(FORTUNE Magazine) – Students, we have here a case history of Electronic Data Systems, the information-technology outsourcer in Plano, Texas. EDS, you may recall, was started in the 1960s by Ross Perot, bought by GM, and split off to the public in 1996. The company is big, with revenues of $21 billion, and it will do anything you want in IT: take over your back office, handle business processes like your HR department, do a sophisticated consulting project, you name it. Remember how often I've told you that business is unpredictable? This case will prove it. EDS has a stock that's been flattened and a CEO who's been humbled: "For the first time as a CEO, I got whacked off the saddle of the horse," he says. He's trying hard to remount. But EDS still has big problems, including an SEC probe that's just been upgraded to a full-out "investigation." So tell me, students: Today, early 2003, would you want to own this stock?

The stage for what happened was set in 2001, a smashing year for EDS and CEO Richard Brown. Profits hit a record $1.4 billion. The stock at year-end was $68 a share, a few points below its record high. The company absolutely knew the stock would climb from there. And Brown, three years into his job and with fine results behind him, had just earned--though this hadn't yet been announced--a fabulous $52 million pay package.

After that, in 2002, EDS had a year from hell. In the weak economy, its stock headed down instead of up; the company got hammered by the bankruptcies of WorldCom, US Airways, and United Airlines; and in the third quarter it had a disastrous earnings miss that sent its already battered stock from $36 to $12 in four days. Although the stock had risen to $17 by late January 2003, it was still off 75% in a little over a year, a fall that drove EDS's market value down from $32 billion to $8 billion. The company hasn't been valued that cheaply since 1990.

Brown, 55, who took his job in 1999 preaching "shareholder value," calls the year "horrible." Using his Sopranos word again--in fact, sounding a little like Tony in a shrink session--he broods about the earnings miss: "It whacked us hard. You get a pit in your stomach, and you feel like a failure, and your people are discouraged and surprised, and the market can't believe you're that far off what you told us you'd do not that long ago."

Brown knows that it's he who has the rat in his pocket and he who must figure a way to get it out: "I own this problem," he says.

White-haired--even before 2002--and intense about EDS, Brown often leans upon what he calls "coachable moments" to instruct executives about how they might better have handled a problem. In 2002, events took tight hold of EDS and consumed its year with coachable moments, some having broad application to business. The question now for the company is whether it has emerged from the coaching smart enough to regain the ground it lost.

If there was a flaw in late 2001, when so much looked so good for EDS, it was that cash flow was not, as CFO James Daley said recently, "very robust." One reason was a multiyear, $6 billion build-us-an-intranet contract with the U.S. Navy and Marines that EDS had won a year before. The company had exulted--raised Navy and Marine flags at headquarters, in fact--over beating its biggest outsourcing competitors, IBM and Computer Sciences, for this megadeal. Nonetheless, barnacles formed on the contract and caused EDS's costs to outrace revenues (as they continue to do). That, right there, is a pinching cash-flow problem, and in late 2001 it was affecting a lot of EDS's decisions.

For example: To keep the exercise of stock options from diluting its shares, EDS needed to buy back stock. But it chose to do much of its buying in 2002 not in the market but by means of derivatives contracts intended to lock in what it thought of as decent per-share prices. The idea was to space out the need for cash in 2002, yet not get caught in what EDS thought was a sure thing: rising prices for the stock.

Like many a racetrack horse, this "sure thing" didn't come in. So last September, EDS closed out its open contracts. All told, it paid parties on the other side a net $325 million to buy 5.4 million shares, for an average price of about $60 a share. That was expensive, considering that EDS's average daily close in the market through mid-September--which seems a fair price to think of as an alternative--was about $51. Taking the difference of $9 and multiplying by 5.4 million shares gets you to a cost of close to $50 million for the derivatives strategy. Says Brown, appearing to state the obvious: "If I had it to do over again, I wouldn't do it. We're wiser now."

