In recent years many high-tech companies have shifted chip design, software development and medical research to countries like India and China, which have a large number of highly educated engineers — many more than are available in the U.S. However, before these newly formed RDCs can be fully functional — and before any hypothetical cost savings can be realized — we need to learn from the initial mistakes of the industry. For example, preserving the corporate culture under several layers of cultural diversity is a big challenge. An even bigger challenge is bringing people together who may have radically different goals while creating an effective communication process that avoids isolationism. The problem is exasperated when teams are separated by both time and space.

Let's take an example.

A US-based fabless ASIC company, with most of its customers in US, started a design center in India to take advantage of a large pool of cost-competitive, skilled engineers. In the ASIC business, there are three ways to grow — build more design teams to service more customers/projects, ship more parts, or escalate price. The company chose the first option and mirrored their customer design team in India, much in the same way it would have added a team at its headquarters.

The RDC failed in its first project since it operated as though all stakeholders in the project were at a single geographical location.

The difference in time and space created a time-constrained, ritualistic communication between cross-functional teams and customers at odd hours of the day leading to low-quality and ineffective technical communication. In the process, customer interface responsibilities shifted from marketing to engineering. The ultimate result: stressed employees, delayed projects, management involvement in customer relationships, and finally loss of customers. The mistake the company made was in creating a self-synchronous, mirrored project team for a real revenue-generating project without going through a training cycle for internal and customer communication and the design process.

In another example, an organization retained the responsibility of creating the product at its headquarters but distributed parts of its project among three geographically different design centers. For example, one of the two remote design centers was responsible for designing high-speed I/O cells while the other remote design center was responsible for test benches and verification for all the logic blocks designed at the headquarter design center. Each RDC was a full-scale organization with similar hierarchical structure to the parent. This meant that each RDC had a design center manager, project leader, technical project lead, and design team. In this situation the business unit manager had to deal with three design center managers, three project leads, and three technical leads, thus making the project team top heavy with individual goals and priorities. This organization model created islands of excellent technical capabilities. However, the model did not support quick time-to-market.

In yet another case the organization started with a modest beginning of a Cube extension model. Rather than focusing on creating a large organization, it just hired a few engineers as an extension of the current design team at a remote location without creating a blue print for its future growth. This opportunistic strategy resulted in a workforce completely unaware of the reason for their existence and the meaning of their work in the big picture. More practically, it became a training ground to benefit the hiring for other companies.

The right way
Most of the situations described above occurred in the last two to three years. What went wrong? Most organizations make the mistake of focusing only on cost cutting and not on creating an extension of their existing organization. So what is the right strategy for creating an effective RDC with a killer instinct? The answer lies in creating a RDC by following a classical product introduction model in marketing — Introduction, Ramping, Consolidation, and Maturity, and creating a work force of global leaders of Blood Group U+.

People of Blood Group U+ are people who have a high degree of ownership (ownership is different from leadership) and have a sense of Urgency when getting things done to meet time to market with an additional degree of resiliency. They are people capable of managing a project working through cross-functional teams and are managers with ownership qualities because ownership is the preferred form of consolidation. True global managers are managers who have moved beyond borders and are willing to move or travel continuously across boarders, have experienced working with multiple cultures in different countries, and understand the needs of people with different ethnic backgrounds under the bigger picture of corporate framework.

There could be several co-ordinates to measure the position of an organization relative to any of the four stages (Introduction, Ramping, Consolidation, and Maturity). The two most important scales to define are the degree of modularity of tasks/design (DoM) and the degree of organizational hierarchy (DoH). DoM is the measurement of the modularity of any project or task from being a single piece / event to highly modular. DoH is the measurement of being a flat organization to layers of hierarchy. The best fit between these two variables defines the stage of the organization out of these four phases (see figure). The relative time scale spent in each of these four stages completely depends upon the size of an individual organizations' business.

In the initial stage of starting an RDC, the DoH must be kept low since the DoM may be very low. DoM is low for both psychological and technical reasons. Project managers are in the middle of a project and it may not be technically possible for them to divide the project into smaller pieces which could be implemented at RDC. In addition, the decision of starting an RDC is most often perceived by the middle managers as a management opportunity and threat to themselves, and therefore they may not be willing to break the project into smaller modules.

In the absence of enough direct data points, a controlled experiment — by hiring a critical mass — is always a good start to understand the cultural issues, hidden costs and true economics of operations, and stress-test the middle management who will be dealing with new cultures and working with another time zone. During this phase it is always a good idea to expect results from the RDC on "best effort basis" rather than a guaranteed service by completely transferring the responsibilities. In the semiconductor business for example, porting an existing standard cell library to a new process, or customer requested simulation of test benches with the existing RTL, could be good start at the RDC during this phase.

As the system becomes more stable in terms of the output being within the acceptable norms of quality, it is the time to tweak the two "knobs" of DoM and DoH. To move into the second stage, Ramping, you may want to keep the DoH the same and increase the DoM. During this time project managers can trust the remote team to understand project specs and their implementation. Caution must be used in transferring responsibility until schedule and quality can be enforced. The ramping stage does not necessarily mean hiring. It should be seen as ramping of operations, which may also be achieved by leveraging the resources available from other local companies.

Meeting the schedule within acceptable standard deviation is the sign of maturity in the Ramping phase. This is the time when the DoH can be varied and responsibilities can be carefully implanted to enter the phase of consolidation. Full chip design such as re-using the proven blocks from previous generation designs, board re-layouts to change the form factor, or writing device drivers can be completely transferred to the RDC with a clean transfer of responsibility. Most managers feel that the first full-scale project a RDC does during the transition phase from Ramping to Consolidation decides the success and acceptability of the RDC. Therefore, before reaching this phase it is important to evaluate the best fit between the DoM and DoH.

It is natural that as we mature, we demand more independence. Soon, these little satellite organizations will demand more participation in the business process, sharing of power, proximity and involvement with customers, and so on. The initial reason for their creation — being cost competitive — may not remain the effective reason for their existence.

At the macro level, globalization improves the quality of life in cost-competitive countries and therefore creates more demand for emerging products and technology. At the micro level, RDCs enable the creation of centers of excellence. In the future, these organizations will be mature enough to move to phase four: Maturity. They should be now be empowered to become independent product lines to meet either local or global needs.

Gopal K. Garg is Senior Director, Technology at Cypress (San Jose, Calif.)

See related chart