(FORTUNE Magazine) – Tom Siebel, it turns out, cannot see around corners after all. It's a Wednesday in mid-July, and the cocky, combative, loquacious CEO of software provider Siebel Systems is hosting his company's second-quarter conference call with financial analysts. Normally such an occasion would be a forum for Siebel to brag that his company's software is telling him all he needs to know about everything from customer behavior to the state of the global economy.

Yet on this day Siebel is morose because the numbers are so utterly and indisputably bleak: Year-over-year sales are down 28%, profits are off by 61%, and the company is eliminating almost 1,200 jobs, or about 16% of its workforce. "It's very weak out there," laments Siebel. "There is just not a lot of business being done in information technology. We don't see any reason to assume it will get better," at least for the balance of the year. He notes that the "conventional wisdom"--meaning his earlier prognostications--called for the tech economy to start improving in the first half of 2002. "The fact is that the market did not turn up and is not turning up," he adds. Then, when asked for detail on potential sales leads, Siebel lets loose a whopper that would have been unthinkable even a year ago, when he was still ridiculing other CEOs for not having "visibility" into their businesses, presumably because they didn't use Siebel software. "We don't really have a thorough analysis of the pipeline here with us today," he replies.

Tom Siebel saying he hasn't brought along an analysis of how sales are shaping up? That's a little like Yo-Yo Ma neglecting to pack his cello for an appearance with the philharmonic.

How has it come to this? A company whose growth had been so torrid that even now it sits at No. 19 on FORTUNE's list of the 100 fastest-growing companies, Siebel Systems has shifted into reverse. Even though it continues to generate ample cash, the technology upstart will post bottom-line losses in the second half of the year because of the $275 million it will spend on severance, lease terminations, and other slimming-down measures. Sales aren't expected to return to 2001 levels this year or next, which is as far out as analysts are projecting. And Siebel's once go-go stock, stuck at around $9 a share, is down two-thirds this year, more than the declines for each of its biggest competitors: Oracle, SAP, and PeopleSoft.

It would be tempting to blame the economy alone for Siebel's troubles. Tom Siebel, for one, does. Certainly the tech-spending slowdown is largely what ails his company now. But he faces longer-term problems as well.

Siebel Systems is far and away the leader in what's known as customer-relationship-management (CRM) software, which, among other things, lets salespeople plan and account for their deals and gives customer-service reps instant access to information about consumers. The trouble is, Siebel software is highly tailored to each industry and hence very expensive--which is why the tech-spending drought comes at an awful time for the company. Its competitors are using the slowdown to improve their products and slash their prices.

After a run in which Tom Siebel seemed to make every right call in an unpredictable industry, investors are no longer willing to give him or his company the benefit of the doubt. In fact, the very practices that made Siebel Systems shine for nearly a decade--its hand-in-glove relationships with business partners and its insistence on buying from its customers whenever possible--now threaten to taint the company with the suggestion of excessive coziness that is plaguing much of corporate America. It's enough to make a guy who is renowned for self-assurance, well, morose.

Throughout the high-tech boom, Tom Siebel and his company were the stuff of Silicon Valley legend. Now 49, Siebel arrived in the Valley in the mid-1980s as employee No. 40 at Oracle, and quickly became its star salesman. Along the way he developed a software program that Oracle sales executives used internally to manage their accounts and that he thought would make a nifty commercial product. After CEO Larry Ellison refused to fund the idea, Siebel left in 1990 and after time off became CEO of multimedia software startup Gain Technology, which he sold to Sybase in 1993 for $74 million. Siebel took his $10 million in earnings from the deal and with Patricia House, an Oracle marketing executive, started Siebel Systems.

Tom Siebel was an early master of partnering as a sales technique. His company raised no venture capital but did accept investments from wealthy and influential friends like Charles Schwab, still a board member, and companies like Andersen Consulting (now Accenture), whose then-CEO, George Shaheen, also continues on the Siebel board.

For eight years Siebel's intense focus on customers led to extraordinary growth. The company's first product, known clumsily as sales-force-automation software, lets salespeople record the progress of their deals. While the software makes life easier for the salespeople, its chief benefit is that it gives real intelligence to top executives. A sophisticated system of colored lights alerts managers to deal status, letting them see at a glance how sales prospects are shaping up across the entire company. Siebel's second major product, call-center software, is a staple of large consumer-oriented organizations, which use it to give reams of information to service reps while they're on the phone with customers.

When Siebel went public in 1996 (at a split-adjusted price of $1.06), its previous year's revenues were just $8 million. Five years later sales peaked at just over $2 billion; the share price topped out in late 2000 at almost $120. Along the way Siebel Systems became renowned for a corporate culture that seemed to have more in common with IBM in the 1960s than Silicon Valley in the 1990s. The company insisted on business attire at a time when dot-commers showed up for work in shorts. And Siebel was a tough place to work. Managers evaluated employees continually and twice a year fired the 5% who ranked lowest in performance.

