Texas Instruments Inc. CEO Richard "Rich" Templeton runs a tight ship and the proof is in the company's second-quarter results. Top-line revenue grew 9% sequentially and TI's order backlog jumped $361 million to $3.39 billion. Dallas-based TI's gross operating margins shot up 2.1% to 47%. Meanwhile, SG&A as a percentage of revenue dropped a half a percent to 11.1%. An electrical engineer, Templeton has risen through the ranks in his 25 years with TI and was named CEO and president on May 1, 2004. His messages stress accountability, innovation, market share and—like many of his professional peers— intense focus on customers. EDN editor-in-chief John Dodge interviewed him.

Q: When you say you put customers at the center of everything you do, can you give me some specifics on how you interact with customers that makes TI different or unique?

Templeton: Success can be measured very simply. And that is: Is your market share growing on an annual basis? That's how customers vote. There's no magic and no shortcut to how we do this. It's about spending time and [that's] everybody up and down the management team. While it's important to have your own roadmaps and to know how you want to evolve the technology, we put a lot of pressure on our folks to ask very obvious questions. Where is the customer input on these roadmaps and how tightly have we coupled them?

Q: What do you personally do to make sure this deep customer interaction happens?

Templeton: The two most visible things is spending time out in the field with customers. Secondarily, we have a culture where you keep track of everybody relative to their market share. That can range from sitting down with the European team and understanding how they're doing on an account-by-account basis or how are they doing in the region. That is the ultimate test of our customers voting for you or your competition.

Q: What percentage of your time do you spend with customers?

Templeton: It's probably well over 40%.

Q: How do you reconcile pushing people to innovate and try new things against strict accountability?

Templeton: I don't think they're different at all. Anybody that's been in the semiconductor industry for any amount of time knows that when [you] innovate well, you win.

Q: You've talked before about three billion new consumers from developing nations. Please elaborate.

Templeton: The PC era involved two billion people from primarily the U.S. or North America, Western Europe and selected economies in Asia. If you think about this era of communications and entertainment, you have three billion additional consumers showing up between India, China, Russia and South America. Those three billion people really were not that influential in the PC era. And these three billion people are ready, willing and anxious to make phone calls, watch television, listen to music and be entertained.

Q: You've also talked about 10 products for every one of those consumers. What do you mean?

Templeton: There was group of a hundred people for one minicomputer. The PC was one person, one computer. If you start adding up people's personal electronics now, it's a cell phone, music player, digital camera, wireless access point and broadband connection. You just start counting up the number of electronic products per person and we could be in a world of ten-to-one.

Q: How do TI's customers break down? Is it 20% of customers constitute 80% of your business? Or do you have thousands of similarly sized customers?

Templeton: I suspect the results in any one year probably have an 80-20 look, but over even a reasonably short period of time, you can have a lot of change in who those top customers are. That's where diversity of customer base and markets really carries us. While Nokia is a great customer, we've got great relationships with Motorola. We've got growing relationships with Samsung. We've got a chip set business that lets us engage with customers throughout Asia. We have a growing business in Japan with [NTT] DoCoMo. When one of those is down, another one tends to be up. When you look at our analog business, you literally find us with 50,000 or 60,000 customers around the globe.

Q: What will be the fastest growing markets in 2006 and what might be slowing down?

Templeton: The way we run the place is not trying to pick what the market is going to do in '06 or in '05. We ask if these spaces have good growth for the next three to five years.

Look at the wireless world. The handset is probably the most important electronic product in the industry. There's a lot of growth in front of it with new 3G and 4G standards with more content going into every handset with three billion new people coming into the consuming population on a global basis. Broadband is still very early in its existence as we put high-performance information into homes throughout the world. Once homes are wired for broadband, the [number] of electronic products that can go into that home vastly increases.

Q: Haven't you said broadband is under-performing from a margin perspective?

Templeton: It is, but it is a strongly performing market from a revenue perspective. Another market is HD TV. When people see it, they don't ever want to go back. The U.S. is out ahead on HD deployment compared to other countries. Over a 10-year period, HD is just going to get larger. We're doing a lot of things to make sure that the cost effectiveness of HD and large-screen HD TV can keep coming down to where it's even more available to the population.

Q: Do you see the functionality in these finished products coming together into integrated packages? Or do you still see more separate products, single-function products?

Templeton: I am not a big convergence fan. I don't believe that there is a single Swiss Army knife that's better than the individual products. Take a look at a cell phone. A lot of people didn't believe cameras on cell phones were going to be anything other than a novelty, but today it's a required accessory. In the future, cell phones will have full camcorders. And you'll probably see cell phones be able to play music; TV or games. That said, people that take a lot of pictures will not have the cell phone as their primary digital camera. My belief is that individual devices will remain healthy and strong.

Q: For all its innovation, why do you think the semiconductor industry is perceived as so mature, not growing comparable to the markets for the finished products that contain electronics?

Templeton: When the market is down, and everybody says the sun will never come up again. Then it swings to the other extreme when it's in good condition.

Q: Do you view regions like China as a threat or an opportunity?

Templeton: I spend very little time on China as a threat. They're a consuming market that will buy cell phones, TVs and players. And as a result, they're buying electronics. Will we see Chinese semiconductor companies emerge over time? It's still to be proven. We've seen more dedicated or targeted companies come out of Taiwan.

Q: With respect to manufacturing, how does your build internally versus outsourcing ratio break down?

Templeton: Around 20% of the total volume is built on the outside. As a percent, it's reasonably small.

Q: You focus a lot on inventories in your earnings statements. How do you manage those?

Templeton: It's back to a very simple belief that the most important thing is to be responsive to customers. While that doesn't always go over well with the financial community, it does go over well with customers. That's the most important way to manage inventory.

Q: How do you strike a balance?

Templeton: If you are in long-life products like in our analog market or our catalog DSP markets, the tradeoff is very clear—more inventory is better because those products will be saleable. It's only a question of when.

With areas that are more customized by socket or customer, you've got to use good judgment in making sure you don't build products that are not going to be saleable over time. That ends up being the greater tradeoff. But the dominant theme is to make sure you can be supportive of the customers.

Q: Could the Dell model of carrying days or just hours of inventory such a model ever work in the semiconductor business?

Templeton: No. Semiconductor manufacturing, which depends on the sophistication and complexity of the process, package and test, lasts anywhere from eight to 12 to 16 weeks of manufacturing. With that type of manufacturing cycle, you will not be at days or hours of inventory. Our job is to work with these supply chains, including customers like Dell or others and make sure we can feed and support them most efficiently.

Q: The five years since 2000 were not kind to the electronics industry. How did you drive innovation during the period?

Templeton: Because we divested the DRAM business back in '97 or '98, we were very focused on analog and DSP going into that horrible downturn in 2001. We stayed true to the strategic direction. If you take a look at our research and development spending, we held it [down] through very tough years in 2001 or 2002. We continued with a 300-millimeter wafer-fab investment. It was our first 300-millimeter fab and brought it up. Even in the face of very tough economic times, we stayed the course on R&D spending and advanced technology spending.