Scott McNealy’s Lonely Battle Against Eternal Night
Silicon Valley has always had a soft spot for Scott McNealy, and it’s easy to understand why. For the nearly two decades he has run Sun Microsystems–the longest tenure of any current high-tech CEO besides Oracle’s Larry Ellison–he has always said and done what others wouldn’t or couldn’t, and he’s often been right. In 1991, after Sun had spent a decade building engineering workstations, McNealy’s peers thought he was nuts to take on Digital Equipment, HP, and IBM in the server business (servers are machines that handle centralized computing tasks like running a database), yet he beat them. In 1995, when he told the world about the Java programming language, they thought it would never reach critical mass, yet it has become an Internet standard. When they told him his broadsides against Microsoft–calling Bill Gates “butthead” and his products “hairballs”–would only harm McNealy’s and Sun’s reputation, the rhetoric helped inspire the biggest antitrust lawsuit in a generation. One way or another, all those gambles helped make Sun one of the biggest success stories–and McNealy one of the biggest business heroes–of the 1990s.
Now McNealy has a new crusade: cutting the high cost of corporate computing. Once again he’s saying what his peers can’t or won’t– that hardware and software companies, including Sun, have been gouging customers. Customers already know that, but wait! McNealy says Sun has a way to reduce the $1 trillion a year U.S. corporations and governments spend on IT by a factor of ten. The pitch goes like this: High tech no longer needs to be sold in pieces that corporate customers must cobble together; it will be sold instead as pretested systems that work with everything else a company owns right out of the box, much like consumer PCs today. “Only in our industry do we create our own unique jalopies,” McNealy says. “You don’t do that when you buy a car. But in high tech somehow we’ve become obsessed with selling customers componentry. We sell them piston rings and carburetors and crankshafts and dashboards and tell them to put it all together themselves. It’s incredible! I’ve yet to see a corporate server room that looks like any other. They’re like fingerprints.”
That’s an interesting way to look at the future of tech. Some would even say it’s right and noble. The only problem is, even if McNealy’s vision of computational Eden does come to pass, there’s a good chance that Sun Microsystems as we know it today won’t be around to enjoy it.
Sun, to put it bluntly, is in deep trouble. Its once-booming server business has hit the skids–hard. Corporations that loved Sun’s servers during the Internet bubble are turning to cheaper machines from IBM and especially Dell. Sun’s sales are down 37% from their peak, and the company has lost money two years in a row. Its stock, which topped $60 a share back in August 2000, now sells for about $3.50.
And things don’t seem to be getting better. Earlier this month, for the second quarter in a row, the company said results would fall significantly short of expectations. Sun lost $587 million in the year ended in June 2002 and $3.4 billion for the year ended last June; analysts expect it to lose another $400 million in this fiscal year. Revenues have dropped from $18.2 billion in fiscal 2001 to $11.4 billion in fiscal 2003.
Even a customer whose name Sun provided as a reference to FORTUNE has a hard time saying nice things about the company. Marc Benioff, CEO of Salesforce.com in San Francisco, says he loves his Sun servers and plans to buy three more, but adds that the company is out of touch with its customers: “They don’t know how to sell and market in the modern world, and it’s killing them.”
McNealy’s new crusade may actually be an excellent example of what Benioff is talking about. The we’ll-sell-you-everything-in- one-tidy-unit pitch is catchy, but it’s hard to find anyone who’s buying it–not now, anyway. Corporate customers certainly want cheaper, easier-to-use hardware and software. They also want something else: cost control, which these days means getting their IT from different companies. “At its core Sun is a hardware company, and hardware companies try to lock their customers in,” says Marc Andreessen, the erstwhile boy-wonder co-founder of Netscape and current chairman of Opsware, a systems-management company. “Customers remember what that was like, they hated it, and they absolutely refuse to go back to it now.”
No one thinks Sun is going out of business. It has $5.5 billion cash in the bank, and $126 billion of Sun equipment is installed worldwide–equipment that should generate upgrade and maintenance revenue for years to come. But what makes Sun’s predicament more dangerous than its previous business scrapes is that when customers have cash to spend, they’re thinking of other vendors before they think of Sun. “Our CIO survey in May said that roughly a third of the 100 FORTUNE 1,000 CIOs we talked to are thinking about eventually migrating their platforms away from Sun. That’s not good,” says Merrill Lynch’s Steve Milunovich, one of the most respected computer industry analysts on Wall Street.