Even after the dot-com bust, Siebel's business continued to accelerate. In early 2001, when Tom Siebel forecast a looming slowdown, he looked like a hero for slashing costs to keep profit margins high. So when he said earlier this year, in a flattering cover story in Forbes, that he saw an improving business climate, many took heart. But for once the famous flashing lights on Tom Siebel's computers were sending a false message.

Siebel wasn't alone in misforecasting the end of the recession. For him, though, the gaffe is particularly embarrassing because his software is supposed to give his customers a window onto the future. What is he supposed to tell them now? That his software is fine for forecasting sales in boom times but not when the worm turns?

In an expansive interview at his company's headquarters in San Mateo, Calif., three weeks after the earnings call, Siebel seems his old self again, relaxed and lighthearted. He admits he erred last year in keeping on 1,500 people more than the company needed in the hope it would beat up on the competition when business returned. Now, Siebel says, he has pared his staff to mid-2000 levels and is ready to move forward again. In fact, he argues, the downturn will wipe out so many competitors that Siebel will emerge even stronger. "The idea that we're in a permanent recession strikes me as unsupportable," he says. "This is a growth company. It will be a growth company again."

But the numbers suggest that growth won't resume anytime soon. According to analyst estimates, Siebel will have sales of around $1.7 billion for the next two years--down from 2000 revenues of $1.8 billion. Siebel is also under pressure from competitors hoping to gain ground with software that sacrifices fancy features for a compelling virtue in tough times: lower prices. Siebel's custom-made software is complex to install and typically costs $2,000 per seat a year for corporations with 1,000 users. Rivals are now offering plain-vanilla software suites that have much of Siebel's functionality for a lot less money. PeopleSoft might charge as little as $500 a year per seat for the CRM application it sells as an add-on to its human- resources and financial-management software. "There's a certain good-enough threshold beyond which extra features are simply nice to have or even add an unneeded layer of complexity," says Charles Phillips, an analyst at Morgan Stanley.

Though Siebel dominates CRM today, other players have an even bigger presence in the corporate world. SAP, which specializes in software that ties together back-office operations of large companies, has 18,000 customers and is now trying to crack the CRM market. "Our key is our installed base," says Michael Park, SAP's vice president for global marketing. Siebel, by contrast, has about 3,500 customers.

Needless to say, the competition is positively giddy about Siebel's perceived weakness. Says Larry Ellison at Oracle, which markets a CRM application as part of its software suite: "The specialty vendors will die over time. I think Siebel falls into [that] category. The suites always win." Craig Conway, CEO of PeopleSoft and like Siebel an Oracle veteran, adds, "The CRM industry has changed. Now not only should CRM not be viewed as the center of the universe, but I question whether it's a separate category at all."

Siebel is dismissive of the something-for-everyone competition. "Larry has been saying this since 1995," he retorts. "Oracle can't point to one large-scale CRM user around the world. It has 1,500 programmers working on CRM and still can't get a good CRM product." As for Germany's SAP, Siebel notes that he counts Siemens, Daimler, Deutsche Telekom, and Bayer as customers "in the heart of the motherland." He compares Siebel's expertise in customer-oriented software to Microsoft's dominance in desktop software. "I've invested $2 billion in this technology," he says.

Still, the company is fighting a two-front war. Besides attacks from SAP, PeopleSoft, and Oracle in the corporate market, Siebel's role model, Microsoft, is among a host of companies trying to sneak up from below. In July, Microsoft introduced a CRM product for small business. That has not been a Siebel market to date but could be a jumping-off point for pursuit of bigger game. Startup Salesforce.com, which sells low-cost subscriptions to its Web-based CRM software, is also popular with smaller companies. With forecast revenues of $60 million this year, Salesforce is a pipsqueak next to Siebel, but Microsoft won't be so easy to shrug off.

On top of its competitive struggles, Siebel faces questions about the quality of its revenues, namely the concern that too many of Siebel's sales are to its own suppliers. Each quarter Siebel reports a figure known as concurrent revenues, basically the sum of any transactions in which customers bought Siebel software within six months of selling goods or services to Siebel. Thanks to the notorious swaps and other types of barter transactions in the discredited telecommunications industry, such deals have gotten a bad name. But Siebel swears that his company's concurrent revenues aren't empty swaps. "This is not something of no value for no value," he says. "So far it's never been an issue."

