Mantra for Effective Chip Design

Date:   Monday , November 17, 2008

Fundamentally, managing technology calls for leadership qualities more than management qualities. It’s all about inspiring an already highly qualified intellectual class of people and getting them to focus on delivering results. One of the biggest secrets of managing technology is putting the round people in the round holes and the square people in the square holes. For example, at Silicon Spice we had a key chip architect who had a hard time fitting in as a team player. In fact, the situation got so bad that most of the staff was ready to quit or have him quit the company. Chip architects in the silicon industry are very independent, innovative, creative people. They are “individual contributors.” They don’t necessarily have the savvy and the cultivation of mindset to be able to work in teams. Nevertheless, they play a very critical role in delivering the solution. Therefore, you cannot afford to lose these people. To remedy the problem with this architect I had him report to me directly as an individual contributor. That kept him in the company to get the full benefit of his insight. But he interacted only indirectly with the “implementation team.” We got the full benefit of his chip architecture expertise without compromising the implementation. So, sometimes you have to accommodate tremendously. When we were doing the Pentium at Intel, the man who had worked on the Intel 486 chip had moved on to a different role. A woman who worked under him was the natural best choice to succeed him, but she wasn’t interested. She wanted to learn more about software compilers. The next best bet was Avtar Saini, who then was just a logic designer (he now heads Intel’s operations in India). I believed he had the potential to manage a large complex design, but he had no experience at leading a complete design team. What was perhaps worse was that the top management didn’t know him. If I had just put him in charge of the program — which was highly visible to the industry and to management at the time — people would have raised numerous concerns. Instead, I put him in a “two-in-a box” arrangement with a man named John Crawford, a very visible chip architect with the credibility I needed to sell this organizational change to top management. John was interested in trying his hand at management, so the combination turned out to be a win-win situation for everyone. It may be true or false, depending on the company, but many engineers associate progress, growth, compensation, and financial rewards with general management. They know if they’re GMs, they’ll have more control and they’ll be paid much more. However, at a place like Intel or even at Broadcom, individual contributors are tremendously valued. That’s their career path. A technical person can make as much money as a general manager. In short, you have to fine-tune each situation, look at the individuals involved, assess their strengths and areas of improvement. Strategies for Success
Time-to-market is becoming shorter and shorter in the chip business, but building complex custom chips takes more and more effort. It is almost like running a marathon. You have to keep in mind when you start these teams running that you don’t want them running too hard and too fast. You don’t want them to tire out too early. Specifically, you want to make sure that you set things up to resemble a relay race rather than a marathon. I have often done this by rotating key team leaders through different roles, some much less demanding then others, to avoid burnout. That requires some very careful orchestrating. Many CEOs don’t understand that in technology, the CEO’s job goes beyond selling the company outside, raising money, and working with the board of directors, marketing the product, visiting customers, generating sales and acquiring and managing partnerships, and son on. While all of this is important, the CEO has to make sure he keeps his eye on what I call “the execution ball.” Execution, or product implementation, is of paramount importance. Most companies in my opinion fail not because they lacked a great idea, but because they failed to execute it in a timely manner with the original performance goals. No matter how strong the management team, the CEO’s or general manager’s responsibility is to get personally involved in implementation detail and act like a Super Program Manager for the product. Marketing
Marketing in this industry has taken a different slant over the years. We no longer just sell chips. We also add value by selling software, even though customers may not pay a lot for that software. The value-added dimension of the chip is the software. It is essential that you identify your markets and key players early, and work with customers from day one. This is even more important in the communications industry, where either there are no standards or there are a multiplicity of standards. You have to work very closely with customers to define what the solution is going to look like. The solution is different in each case. Different customers are following different protocols and different algorithms. People are mistaken to think that selling silicon chips has any similarity to selling corn flakes. You’re selling a very sophisticated product generally to an engineering audience. They will not get turned on by a sexy or appealing company name. Customers want to know the underlying value proposition. What is your technology? Is the value proposition the density, the low power consumption or the flexibility of the architecture? What does the platform look like? How will they be able to debug it? How will they get customer support? How are you doing the testing? Customers look at many such factors. Big vs. Small
The biggest lesson I’ve learned is that the success of a big, established company brings along its own drawbacks. For example, at Intel the core team did 386, 486 and Pentium. The good news was you did not repeat the mistakes you made in the previous program. But the bad news was that we began to think linearly about what it takes to do the next one and the next one and the next one. In a startup environment, where you start from scratch, you’re not encumbered by history. You lack resources, money, time and people. So you’re really forced to think out-of-the-box to come up with the most creative solutions you can. Therefore, startups will always be around to challenge the well-established Intels of the world. The second lesson I’ve learned is that working with customers really early on is essential for success. You can no longer create technology, take it to the customers and expect them to buy it. You have to work with them, early on. You have to make sure that whatever you’re creating is a true value-add. When you’re done with that, they’ll buy it from you in large volume. Lastly, from working in large and small firms over the years I’ve learned that large companies can learn from small ones. You have to be extremely flexible, because your workforce is in great demand. You can’t treat them like gear or equipment. They have emotions and feelings. You have to personally get in touch with them. They have to know that you care about them.
The main challenges for technology management arise from the need to continuously find ways to add value for the user. Semiconductor technology has become a commodity. Technology for the sake of technology is not interesting anymore. The user has to realize a significant value in terms of performance, functionality and cost. That’s where the new generation of applications, appliances and solutions will come from.

Vinod (Vin) K. Dham is vice president and general manager of the Carrier Access Business Unit of Broadcom Corporation. He came to Broadcom from its acquisition of Silicon Spice Inc., where he was chairman, president and CEO. Recognized as the “Father of the Pentium Processor,” he held the positions of vice president of Intel’s Microprocessor Products Group and general manager of the Pentium Processor Division.