Electronic Design

As Linear Technology Turns 30, Electronic Design Interviews Co-Founder Bob Swanson

 
1 of Enlarge image
 

 

 

As Linear Technology celebrates its 30th anniversary, Electronic Design editor-in-chief Joe Desposito spoke with Linear’s co-founder Bob Swanson (see photo) about how the company began, the challenges they had and how they view today’s and future prospects for the company.

Desposito: First, let me congratulate you on the 30th anniversary of Linear Technology. That's quite an accomplishment. But let's go back to 1981, the year that you and Bob Dobkin founded the company. At that time, the computer revolution was starting to pick up steam with products like Apple II Plus and Commodore 64 gaining wider market acceptance. But the big news in 1981 was IBM's entry into the personal computer market with the IBM PC. So, what gave you the idea that this was a good time to start a company focused on analog?

Swanson: Wow. You may be surprised to hear that founding the company had nothing to do with those events. We actually paid very little attention to the computer business at the time. The PC seemed to us to be an all-digital kind of product which, of course, it wasn't. But the reason Dobkin and I left was sort of the standard story in Silicon Valley, frustration overcame loyalty. The one thing we knew was that analog still had a future. Analog was already a pretty crowded field. In 1981, analog was considered to be an existing, kind of old technology, so to start a company in analog seemed like a challenge. We basically said, "Look, we're going to select some product areas, and only those product areas, that we think that we can design and deliver products better than everybody else, the whole pack.”

Then, as time went on, we were hoping to expand our area of expertise so that one day, we could say, "Hey, we are a leader in high-performance analog kind of across the board." So, that was our strategy. We certainly were aware that battery-powered portable products of all kinds were on the horizon, but we kind of had our eye on things like digital volt meters and medical instruments. This was before we realized that the PC had a lot of power content and, in fact, we were in all the early Apple desktop and portable machines because we basically did the analog part of it. So, that's how the company started at. More out of frustration than, "Hey, we have this great vision on an emerging market."

Desposito: During the 30 years that have passed since you started the company, digital technologies have slowly but surely chipped away at some of the tasks analog circuits used to do. I'm specifically talking about digital signal processors, but there are other examples as well. Do you see a declining role for analog in the future due to these advancements? Or do technologies like DSP actually expand the opportunities for analog electronics?

Swanson: This has been an issue really since we started the company. In fact, if you took a look at our first business plan, we talked about the data converter area. At the time we were starting the company, Intel had just announced that they were going to put ADCs on the front end and DACs on the back end of their microprocessors. That was kind of scary to us. We had a lot of respect for Intel. And so, we put in our business plan that we were going to watch to see how that went before we jumped into the data converter area. As it turns out, and you probably know this, Intel didn't have very much success with that. They were throwing away perfectly good microprocessors because they couldn't build an 8-bit A to D. And about two years into Linear, maybe three, I actually had a meeting with Dobby (Bob Dobkin, co-founder of Linear Technology) and a key strategic guy at Intel. We talked about this issue. And the Intel guy said, "You know, we have now come to the conclusion that there are certain things that should remain analog, and certain things that should remain digital.” And we kind of cheered and said great, we're going into the A-to-D business now. And we did.

But I think there’s a broader answer to your question. In the beginning people would say, “If you're not in the DSP business—even as an analog company—you're going to get packaged out.” In reality, we got to a billion and a half dollars in sales without having done that. I think that the data shows that for every socket that we lose to a microcontroller-based system or a DSP machine, a product is created that wasn't practical or viable until this DSP machine came around. If you look in most of these boxes that have a DSP at the heart of it, there are multiple analog chips surrounding it. So I think on balance, for every dollar that we've lost, we’ve gained a dollar or more. You know, when we started the company, the analog business was a grand total of $2 billion—a big market in '81. As you know, the market blew right through $42 billion in 2010 and is predicted to be $50 or $60 billion in the next five or six years with even modest growth. So I think the bottom line is that almost everything electronic needs analog, and a lot of these digital approaches turned out to be inefficient duplications of analog functions. Where they’re needed, they’re absolutely required, but I think there are still many analog functions where the digital approach just isn’t efficient.