Nor would he, given a second chance, say one thing he said at EDS's annual meeting last April. The stock had fallen by more than 20% since the year began and some shareholders strongly objected to Brown's $52 million pay package. The proxy explained that the directors were rewarding his "outstanding performance and contribution" during 2001. A second sentence rephrased things to "extraordinary contribution and performance." None of that silenced the critics, one of whom wanted to know why anyone could possibly need that much money. Said Brown, trying to make a joke: "I have a very expensive wife." Now, there is a coachable moment!

That was a sideshow compared with the bankruptcies that followed. First came WorldCom, with which EDS had--and still has--a close relationship in which each sells to the other. Next was US Airways, whose IT business EDS picked up when it bought Sabre's outsourcing operations in July 2001. Besides that, as part of its "investments," EDS has leased aircraft to US Airways as well as to United, another bankrupt airline. All told, these three failed businesses cost the company $281 million in charges in 2002. Although you can make a case that EDS was foolish to get itself so deeply in league with the long-troubled airline industry, you can't blame it for being blind to WorldCom's fraud. So was the rest of the world.

That point was, in fact, one that EDS wanted to make to investors after WorldCom declared bankruptcy on July 21. So in early August, Brown and other executives went on a six-city tour to reassure institutional investors. William Nygren of Oakmark Funds remembers what he heard from EDS in Chicago: "Business is great, and we're still comfortable with the guidance we've given you."

Just weeks later, on Sept. 18, EDS made a conference call to analysts after the market closed and announced that its third-quarter earnings would be 80% below what the guidance had said! The company's per-share target for 2002 fell to between $2.05 and $2.10, down from more than $3.30 when the year began. The next day the stock dropped from $36 to $17. Among the investors bailing out was just about anybody who'd been thinking of EDS as a growth company.

Nygren has since said you can explain EDS's September shock in one of three ways--none of them comforting to investors. First, maybe management didn't know in August what it had to announce in September. Second, maybe it did know and told investors otherwise. Third, he says, maybe "the business model is worse than we thought it was." A management would just as soon not acknowledge any of Nygren's hypotheses to be true. But needing to concede one point or another, Brown has from the start opted for ignorance: "I am adamant," he told FORTUNE. "We didn't know." He's not pleased with that fact; he simply says it's true.

But the story of the earnings surprise seems to raise questions about both EDS's business model and what the company knew when. At the very least, it saw smoke signals earlier in the year when two of its competitors, IBM and Accenture, announced that their business prospects had weakened. EDS itself, says Brown, had observed that contracts were taking an unusually long time to close.

Although July was another omen, with revenues coming in "light," says Brown, past experience told him not to worry. EDS's quarters tended to be back-loaded, meaning that a lot of business --in the form of discretionary spending by existing clients--came along late. These add-ons--roughly comparable to change orders a homeowner might give his contractor--typically had better profit margins than EDS's base contracts. Given these facts, the company didn't agonize if the first month in a quarter, like July 2002, was disappointing. EDS simply looked forward to August and September to provide the "uplift" it needed.

Again stating the obvious, Brown says, "In retrospect, I think we were too optimistic." When August's figures came in from 60 countries and were "tiered up" into totals that management could look at, the results were dismal. EDS was now in the second week of September--and panic set in. Brown and his team did what anyone would: They checked back to make sure the data were accurate. No salvation there. No prospect either of September's being powerful enough to save the quarter.

When EDS faced Wall Street on the 18th, it reeled out a string of reasons for the earnings miss. But Brown and CFO Daley focused especially on three: a slowdown in new sales; cost problems in some existing European contracts; and a lack of the add-ons that normally provided quarter-end "uplift." And that, says Bill Nygren, is a business-model problem. If the add-ons don't materialize, EDS is in the soup.

Once the company went public with its bad news, its problems only multiplied. Procter & Gamble, which had been on the verge of signing a megadeal with EDS, suddenly backed off. Elsewhere, customers negotiating contracts with EDS "froze," says Dennis McGuire, CEO of Houston's TPI, the leading consultant to corporations hiring outsourcers. While the customers were on ice, Merrill Lynch analyst Steve McClellan publicized EDS's misguided derivatives contracts. That appears to have been the knockover for the SEC. On Oct. 1 it launched an informal inquiry into EDS, asking the company for information about the earnings miss and the derivatives.

