Insanity, someone once said, is doing the same thing over and over again and expecting a different result.
Scott McNealy, the brash, combative CEO of
, has been called a lot of names, but crazy isn't one of them. And after years of doing the same thing with increasingly disappointing results -- his company has reported nine straight quarters of year-over-year revenue declines -- McNealy is doing something different.
Sun, long thought of as a seller of high-end servers, is positioning itself as the company that will finally make enterprise computing rational, dependable and affordable. Instead of buying dozens of "moving parts," customers will be able to turn to Sun for hardware and a bargain-priced software bundle that will make a business network run, the company says.
Many analysts give that vision high marks -- in theory.
But convincing customers to bet their business on Sun software is going to be very difficult. "Leadership in the software infrastructure market is essential for the success of a 'systems' company, yet Sun is still not perceived as being credible by many of our clients," says Nick Gall, senior vice president of the META Group, an IT consultancy.
Sun has sold software for years, of course. And 50% of the company's $1.84 billion R&D budget is devoted to it, says John Fowler, CTO of the Santa Clara-based company's software group.
But in the last two years, Sun's software revenue (which the company doesn't break out) has declined by an estimated 25%, to $800 million, according to Gartner analyst Joanne Correia. And the last time Sun reported revenue for middleware -- the fiscal year ended June 2002 -- it lost $314 million on sales of $254 million via its former iPlanet subsidiary. (Total revenue that year was $12.5 billion.)
Moreover, Sun's Java programming language has produced very little direct revenue because it's free. "It's given Sun some momentum and a certain amount of indirect revenue, but it has not produced a sustainable business model," says Gartner vice president Yefim Natis.
The downward trend in Sun's software revenue isn't likely to reverse itself any time soon. Indeed, Sun's Project Orion, slated to debut in September, will essentially give away key components of its middleware stack to customers who buy its operating system, Solaris. The company figures that the bundle and a new, simplified pricing scheme will entice enough paying systems customers to more than make up for the loss of middleware revenue.
Sound like a dire state of affairs for Sun's software business? Not to Sun software czar Jonathan Schwartz. "Not a thing has changed in our strategy: create new market opportunities with software; monetize them with systems," he said in a recent interview. "I have no doubt in my mind that we will innovate our way into the market."
In its most ambitious bid to create new markets, Sun plans to make Java (as well as Sun's developer tool sets and other software) the platform for an emerging class of services that will run over the Web, mobile phones, smart cards or any other networked device.
Although Java is free, Sun wants to sell the directory software and servers that will power those services. And because the SunOne directory server is widely regarded as excellent and by some measures ranks as the market leader, Sun is well-positioned to make a go of Web services.
Besides creating new markets, Sun is targeting existing ones with greater precision.
Schwartz, who tends to talk faster and sound even cockier than his boss, has more going for him than bravado. He aims to capitalize on an ugly fact: Business software is complicated, expensive and often fails to deliver a significant return on investment.
Moreover, there is a deep reservoir of resentment against
high prices, leaky security and often arrogant behavior. Put simply, business buyers are looking for alternatives.
But in the important market for middleware -- the stuff that glues together operating systems and applications -- Sun has a lot to prove. Sun's middleware is "second- or third-tier," says Richard Williams, a strategist and senior software analyst at Summit Analytic Partners. "That's the perception we get from the developers we talk to -- the customers in the market -- which is ironic because at various points they had what was considered to be the very best app servers of their day."
Few analysts believe the company has a winning share of middleware, with the exception of the directory server. In application servers, a key part of any software deployment,
hold 66% of the market by revenue, compared to Sun's 4%, according to Gartner.
Then there's the desktop market. Here, Sun proposes to sell a low-cost open-source office suite running on a flavor of Linux -- offering to undercut Microsoft's prices by 50% over the cost of a total deployment. Asks Schwartz: Why should a worker doing transaction-oriented tasks or data entry sit at an expensive Windows PC running pricey Microsoft Office software?
To be sure, that means kicking off a price war with the world's largest software maker, but Schwartz shrugs off such worries. "It's all upside revenue to me," he says. But Schwartz won't say how much revenue the so-called Mad Hatter desktop initiative -- not to mention the rest of Sun's software's efforts -- is expected to generate.
Looking to the future, Sun also wants to stake out a big piece of the nascent market for Web services -- Net-based applications that automatically interact with each other to help make businesses more efficient. But no one has yet made serious money selling them.
Wall Street Says 'Show Me'
In the meantime, Sun is struggling to keep its head above water. On a per-share basis, earnings have been in the black only once over the past nine quarters -- in the June quarter of 2002, when Sun eked out a one penny profit. In the third and fourth fiscal quarters of 2003, net income of $12 million and $4 million proved too dinky to translate to positive EPS, which remained stuck at the breakeven point.
Sales, too, have been dismal. In fiscal-year 2003 ending in June, revenues from hardware -- which accounts for about three fourths of total sales -- dipped 14%. The company reported a massive $2.4 billion loss, though $2.1 billion of that reflects a write-down for impaired goodwill.
Sun's management is confident it's on track for a turnaround, but Wall Street has heard similar promises before.
Steve Milunovich of Merrill Lynch says Sun's new initiatives are "interesting long term but unlikely to impact the company's growth rate until fiscal year 2005 to 2006 at the earliest." Sun hasn't made a good case for why its products are any different or better than those of IBM or
, he says. (His firm has done banking for the company.)
One fund manager who sold off his Sun holdings a few years ago says he remains skeptical. "How excited can you get about a company with declining sales?" he says.
"It seems to me
CEO Scott McNealy just didn't have a mature, realistic appreciation of the fundamental nature of some of the challenges he was facing, and this relates to why I follow Sun with only half an eye," the fund manager says. "I don't see much of that realism yet."
Even Sun's current investors are far from bullish. Lured by Sun's bargain price, John Rutledge, manager of the
Evergreen Technology fund, opened what he calls a "tiny" position in Sun in May. "The stock price is cheap, they have plenty of cash and they're not losing money. Therefore, they're not going to zero," he sums up.
When that's the best an investor can muster, a company had better do a lot more than talk about success. It has a lot of proving to do. But if McNealy succeeds, he might be asking the Street, who's crazy now?