Desposito: What you said about ADCs concerns one of my other questions. But before I get to that, let me ask you this. The world is changing rapidly with regards to electronics—I like to say that we're living in the age of disruption. Some disruptive technologies that readily come to mind are solid state lighting, the smart grid, alternative energy, electric vehicles, wireless communications, energy harvesting. How big a role does analog play in these technologies from your perspective, what do you think the prospects are for the future, and which of these technologies are the ones that you're really trying to focus on. Or are they all in your sights?

Swanson: I think the answer is they're all in our sights, and a sort of more general question, how big of a role does analog play? Most of these are not going to become viable businesses without good analog and advances in analog, especially in the area of electric vehicles, wireless communications, and energy harvesting. In many cases, analog is the missing link for these areas. Of those you mentioned, the only area that I looked and said, "What role does analog play and what new advances are needed?" is the smart grid. Off hand, I can't think of some great advancement that's going to be needed in analog in order to make the smart grid viable. I think the building blocks available today in analog are quite adequate to serve this market. But when it comes to technologies like energy harvesting, which is a brand new area in its infancy, I think it was true for us to claim that the analog block that we came up with—and now we have a whole series of them—was the missing link. We've always known that there were transducers that could convert delta temperature and vibration into electricity, but the amount of electricity is always so minimal that it wasn't adequate for most practical solutions. And then came along some devices from Linear that were so efficient they could take 20, 30, 40 millivolts, start themselves up and have enough energy left over to drive other things, particularly remote sensor applications. So, I think for all of the areas you mentioned, with the exception of the smart grid, analog breakthroughs are going to be part of the enabling technologies.

Desposito: Let’s get back to analog-to-digital converters. Microcontroller companies are sticking ADCs on the front ends of microcontrollers. They may be low-performance ADCs at this point, but do you ever see a time when the performance would be similar to a standalone ADC?

Swanson: Well, I think the answer is yes, they could if we stood still, but the analog companies are right in the middle of this data converter revolution-evolution. We're not standing still. And I think the analog part is still the difficult part. It's a core business to analog companies and as long as we don't stand still, I don't see this sort of evolution that you’re referring to as a killer for the analog companies in this particular product area.

Desposito: I took a look at your annual report and noticed that industrial is the largest end market by almost two-to-one over the next largest one. Can you explain a little bit more about the market from your perspective and tell us why the market demands Linear's expertise in analog?

Swanson: This has been a difficult question for me to answer since I was pitching the company to investors, sort of like, what is this Industrial, what does it mean? And the good news about it is that this segment purchases all our products from A to Z. The number of customers is in the thousands, so therefore the competition is diffused and the applications are numerous. But that's what makes it hard to identify it for either yourself or an investor as opposed to say a cell phone handset or MP3 player, which anybody can get their head around. It's all of this process control, all of this automation, all of this energy management that's a big deal to factories, collecting more information and improving the communications in factories, a lot of it wireless.

And then for us, industrial also means medical—the medical industry. And it includes test and measurement. A recent observation I made during this last economic crunch was that automotive and industrial for Europe and for Japan, at least in part, and maybe a big part, are kind of joined at the hip. So as automotive goes down in Japan, so does industrial. When automotive picks up in Europe, so does industrial. So, these are the two markets that we're focused on. Automotive and industrial are key markets for us now and for a lot of reasons and they're actually joined at the hip in some major countries.

You look at the industrial market and you just say, "So what could we put into it?" But it's kind of a long tail, and it surfaces almost every place you look.

And energy management has become a bigger and bigger deal as energy costs go up. So solutions like the so-called digital power management, which is a misnomer because most of the digital power management systems are analog loop with telemetry, allows us to monitor and manage certain things in power supplies. This whole drive to improve energy at the board, at the system, at the facility level, we’ll throw that in, too, that all goes into industrial the way we categorize it.

Desposito: We just looked at Linear from a market perspective. From a product perspective, which products drive the most or the least revenue for Linear and do you expect it to change or stay the same in the future?

