When a vision vanishes

Sections: Have a plan B Vision and risk Timing is everything Hedging bets Where to grow from here CEO survival

Brian Halla (pictured) once had a vision. He would transform National Semiconductor Corp., Santa Clara, CA, from a steady but stodgy producer of analog and low-cost commodity chips into a passionate purveyor of a system on a chip (SoC) that would be the heart of a whole new category of electronics. With the phenomenal explosion of the Web, dozens of companies were designing low-cost devices that would connect to the Internet. National could rely on its longtime analog markets to remain profitable while it achieved fortune and fame by hitching its star to this sexy new Internet appliance (IA) market.

National built the SoC, combining digital and analog features in a chip called the Geode. The market, however, never materialized. Nobody could figure out how to make an IA that offered anything more practical and low-cost than the tried-and-true PC, especially as PC prices were plummeting and the frenzy over all-things-Internet was dying down. After a five year effort, Halla finally threw in the towel in February, when he announced a financial restructuring that included the sale of the Geode and other cost-cutting measures.

Investors and analysts expect a CEO to have a vision that rallies the troops and gives everyone a common goal. If that vision pans out, the CEO is a hero, says John Stieber, management consultant with Hagberg Consulting Group, Foster City, CA. But what happens when that vision vanishes? Then, the CEO has some explaining to do. A close examination of Halla's experience at National provides some lessons on what CEOs should and shouldn't do when their vision doesn't bear fruit.

Have a plan B

The most important lesson: Make sure you have a backup plan. "To his credit, Halla didn't bet the entire company on this IA strategy," says Megan Graham-Hackett, an analyst at Standard & Poor's Investment Advisory Services LLC, New York. Even as Halla was chasing Internet-era glory, he did not turn his back on National's analog business, which accounts for 75% of its revenue. Halla continued National's R&D in power management and other profitable analog areas. Now, Halla is relying on a growing demand for analog to return the company to profitability and improve shareholder value. "Our core business is analog, which remains a thriving business," asserts Halla, who responded in writing to submitted questions.

Wall Street seems to like this new plan. National's stock rose two points on news of the Geode sale. "National is a strong analog company hidden by all the SoC junk that never went anywhere," says Eric Ross, an equities analyst at Investec Ernst & Co., New York. National now stands a better chance of improving shareholder value. "The good thing about the analog market is that it is quite fragmented," says Mark Edelstone, an equities analyst at Morgan Stanley & Co., Inc., New York. "You don't have to be the biggest player. If you're innovative, disciplined and a good manufacturer, you can be a solid, profitable player."

At least one big investor agrees with Halla's new strategy. Ralph Whitworth, who buys stocks in companies he thinks have mismanaged assets, has acquired 7.2% of National shares, including a block after the Geode sale was announced. Whitworth wants to get two board seats, for himself and another colleague from his investment firm, at the shareholders' meeting this fall to make sure National stays the course.

Vision and risk

CEOs are paid the big bucks to offer bold visions. But they also must ensure that they don't become so attached to that vision that they are blind to reality. "There is uncertainty in any decision," says Stieber. "Sticking to your knitting also has a downside. You can get out of touch with a changing environment."

When Halla became CEO, in May 1996, the pioneering chip company had hit bottom. The former CEO, Gilbert Amelio, had initiated many changes, only to leave a turnaround half finished when he abruptly departed for Apple Computer Inc., Cupertino, CA. National executives at the time privately expressed frustration and a sense of betrayal. The company desperately needed energizing.

Silicon Valley was gripped by Internet fever in 1997, the year Halla paid $550 million in stock to acquire Cyrix Corp., Richardson, TX, an x86 microprocessor company. He knew there was risk, admitting at the time that analysts and his own employees might think some of his decisions, including the Cyrix acquisition, were crazy. (See EB, "The Vision Thing,"  March 1999.) But Halla believed that National's IA chip had to include a microprocessor core based on the x86 architecture that was so pervasive in the PC market.

"At that time, in 2000, we thought we had a great opportunity to build a new category of devices that would connect people to the Internet and become a profitable business." —Brian Halla, CEO, National Semiconductor Corp.