Yet the concurrent revenues are becoming an issue. In 2000 such transactions accounted for about 1% of Siebel's licensed software revenues. The figure then fluctuated between 4% and 11% during 2001 and early 2002--and spiked to 18% in the second quarter of this year (see chart). That caused Wall Street to wonder what Siebel was doing to make its sales targets--and sent down the price of Siebel stock. Tom Siebel says the second-quarter increase was triggered by higher-than-normal spending on data centers that Siebel is building to hedge against a disaster at its main facility. He won't name the customers, but he's essentially arguing that the concurrent transactions, whose disclosure is required by the Securities and Exchange Commission, are coincidences. "It's highly likely going forward it will be a lower number," he says. PeopleSoft and SAP, however, promptly issued statements that their concurrent revenues are negligible. "I did barter transactions at Oracle for many years," says PeopleSoft's Conway. "Don't let anybody tell you barter transactions are happy coincidences."

Siebel also faces a credibility problem in its customer-satisfaction surveys, which the company relies on both to report how happy its customers are and to reward its employees. Siebel executives often note that its surveys are conducted by a rigorous, independent third-party firm called Satmetrix in nearby Mountain View, Calif. What Siebel neglected to point out is that it is a minority investor in Satmetrix, and that Siebel board member James Gaither is also on the Satmetrix board. Tom Siebel professes no knowledge of Satmetrix's finances and says his company's dealings with Satmetrix are at arm's length. But the coziness has irked analysts, who generally weren't aware of the connection. "They never told us when we were looking at the slides," complains Morgan Stanley's Charles Phillips.

Analysts and their investor clients are just as grumpy about Siebel's share price. Though its balance sheet is rock solid--Siebel has $2 billion in cash and equivalents--the company will face a math problem of its own making when business improves. Perhaps more than any other Silicon Valley company, Siebel has relied heavily on stock options to compensate employees. It has issued so many options--247 million at an average exercise price of $23.81--that at the end of last year, employee options amounted to 52% of undiluted shares outstanding. This so-called overhang, big even by Silicon Valley standards, means that the higher the stock price, the more likely employees exercising options will dilute shares owned by public investors. Analyst Patrick Walravens of JMP Securities in San Francisco estimates that at $10 per share--near the current level--options represent a 10% dilution of Siebel shares. At $20, the dilution would be 16%, and at $30 it would be 31%. In other words, says Walravens, "the dilution gets worse the better Siebel does. To justify a $30 stock price, they require 31% more earnings than otherwise would be the case."

Characteristically, Tom Siebel is not only unrepentant about the options grants but proud of them. He notes that the company raised a total of $60 million through its IPO and a subsequent offering, in the process creating what's currently about $4.2 billion of value for stockholders. Sharing the wealth with employees was always part of the plan. "The reason I didn't use venture capital was that I really wanted to have an employee-owned company," he says. "It's a great capitalist story. At the same time it's a Marxist dream, the workers owning the means of production. I wouldn't do it any differently."

Even if his name weren't on the door--or on everything else, for that matter--there'd be no disputing that Tom Siebel embodies Siebel Systems. In an interview in the large conference room adjoining his office, he sprinkles his comments with phrases like "my software" and "I have a small investment in" when it's the company's software and the company's investment he means. As he reaches for a pen to illustrate a point, a press aide quickly slides a blank sheet of paper underneath his raised hand so that the CEO can draw away. His Mercedes CL 600 coupe is parked in an unmarked spot next to the entrance of headquarters, making it easy for aides to check from their windows whether the boss is in the building.

Because Siebel has been outspoken in the past and willing to knock his rivals, the competition is happy to gloat today. Noting that Tom Siebel declared publicly in January that he could buy PeopleSoft--now worth over $1 billion more than Siebel Systems--but wasn't interested, PeopleSoft CEO Conway cracks, "Tom saying he's not interested in buying PeopleSoft is like Canada saying it's not interested in buying the United States."

Despite the stock's plunge, Siebel's 13.6% stake in his company makes him worth nearly $600 million. An undergraduate history major and still a history buff, Siebel is philosophical that the downturn is temporary and that the business climate has become unduly harsh. "Every CEO is a crook and every priest is a pedophile," he muses.

Longtime believers still believe in the company's prospects. "I would never bet against Tom," says David Ketsdever, a software entrepreneur who once was one of Siebel's investment bankers. Adds Emeric McDonald, a former portfolio manager with Amerindo Investment Advisors when it was a large Siebel shareholder: "Siebel's top management is chameleon-like. They're just good athletes."

For his part, says Siebel, all that's required is patience. When business turned bad, he replaced his goals, which every employee can read on the company intranet, with bare-boned directives: Continue to generate cash, keep product quality and customer satisfaction high, maintain market share. "That's it," says Siebel. "Keep it real simple. We're hunkering down and waiting for the storm to be over."

In the meantime, he has gone out of the prediction-making business. "I don't pretend to be an economic guru," he says. "It was my best professional judgment that the market would pick up in [the first and second quarters], and I was wrong. I'll take my shots on that. In hindsight, things are pretty clear to everybody, right?"

Fair enough. In the technology industry, the past doesn't count for very much. As for the future, it seems that nobody, not even Tom Siebel, has a crystal ball.