easy and cheap. “It’s a lot easier to get to success with transactional-type outsourcing [than with other types] if the customers have well-defined requirements,” says Mary Lacity, professor of information systems at the University of Missouri. “If the customers know exactly what they want, they can tell the vendor, and the vendor can price [the contract] correctly. The expectations [for the outsourcing arrangement] can then be well defined in the contract.”
The suggestions for improvements can sometimes come from the vendors themselves. Of course, the outsourcer has its own motives, such as cost efficiencies, for making such advances, but given the nature of transaction relationships, these can work out well for the customer too. For example, under its original contract with International Finance Corp., ACS had been providing onsite help desk service during work hours in IFC’s Washington, D.C., office. At night, a single ACS employee took a pager home to handle off-hours problems. It worked out OK, and there were no major complaints. “But frankly,” says de Poerck, “the service was dependent on the ability and commitment of that one person.”
Nine months ago, ACS approached de Poerck with a proposal: ACS could now offer 24/7 help desk support to IFC for the same price. The reason? ACS was now able to transfer off-hours calls to its Bangalore office. De Poerck had significant experience with offshoring application development and maintenance to another outsourcer in Chennai, and was game. After a pilot, he signed up.
De Poerck says that the impetus for the change was the looming end of the multiyear ACS contract, which the World Bank requires IFC to put up again for bid at the end of a contract period. If ACS could offer IFC extended help desk service for the same price, perhaps it could outbid all comers. Regardless, “now that we have that team in Bangalore, there’s a level of performance and processes that weren’t in place before,” de Poerck says. “We’re getting a lot better service.”
Clean House Before Outsourcing
If the success rates and benefits of transaction relationships have you wondering why you don’t just outsource everything that you can characterize as extractable, beware. There is a potential trip-up. Emphasizing transaction relationships can make a company less innovative with its IT architecture. In the short term, transaction outsourcing can help a company clean up isolated processes, but in the long term these deals may actually reinforce application silos in a company, according to the CISR-CIO research.
This potential problem makes the need for enterprise architecture planning even more important for companies engaging in transaction relationships, Ross says. “We think of a transaction as a smallish piece of IT that the vendor ought to be able to provide value on,” she says. “But that can be dangerous. Until companies have cleaned up internal processes, outsourcing transactions is going to reinforce silos. And that can make your internal architecture messier and messier.” Over time, a weak enterprise architecture could inhibit a company’s ability to respond to changing market conditions.
Cinergy could have fallen into that trap. Before CIO Gaines’s arrival in 2002, the company had commenced a transaction relationship with DBA Direct to do database administration for its commercial energy business unit only. At the time, Cinergy IT was decentralized, with each of the company’s four business units operating autonomously. The DBA Direct deal “worked reasonably well and served its purpose,” Gaines says. “But there was not a lot of visibility into it. It was filling a rather specific need in that business unit, and it wasn’t scalable.”
When Gaines took the reins, he was charged with centralizing IT for the $4.7 billion company. One of his first acts was a review of Cinergy’s entire IT service delivery model. He also looked ahead to future needs. “We had an opportunity to get into some new emerging technologies, but that would require incorporating an enterprisewide data warehouse,” he says. Gaines knew he did not have enough internal expertise to tackle that job.
Gaines began to look at outsourcers. He liked that DBA Direct had experience with his company. “The database work they would be handling was complex, and they were already managing a highly reliable database for us,” he says. So he hired the vendor for database administration enterprisewide. DBA Direct started with the company’s Oracle databases. And as Gaines continued to centralize IT, the outsourcer took responsibility for most of Cinergy’s databases.
By reexamining the outsourcing relationship with DBA Direct and expanding it as part of a larger enterprise architecture strategy, Gaines avoided the silos that transaction outsourcing sometimes produces. The vendor has also proved valuable working directly with Cinergy’s other outsourcers, with whom Cinergy cosources much development and maintenance of applications that in turn rely on database functioning and availability.
But most impressive to Gaines has been how quickly DBA Direct, with its well-honed database experience and best practices, has been able to move Cinergy from its old data mart model to an enterprise data warehouse. Far from stagnating Cinergy’s architecture, transaction outsourcing has improved it. “I would never have been able to bring that [enterprise data warehouse] into this company as quickly if we tried to do it ourselves,” says Gaines. DBA Direct “had that core competence because it’s their business. It wasn’t rocket science. But the need was so urgent for us—we had customers in the business who were building new applications dependent on having a data warehouse—that finding a partner like them was important.”
CIOs should not be lulled into believing that success with transaction relationships means they’re ready for riskier adventures in outsourcing. According to Ross, knowledge from transaction outsourcing does not necessarily transfer to other, more complex types of outsourcing. “You’ll never find the level of success you have with transaction relationships in other types of outsourcing because nothing else is as extractable,” she says.
Smart CIOs who have found success in transaction relationships know this, and consequently, they approach all outsourcing—even further transaction relationships—with due caution. “With transaction relationships, you tend to find a lot of happy people. They’re outsourcing smaller pieces of IT,” says Ross. “But the bigger it gets, the riskier it gets.”