You don’t trust surveys or analysts? Talk to three of Sun’s once- key business partners, Oracle, BEA Systems, and Veritas. These software companies used to sell their products almost exclusively for use on Sun gear, which drove billions of dollars in sales for the company. Now they are aggressively rewriting their programs to work on other platforms, namely servers running Intel chips and the Linux operating system. Oracle now uses Linux, not Sun’s Solaris operating system, to develop and test its database applications. It plans to move its own corporate database off Sun machines to Linux too.
Milunovich is so concerned about Sun that this month he took the extremely unusual step of writing a three-page open letter to McNealy and the board questioning McNealy’s leadership. The letter, distributed to Merrill’s entire institutional and retail investor system, publicly states what you hear privately from industry insiders–that McNealy needs to hire a chief operating officer, lay off at least 14% of Sun’s staff, or roughly 5,000 people, and cut and better focus its $1.8-billion-a-year R&D budget. If Sun wants to survive as an independent company, Milunovich says, it needs to shrink its aspirations dramatically and accept a role as a much smaller niche supplier, perhaps using its superior technological expertise to serve the high-end computing market–airline- reservation systems, telephone networks, that sort of thing–where it still has a loyal following.
What’s happening to Sun is the same thing that happened to companies like DEC, Apollo, and Silicon Graphics in the 1990s, and to Wang, Data General, and Honeywell in the 1980s. Namely, faster and cheaper technologies came along and captured buyers’ imaginations before the companies could adapt. Sun’s value to customers in the 1990s was that its hardware and software were designed to work with each other, and were built from the ground up with the Internet in mind. Its servers were expensive, but they were also the fastest, most reliable, and most versatile machines on the market.
But just as the bubble burst, servers driven by cheap Intel processors–the ones that go in PCs–became in most cases faster than machines with Sun’s expensive SPARC chips. On top of that, the Linux and Microsoft server operating systems became reliable enough for corporate use. Customers, faced with a cost crunch from the recession, found they could buy IT systems for a fifth of what they’d been paying and turned away from Sun.
Even if you don’t accept McNealy’s vision for the future, you have to give him credit for tenacity. In the past 18 months he has turned Sun inside out. First he took back day-to-day management, parting ways with longtime president Ed Zander. He cut out two of five management layers. Perhaps most important, he appointed two executives, reporting to him, to be in charge of software and low- end systems. Sun has never made much money in either market because its salesmen always knew the juiciest commissions were in big, expensive hardware. Now a new commission scheme is making it worth the sales staff’s while to move software and the cheaper machines as well.
Customers are already seeing the results of the changes. In May, Sun unveiled two inexpensive servers that are competitive with Dell’s. Sun had long resisted selling machines not equipped with SPARC processors, yet the new models sport Intel processors and sell for $2,500 each, which is about the same price as Dell’s.
On the software side Sun has lowered its prices. For $100 an employee, corporate customers can now get not only Sun’s top operating system, Solaris, with their gear but also Linux and Sun’s entire suite of 224 “middleware” products. (Those are programs for running corporate e-mail, websites, security and authentication, transaction processing, and so forth.) The thinking, says Jonathan Schwartz, the newly minted software chief, is to offer a price so compelling that customers won’t be able to resist trying the package. Sun also announced, for $50 an employee, a Linux-based equivalent of Microsoft Windows and Office, programs that Microsoft, even with corporate discounts, charges hundreds of dollars for. “Just as Dell commoditized hardware, we are going to commoditize software,” says Schwartz.
McNealy has given his people another sales tool: a briefing center in their Silicon Valley headquarters where customers can visit with top executives, explain what they want, and take part in the design process. If a customer commits to buying a system, Sun will design it to his or her specs, test it, and install it for free. Sun executives claim their success rate at selling big systems using this approach is between 80% and 90%. The briefing center is so important to McNealy that last year he moved his office next door just so he can be a part of as many sessions as possible.
Other initiatives up McNealy’s sleeve are much more in keeping with his company’s supergeek reputation. There’s something called N1: a data-center management service that will enable Sun customers to manipulate their computer systems with the flexibility of an electric utility manipulating electricity flow. It’s a bigger deal than it sounds. The theory is this: Suppose you could pool the power of all your company’s far-flung computers and apply it anywhere it’s needed on demand. If you could do that, you wouldn’t have to buy so many computers.
