Establishing a dedicated offshore development center (ODC) is the latest trend in Indian outsourcing—witness United Technologies subsidiary Otis Elevator, which created one with Bangalore, India-based Wipro Technologies. After two years of sending software work to India on a project-by-project basis, the Farmington, Conn., company had accumulated a portfolio of 30 applications, all requiring ongoing maintenance work. Each time work needed to be done, Wipro had to assemble a new team. A month after the project was complete, the team would disassemble, taking with them their knowledge and expertise. Today, Otis instead has a permanent 20-person Wipro team devoted to its application and development needs.

Although the capability to set up such ODCs for clients has been available in India for years, this delivery model is just now catching on, as CIOs’ comfort levels with sending software work to India have increased. "This has been one of the hottest growth areas [in Indian outsourcing]," explains Marty McCaffrey, executive director of Software Outsourcing Research. "Besides the potential cost savings and other benefits, a major reason this is increasing is because clients are making larger offshore commitments. More and more deals are at 100, 200, 300 [full-time equivalents offshore] right from the start and are a result of the maturing of the industry."

The benefits of a dedicated offshore center are clear: sustained knowledge retention, better rate negotiation, speedier ramp-up times for new projects and increased productivity. Additionally, foreign workers even halfway around the world can feel more connected to the client company, as they are often located in a private building or floor with the client company’s logo adorning the wall. The feeling of being part of the customer’s company can often translate into reduced turnover and better quality.

But a dedicated center is not the best choice for every situation. Setting up an ODC at an Indian vendor comes with its own unique contractual requirements. For instance, in such an arrangement, most vendors require customers to commit to a certain number of man-years of work—in Otis’s case, 20 man-years. The commitment makes the deal worthwhile to Indian companies, which have to set up the labor and infrastructure requirements necessary for an operational ODC. If you cannot adequately forecast your minimum annual manpower requirements and clearly identify the work that has to be done, or if manpower loading will be inconsistent or hard to project, an ODC is a bad idea. And you’ll pay a price. "If you overestimate the amount of work and don’t end up utilizing all that manpower, you’re going to incur certain costs," McCaffrey says. "The vendor may work with you, but there will wind up being delays as they have to shift people around."