Swanson: Obviously, power is in that broad category. Power is 60-plus percent of our business. Signal-chain products, which we break into two groups here—signal conditioning and mixed-signal products—drive about 40% of our business. But about three or four years ago, I kind of predicted—in more of a prayer than a prediction—that the things we were doing in the signal chain side, in A-to-D converters, in really complex drivers for A-to-Ds and for battery stack monitoring chips, and everything else we're doing in the signal chain side, could start to grow faster than the overall company. Not at the expense of power, but we’ve found some exciting  markets like base stations and networking. They all take power but they're also driven by signal chain kinds of products.

And so for us it's always open. We've been described as a company that skims the cream. We're high performance. We aim at high performance whatever the function is. And so people say, "Well, that's kind of a limited market." Over the years we’ve estimated that somewhere between 25% and 30% of the overall analog market is our sweet spot. But in order for it to be 25% and 30%, not 10%, we've got to have product offerings from A to Z, which I think as a company we do have now. And so, it's sort of a broad-based kind of approach to the market. We talk about three legs that can give a company like Linear diversity, which sounds like an oxymoron, since we're a high-performance analog company. But there is customer breadth, there's end market breadth, and then of course there's product breadth. So we try to optimize all three of these things so we're not totally dependent on any one of them, any one customer, any one end market, any one product area.

Desposito: Every company that's been in business for as long as Linear has had its share of ups and downs. What are some of the high points and low points during this 30-year span and some of the crucial decisions you've made that have helped Linear attain the leading position in the industry that it finds itself in 2011?

Swanson: Like all start-ups, we've had our ups and downs. When people ask me that question, the first incident I'm reminded of was just before we went public. Our sales were struggling around four million a quarter. We were going through cash at a rate of $250,000 a week. We had $1.8 million in the bank. I quickly figured out that a quarter of a million doesn't go into 1.8 million very many times, and I panicked. We had pay cuts and so forth and so on. We were actually thinking about going public then, but we weren't making a profit, and we hadn't broken even from a cash flow point of view. And then almost like magic, a couple of quarters later—this would've been about 1985—our sales jumped from four million a quarter to five million a quarter. All of a sudden, we were cash flow neutral and were actually break-even from a profit point of view. We were able to go public in 1986 with a $20 million a year run rate. We had broken even on cash at under $15 million, which was spectacular, since we had our own fab and everything. And so that was the first time I had nightmares about boards being on the windows. Then things went pretty well for us. It was up and to the right... Until the dot-com bubble burst, and we saw our sales fall from basically a billion dollars a year to half a billion almost overnight.

We always had a very healthy financial model, because we were committed to profitable growth. We saw no wisdom in taking a piece of business that didn't make any money just so we could brag about sales growth. And I think that philosophy exists today and it's existed throughout the thirty years. So when we were faced with this problem, back after the dot-com bubble burst, when half of our sales went away, half a billion dollars a year, we had to scramble, we had to cut costs. We actually had to shut down our four-inch fab, not because it wasn't profitable, but we had just built a bigger fab and we had a choice of supporting the big fab or the small fab, and obviously we supported the big fab. We had some minor layoffs. But what we didn't do was panic. We made some adjustments. We had to cut our spending. We have a profit-sharing bonus pool at the company that’s variable. When sales and profits go up, so does the profit-sharing pool, and when the opposite happens, the pool goes down.

At the end of the day, we turned in sizable profit margins at half a billion in sales, which was half our sales two quarters before. None of our competitors did this in the best of times; we did it in the worst of times. Then because of the way we handled our factory and our people during this crunch from about 2001 to 2002—when business started roaring back, we actually outperformed the market, because we kept our powder dry and were able to crank up production faster than everybody else. That led us to the crunch of 2008, same thing happened. Our sales went from $310 million a quarter to $200 million a quarter over a five- to six-month period. Same thing we learned from the dot-com bubble. We're financially healthy, we're still going to make a profit, we're going to have positive cash flow, but just on a much smaller sales base, so let's not hit the panic button, let's not shut our factories down.

Yes, we had to deal with expenses. Yes, we had to have factory shutdowns. Yes, we had to have pay cuts. But we kept the team together. And as luck would have it, what a lot of people don't know, is if our sales had fallen below $200 million a quarter, then the steps that we took wouldn't have been sufficient. But as luck would have it, we bottomed at $200 million and then quickly roared back to $400 million. Again, after this 2008-2009 crash, for six to seven quarters, we outgrew—everybody was going crazy as you know, there were shortages, which was the story of the day—we outperformed the competition in shipments by 50% during this six to seven quarter recovery period. So those are the most trying periods during the 30 years of Linear.