 

In the heat of the frenzy, Halla wasn't the only one to envision a future for the IA. "I also believed in the IA concept five years ago," says Krishna Shankar, an equities analyst for JMP Securities Inc., San Francisco. "In hindsight, everything associated with the Internet now seems extremely overhyped."

Halla also thought National might be able to compete against Intel Corp., Santa Clara, CA, and Advanced Micro Devices Inc., Sunnyvale, CA, in the low-end x86 PC chip market. But it quickly became obvious that it couldn't, so National kept the IP it needed for the Geode—the x86 core—and sold the rest of the Cyrix assets for $167 million in 1999. "History shows that Cyrix was too expensive," says Shankar. "Maybe National should have focused elsewhere."

National then used the x86 technology and its own analog expertise to make a reference platform for a Web-surfing appliance, which was well received at Comdex in the fall of 1998. Two years later at Comdex, Sony and several others introduced Web-surfing appliances built on the Geode. But they never caught on.

A primary reason was that PC economics had overtaken the IA. In 1997, low-end PCs cost $1,500. By the time the first Web-surfing appliances came out, PCs were selling for $600 or less, says Shankar. "I think we were all surprised at how PC prices crashed," he says. "We've also seen the evolution of the PDA and the cell phone. What would have been the IA market broke into these other categories."

The Geode also had some sales in set-top boxes and thin clients, but total Geode sales never amounted to more than about 4% of National's revenues.

Timing is everything

Correcting strategy is one of the hardest things a CEO must do, says Irv Grousbeck, a consulting management professor at Stanford University Graduate School of Business, Palo Alto, CA. When goals are not met, the CEO must determine whether the people, the processes or the goal itself is the problem. "At some point, you have to say, 'This isn't getting done,'" says Grousbeck. Most companies don't make strategic corrections quickly enough, he says.

Even though many analysts were initially enthusiastic about IAs, some of those same analysts now fault Halla for not getting rid of the Geode sooner. "CEOs should be willing to make bets," says S&P's Graham-Hackett. "But the writing has been on the wall for some time now. Halla should have cut loose a year ago."

Criticism from equities analysts oriented to the short term doesn't surprise Halla. "Everybody loves to second guess," he writes. "We nurtured the Geode business because our customers wanted our products. At that time, in 2000, we thought we had a great opportunity to build a new category of devices that would connect people to the Internet and become a profitable business. The prolonged economic downturn…is the overriding factor in understanding what happened to the IA market."

Even after selling the non-Geode Cyrix assets, National still "had to spend a lot on R&D to advance the cores," notes Morgan Stanley's Edelstone. "It proved to be too expensive." Edelstone applauds Halla for finally acting this year.

While National continued to spend on the Geode and other projects, its revenues fell. Because of that, R&D costs grew to 30% of sales in 2002, compared to 21% in 2001. National had a net loss of $121.9 million on revenue of $1.49 billion in the fiscal year ending May 26, 2002, compared to net income of $245.7 million on revenue of $2.11 billion the previous year. National improved its results in fiscal 2003, posting a loss of $33.3 million on revenue of $1.67 billion.

One reason Halla may have hung on to his vision was a faith—shared by those in many other chip companies—that the market was bound to turn around soon. "Until this downturn, we've been used to 20% growth in every second year," says Don Macleod, chief operating officer. "Our industry is now into the third year of zero or anemic growth."

National and other chip companies entered the downturn with quite a bit of cash from the Internet boom, Macleod adds. This made it easier to stick with the Geode longer than it otherwise might have been. National had $850 million in cash and less than $25 million in debt in its most recent quarter, Macleod says.

But National executives reached their limit after ending fiscal 2002 with a loss and the lowest revenue under Halla. Last summer they determined that they would need to cut R&D back to 20% or less to become profitable at current sales levels. If the economy didn't improve by the end of 2002, they agreed, then they would sell the Geode. It didn't, so they put Geode up for sale.

"The decision to sell the IA business was based on a sustained depressed market and a payback scenario that was too far out in the future," writes Halla. "We [finally] decided to sell the business because it wasn't giving National the kind of return we wanted as soon as we wanted it. Simple as that."

In another cost-cutting measure announced in February, National put its cellular baseband unit up for sale. The loss of revenue from it and the Geode won't affect the company much, and it will improve its bottom line from the savings on staff, R&D and other costs, Macleod says. In May, National closed the cellular baseband unit, saying it could not find a buyer for it. The staff reduction from this action was about 300 or 4%.

National also renegotiated its contract with Taiwan Semiconductor Manufacturing Company Ltd. (TSMC), Hsinchu Park, Taiwan. It will no longer pay licensing fees to use TSMC's advanced processing technologies in National fabs. Instead, it will outsource to TSMC all 0.15-micron and smaller production—which means leading-edge digital silicon. As a result, it will be able to cut 500 positions from its workforce of 10,000. (It already reduced its workforce by 10% in May 2001.) These moves save about $15 million per quarter and will reduce the capital expense-to-sales ratio from 20%–25% to 10%–15%, Macleod says. National will continue to make analog chips at two fabs in the United States and one in Scotland.

Macleod estimates the total savings from all the steps at $20 million per quarter, which would make National profitable without any revenue growth. It would also free more resources for developing analog products and analog manufacturing processes, for which it already wins praise from analysts.

Hedging bets

The history of electronics is replete with examples of companies investing in markets that didn't develop. Smart CEOs hedge their bets. "A CEO should not assume that the wave he is trying to catch is a good one," says John O'Neil, president of The Center for Leadership Renewal, San Francisco, who runs seminars for electronics executives. He says partnerships and several smaller investments, especially in start-ups, are preferable to spending a lot of money on a single market.

Even though his focus on the IA market grabbed the spotlight, Halla was doing other things to improve National. Under his leadership, National sold its commodity businesses, which included logic, discrete and nonvolatile memory devices; stepped up efforts in analog markets with higher margins; improved reuse of its intellectual property; shortened its design cycles; and developed a premier Web site for design engineers.

National will never know how it might have spent the money it invested in the Geode, but to its credit, it did not stop investing in its bread-and-butter technologies. Although the company declines to specify how much of its R&D spending has been on analog and non-Geode projects, Halla insists that "during the IA era, we also more than doubled our investments in analog."

During the IA years, National made several analog investments. For example, it acquired Vivid Semiconductor Inc., Chandler, AZ, a firm that designed flat-panel displays, for an undisclosed amount. It also entered into several analog-related partnerships. Last fall it announced one with ARM Holdings Ltd., Cambridge, England, the leading developer of microprocessor IP for cell phones and other devices. They are collaborating on a microprocessor core with power management that should appear in cell phones beginning in 2004, says Peter Henry, vice president of power systems. They'll share royalties.

Thanks to moves such as these during the IA era, many analysts believe that National is positioned for growth. "I'm excited about its direction," says Shankar. "Now we need to see good execution on this new strategy."

National ranks sixth (see "Top 10 analog IC vendors," below) in the $24-billion worldwide analog market (see "Worldwide analog market," above). It would no doubt like to improve that standing, having slipped from fifth place during the IA years—although a return to profitability is its first priority.

Where to grow from here

National's best shot at high growth is in power management, audio and display. It sells all three into the wireless handset market, which accounts for 30% of its sales. It is counting on the cell phone to continue to grow (see "Wireless handset market, worldwide," above) and to produce even more revenue as manufacturers add color screens, digital photography, e-mail, music, games and other features. Power management IP, which will be crucial to the feature-laden cell phone, is arguably the crown jewel at National, which is No. 1 in the analog power IC market (see "Worldwide Analog Power IC Suppliers, 2002," below).

Two years ago, National had an average of about 30 cents of the cell phone bill of materials (BOM), says Henry. It now has an average of $1 and as much as $3 on high-end sets, in power management alone. The display is another potential $6 to $8 per handset, says Macleod. "National is still gaining content on the cell phone," says Shankar. National "can double or triple its content."

Flat-panel displays are also a key market, accounting for 11% of National's sales. The market is growing fast (see "Worldwide market for flat-panel displays," above), and Macleod says National now offers all the analog IP needed for a display, about $20 to $35 of a display's BOM.

Many analysts are optimistic. Jim Feldhan, president of Semico Research Corp., Phoenix, AZ, says National has innovative technology, solid manufacturing, good analog design and a long history of delivering products. Graham-Hackett says National's standing in the analog industry has been maintained or even improved some during this downturn.

CEO survival

Halla gets credit for giving the IA his best evangelism. "Whatever the vision is, you have to get out there and preach it," says Stieber. However, he also catches flack for it. "Halla has a big mouth," says Investec's Ross. "When others had shut up, he was still predicting when the economy would take off."

Even now, Halla continues to argue that the IA market will come to pass, that it has just been delayed by the sour economy. This may have cost him some credibility, and he might be better off simply owning up to the failure. "He is not well regarded among institutional investors," says Adam Parker, an equities analyst at Sanford Bernstein, New York. "You'd probably see the stock bounce up a couple of bucks if Halla were replaced."

"The good thing about the analog market is that it is quite fragmented. You don't have to be the biggest player." —Mark Edelstone, equity analyst, Morgan Stanley & Co. Inc.

(Several weeks later, however, Parker issued a research report that applauded Nationals restructuring and upgraded his rating on the stock.)

Edelstone, Graham-Hackett, Ross and Shankar all believe that Halla's job is secure unless he fails to execute his new strategy. "As long as Halla executes on this vision of growing analog in cell phones and other devices, everyone will be happy," says Shankar.

None more so than Whitworth, managing partner, Relational Investors LLC, San Diego. In his U.S. Securities & Exchange Commission filing, Whitworth says National is underperforming the industry and cites the IA and cellular baseband businesses as being among the reasons. At the time, Whitworth applauded the restructuring. He declined to be interviewed for this story.

"Until this downturn, we've been used to 20% growth in every second year. Our industry is now into the third year of zero or anemic growth." —Don Macleod, chief operating officer, National Semiconductor

 

Whitworth targets companies that have underperformed because their assets are mismanaged, says Ross. "He invests in only a few companies per year, he gets on the board and he changes the way management does things." He says Whitworth doesn't invest before talking to other institutional investors to make sure they agree with his direction. Ross doesn't believe that Halla is at risk. Whitworth, who rarely removes CEOs, endorses the restructuring. "Some of the press want to label this as 'Halla versus Whitworth,' but it isn't that way. Whitworth is a very savvy investor who operates in a straightforward manner."

O'Neil says a CEO shouldn't automatically become the fall guy when a vision fails. "If the CEO fully engaged his team and the board fully supported that move, then there should be no punishment." By all accounts, Halla engaged his team and had board backing and even some Wall Street enthusiasm for his IA vision.

After any such failure, O'Neil says, the CEO, his executives and the board "should sit down and spend a lot of time together asking themselves what they want to have happen and how they can learn as much as possible from this."

Halla and his team appear to be doing that. And if they aren't, Whitworth is bound to nudge them.

Is Halla making the right moves with his restructuring of National? Send your comments tofeedback@eb.reedbusiness.com.

Bill Roberts is a contributing writer for Electronic Business. Send him e-mail atbrobert1@ix.netcom.com.

WORLDWIDE ANALOG POWER IC SUPPLIERS

Rank	Company	(millions of $)	2002 Revenue 2002 Share
1	National Semiconductor	$430	12.1%
2	Texas Instruments	$355	10%
3	Fairchild Semiconductor	$316	8.9%
4	Linear Technology	$275	7.7%
5	ON Semiconductor	$244	6.9%
	Other	$1,941	54.5%
	Total	$3,562
SOURCE: DATABEANS INC.

TOP 10 ANALOG IC VENDORS

2002 Ranking	Company	Revenue (millions of $)
1	STMicroelectronics	$3,360
2	Texas Instruments	$3,100
3	Infineon Technologies	$1,791
4	Philips Semiconductor	$1,438
5	Analog Devices	$1,274
6	National Semiconductor	$1,244
7	Maxim	$917
8	Sanyo	$889
9	Motorola	$873
10	Toshiba	$839
SOURCE: GARTNER DATAQUEST