The SEC inquiry, which was upgraded to a formal investigation on Jan. 17, is puzzling. As far as the world knows, the agency is solely interested in EDS's problems with earnings and derivatives. But because both seem to have been caused by ineptness--which is not ordinarily actionable--it is tempting to think the SEC is on to something else. Whatever the case, the investigation remains a wild card for EDS.

Otherwise, the fundamental question is, How does this company climb back from its year from hell? As a starter, there will be no executive bonuses for 2002. Headcounts are falling throughout the company, and it seems safe to assume that EDS's next proxy will show Brown earning something way south of $52 million. Beyond that, Brown avers what you'd expect--that confidence will be restored. "This is a good company," he says, "and a solid company." EDS got some support for that proposition in December, when two clients it had been negotiating with well before Sept. 18--Bank of America and ABN Amro--"unfroze" and signed large multiyear contracts, with a total value close to $6 billion. TPI's McGuire says the two companies had no doubt been calling each other, checking with other EDS customers, and digging through EDS's books. The apparent result, he says, was "no smoking gun."

Some skeptics think EDS will now cut prices to "buy business." Brown says that won't happen, and he's bringing in a new CFO, Robert Swan, 42, from TRW and before that GE, to make sure it won't. (The incumbent CFO, Daley, 61, is moving to a different EDS job.) Meanwhile, everyone at EDS has become a believer in cash flow. Yes, the company still wants megadeals, usually defined as contracts worth $250 million or more. But it doesn't want the kind that require it to make heavy capital expenditures up-front and wait for cash flow. Says Brown: "We need to show the market that this machine generates cash."

Chief sales officer Steve Smith says that EDS is generally more selective now in the contracts it goes after, and that it is definitely not "chasing everything that moves." He heads a new megadeal team trying to speed up the rate at which these deals go though the pipeline. Smith says the company also is seeking balance in the size of its contracts, wanting "singles and doubles to go along with the triples and home runs."

Will the new contracts have good profit margins built into them, margins that will temper the need for add-on business? Who knows? To investors, outsourcing contracts are basically impenetrable "black boxes"--maybe nicely profitable for the outsourcer, maybe not. "Not" seems more likely these days because around 80% of IT outsourcing is a commodity business, subject to stiff price competition. In that contest, says TPI's McGuire, EDS's edge is a fine reputation for "operational excellence."

Since the company does not currently have an equally fine reputation for forecasting, Brown has put an executive to work to improve that skill. He's also bent on having top management look deeply into EDS's biggest contracts--maybe several hundred of them--to search out trouble. Even so, he says, forecasting is difficult for a company that has 15,000 contracts and a worldwide business. "Things happen that you just can't predict," Brown says.

So he'll be giving up guidance, right? Wrong. When EDS reports its 2002 earnings on Feb. 6, it will very likely give guidance--probably not rosy--for at least the first quarter of 2003, and maybe for the year. Well, EDS is hardly the only company where guessing what may or may not happen is a habit tough to kick.

As in many case studies, knowledgeable people have varying opinions about how things will turn out at EDS. Here are a few:

Ross Perot, who founded the company 41 years ago but now owns only around $3 million of its stock, thinks EDS will be stronger for having gone through adversity: "When you beat on iron, you turn it into steel."

Kenneth Langone, a Wall Streeter who backed Perot when he was getting started and remains a large investor in EDS, says, "Dick has to prove he can get contracts that throw off cash."

An executive of a competitor who thinks EDS has a poor grip on its financial controls nevertheless bought the stock recently. "EDS is better than a $17 stock," he says. "It should be $40."

Rod Bourgeois, a Sanford Bernstein analyst who has followed the company closely, believes that, "unless EDS is completely broken, which I don't think it is, or the SEC finds major wrongdoing, this is a buying opportunity."

And then there's Bill Nygren of Oakmark. In December, with EDS around $18, he was kicking himself for not having sold when it was $20. The stock got back there in early January, and though Nygren's not saying, he probably bailed out--glad to be gone.

FEEDBACK cloomis@fortunemail.com