There’s also Niagara, a chip Sun plans to release in the next two years that’s already getting extremely positive buzz in Silicon Valley. It is a complete rethinking of how processors work, and the word is that it could achieve at least a 15-fold increase in processing power over today’s most powerful chips.
The trouble with these initiatives is that it’s difficult to see how they’ll rekindle Sun’s growth anytime soon–and investors probably won’t give McNealy much more time. N1 and Niagara may be fascinating, but any significant contributions to earnings are years away. Sun’s middleware products may be the cheapest around, yet Sun has tried and failed twice in this market in the past five years. Schwartz says the company has redesigned its product line so that it all works together seamlessly this time, but BEA and IBM, the largest players in the $4-billion-a-year middleware market, are also investing heavily to improve their offerings.
Sun’s bid to compete with Dell for low-end servers seems equally high-risk. Dell is not averse to slashing prices to defend or grab market share. Schwartz contends that Sun can keep its own prices low by bundling together its software and hardware. (With Dell machines, most of the software is sold separately.) So Sun could win if customers were willing to convert en masse to its software– a big if. Competing with Microsoft on the desktop seems like an even tougher fight.
The bottom line is that McNealy’s turnaround strategy may be too late. The last time McNealy made a big strategic shift, from workstations to servers a decade ago, he started before Sun’s share of the workstation market eroded. When the Internet took off he had the best product on the market at the best price. Sun’s current initiatives come after nearly three years of financial ugliness and face a market in which neither McNealy’s vision nor his technology are cutting-edge. “They’re not in control of their environment anymore,” says Laura Conigliaro, an analyst at Goldman Sachs. “And that’s a big problem.”
What happens to Sun, long term? Silicon Valley loves to speculate about Sun’s getting acquired, which is what happened to DEC (sold to Compaq in 1998). But that’s unlikely for now. It’s hard to imagine McNealy, who owns 2% of the company, selling out to anyone. A much less horrifying scenario for McNealy might be to radically restructure the company, much as Apple did in the early 1990s when it gave up on the idea of going toe-to-toe with Microsoft and Intel in personal computing. This means doing a lot of things that McNealy up to now has said he would never do, like getting out of the chip business and ending his feud with Microsoft. Yet Sun’s N1 seems a particularly good project on which to focus, given that Sun’s core strength is solving extremely thorny computer-science problems.
McNealy (no surprise) thinks he can pull off another miracle. “When we get focused, when we get motivated, when we put all the wood behind that arrowhead, we’re going to cause some damage out there. A lot of people keep forgetting that,” he says. Actually, that’s not quite right. People haven’t forgotten; they just no longer believe him.
FEEDBACK fvogelstein@fortunemail.com
McNealy says Sun has a way to cut spending on IT by a factor of
ten.
The bottom line is that McNealy’s turnaround strategy may be too
late.
DINOSAUR EXHIBIT
Computer companies throughout the industry’s short history have
been forced to adapt or die.
COMPANY: Apple
BUSINESS: Desktop computers
PEAK REVENUES YEAR: $11 bil. 1995
WHAT ATE ITS LUNCH: Microsoft and Intel
WHERE IT IS NOW: A successful niche player
[COMPANY]: Digital Equipment
[BUSINESS]: Minicomputers
[PEAK REVENUES YEAR]: $14.3 bil. 1993
[WHAT ATE ITS LUNCH]: PCs
[WHERE IT IS NOW]: Sold to Compaq in 1998; now part of Hewlett-
Packard
[COMPANY]: IBM
[BUSINESS]: Mainframes
[PEAK REVENUES YEAR]: $87.5 bil. 1999
[WHAT ATE ITS LUNCH]: Minicomputers and PCs
[WHERE IT IS NOW]: The world’s largest computer company
[COMPANY]: Sperry
[BUSINESS]: Mainframes
[PEAK REVENUES YEAR]: $5.7 bil. 1985
[WHAT ATE ITS LUNCH]: IBM
[WHERE IT IS NOW]: Acquired by Burroughs in 1986, creating a
combined company called Unisys
[COMPANY]: Wang
[BUSINESS]: Word-processing minicomputers
[PEAK REVENUES YEAR]: $2 bil. 1984
[WHAT ATE ITS LUNCH]: PCs
[WHERE IT IS NOW]: Reorganized after bankruptcy in 1993; sold to
Getronics NV in 1999
Quote: “Just as Dell commoditized hardware, we’re commoditizing software.”