We made one other significant decision, if I backtrack to 2005. When we came out of the dot-com bubble, some people around here convinced some of us that there was a new thing called high-end consumer. Consumer was always a dirty word around here, and you washed your mouth out with soap if you mentioned consumer. We never thought consumer was a market where we could defend sockets the way we normally do, which is always to have a better IC when the competition catches up. But these were the days of the first MP3 players, the first cell phones with a couple of displays and a camera and so forth, and so we thought, okay, and we basically ran the table with some of the big players like Samsung. And by 2005, as it turned out, if you look at consumer products like MP3 players and digital still cameras and personal navigation devices and then the cell phone handset, which was a communications category, that represented 28% of our business.

But around 2005, we said, "Oh god" you know? This is a business that we can't defend the way we normally do. I was promised by our engineers that by the time the competition caught our first-generation products, we would have second- and third-generation products, and we did. But the big lesson we learned is that most of these customers didn't care about a second- and third-generation product if they could buy a first-generation product at half the price. So we learned that this is a market where price has overall importance for them to drive their market share. And so we made this tough decision around 2005 to basically de-emphasize all these consumer products, like the cell phone handset business which was booming then, but we could see no clear path for us to stay competitive in that area. And now some of the big stories in the papers, like The Wall Street Journal, are about how competitors are un-hinging themselves, or suppliers are un-hinging themselves, from the handset business. I think we were five years ahead of everybody.

In the areas that we said would appreciate our innovation, our support and our quality, there is price pressure in every market, but it's not number one with the industrial market. The automotive market has had a surge of electronics integration. They have everything from navigation systems to stop-start systems to safety systems and then the special set of automotive hybrid and electric cars requiring some very sophisticated battery stack monitoring chips. So we said, okay, we're going to do that. And we're also going to continue to be interested in communications, but not handsets. We'll focus on the build-out of the base stations for 3G, 4G, LTE or concentrate on the Ciscos of the world and the Huaweis of the world in the infrastructure business. We also think there are still opportunities in space and harsh environment markets, but these will never be a huge part of our business like they were at one time in the early days.

So those are the areas that we said we're going to concentrate on, and we're going to abandon this high visibility consumer side. And, of course, nine out 10 Wall Street guys thought we'd lost our mind, okay? And in fact, there were some people in the company that weren't sure if we’d lost our minds. But we've successfully executed this strategy. And this year that we just ended, which is our 30th year, we grew from $1 billion to $1.5 billion, and we did that by giving up $300 million worth of business in 2005, when we made a decision that we could do no longer defend those sockets.

Desposito: Yes, I remember when I visited Linear back in April, they did present the fact that you guys had pretty much exited the consumer business, and I was taken aback by that statement. But yes, when you look at that business, it's really cutthroat, and you're right, price is king. So, if you can be in a business where your quality and support matters, that's really the way to go.

Swanson: You know, Joe, that's our only strategy. We have no choice. We can’t compete on price. There's always somebody who's bigger than us or dumber than us, so we have to compete in markets, and fortunately, they're big ones that will pay us for our better designs, our better performance, our better delivery, our better quality. And so the industrial and the automotive markets are two that have turned out to want these kinds of characteristics in a supplier.

Desposito: You haven't mentioned this at all, but are you thinking about the infrastructure that will be needed to support electric vehicles. I realize that this is probably a small segment right now, but I'm starting to read more and more about it, where certain municipalities are putting in charging stations, basically.

Swanson: I'm not personally aware of it, but I'm sure our salespeople are all over it. I'm sure there's an analog opportunity there for us. But right now, in my kind of visibility, it's the electronics that go into to monitoring and managing the battery pack that is a critical application, and we seem to be well positioned right now.

Desposito: Thank, you Bob.

Swanson: Thanks, Joe.


For more about Linear Technology’s 30th anniversary, Linear has posted a video with Co-founders Bob Swanson (Executive Chairman) and Bob Dobkin (Vice President, Engineering & CTO) and CEO Lothar Maier, which can be found at www.linear.com/30yearinterview

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish