Eckhard Pfeiffer is a happy man these days, and it shows. In only a year as president, he has managed to restore the glitz to Compaq Computer Corp. Once the victim of price wars, Compaq is now the aggressor. Its low-cost ProLinea PCs, introduced last June, were an immediate hit, spurring industrywide price cuts and generating a three-month order backlog. In the company's older Deskpro lines, new models and lower prices have produced record sales. In August, Compaq branched out into laser printers and this fall brought ProLineas to Japan, undercutting market leader NEC's prices by 50%. Then it extended the price battle in the U.S. with cut-rate "servers" for PC networks.

No wonder Pfeiffer is happy. But not, repeat not, complacent. On the eve of reporting third-quarter earnings--the first to truly show the payoff of the new strategy--he sprang his October surprise: price cuts of as much as 32% across Compaq's already aggressively priced product line. The move, which Compaq said reflected the economies of scale it's achieving with its new high-volume strategy, brought its cheapest machine down to $799 and is bound to provoke more price cuts. It also came just in time to upstage IBM's announcement of its belated entry into the bargain PC business. "I know that everybody out there wants to be where we are," says Pfeiffer.

SINISTER SPIRAL. Indeed, in remarkably short order, the 51-year-old German-born executive appears to have put Compaq back on top. But the industry that Compaq is striving to once again lead has changed forever--and has become far more dangerous. The reason: With constant pressure to cut prices, margins are plunging.

So, Compaq is leading the charge, but at a price. Gross margins are down almost a third, to 28%, and net margin, at about 5%, is half what it was in the old days (chart)--and may never rebound. Even if the company sets a new record of $4 billion in annual revenues this year, as Pfeiffer says it will, Compaq's earnings will be less than half what they were in 1990 when sales were $3.6 billion. That means Compaq could have less money for the research and development that gives its PCs an edge. And it means any missteps will be costly. "There is not a whole lot of room between the guys losing money and us," Pfeiffer acknowledges.

It's not as if Compaq had much of a choice. A year ago, the company was jolted into that realization by its first-ever quarterly loss--the result of a 17% revenue slide and a belated cost-cutting campaign to match such upstarts as Dell Computer Corp. and AST Research Inc. Co-founder Rod Canion had misdiagnosed the company's troubles as mainly the effects of worldwide recession. As a result, he was unceremoniously dumped and replaced by Pfeiffer. Pfeiffer had concluded that there was something more sinister at work--the kind of corrosive downward pricing spiral he had seen in calculators in the 1970s, when he was a marketing executive for Texas Instruments Inc. Canion could not be reached for comment.

First as head of Compaq's enormously successful European subsidiary and later at Houston headquarters, where he became chief operating officer in January, 1991, Pfeiffer told anyone who would listen that it was time to slash costs and stop losing sales to cheaper brands such as Dell. By focusing almost exclusively on improving PC performance, the company was ignoring broad chunks of the market. "The business we left behind was wide open to the competition, and they went after it with a vengeance," says Pfeiffer.

But the message didn't get through. Compaq's roots--it was founded by a group of TI engineers in 1982--and its meteoric rise as the most competent designer and builder of IBM clones made it hard for Compaq executives to grasp the change. They couldn't acknowledge that their PCs had become overengineered, overpriced, and an easy mark for cheaper clones from Dell, AST, Gateway 2000, and others. "The Canion line was that profit problems were due to the recession," says Andreas Barth, Pfeiffer's successor as head of the Munich-based European division. "The first thing Eckhard did when he took over was to lay down the new law."

The law according to Pfeiffer: Slash bloated costs, forget gold-plated technology, except where it really makes a difference, and, most important, redirect Compaq's vaunted engineering to building high-quality computers at low prices. "The call to action was clear," says Ross Cooley, head of Compaq's North American operations. For Cooley, a defining moment came last January, when Pfeiffer declared that, no matter what, Compaq would match any manufacturer in the world on price. After that, says Cooley, "none of us ever looked back."

Now nobody has the time. With 41 new products announced in June alone and an explosion in distribution channels--from 3,300 to 8,300 outlets worldwide--Compaq's unit shipments are leaping at a 40% annual clip, rebuilding market share, which had slipped from 5.3% to 4.5% last year (chart). In the third quarter, revenues climbed 50%, to a record $1.07 billion, and earnings hit $49.4 million, beating even the rosiest projections on Wall Street. The stock now trades near 40, far from its all-time high of 74 1/4 in early 1991 but a nice bounce from the 22 1/8 of last December, and most analysts have switched to buy recommendations. "It's one of the great turnarounds in the world," says one of those analysts, Charles Wolf at First Boston Corp. "It is a rare example of a company that recognized a change in their environment and responded in the absolutely correct way."

LONG ROAD. The execution isn't nearly complete, however. It took only two years for Compaq to turn from the star of the PC industry to the goat. And Pfeiffer acknowledges that there's a lot to do to ensure that it doesn't happen again. "We have done phase one," he says. That involved two rounds of layoffs that cut the work force by 25%, to 9,000, lopping $225 million off the annual payroll. "They're doing the right things for the time being, but they still have a hell of a long way to go," says Kimball H. Brown, analyst with International Data Corp. "We need to see six to eight quarters of growth before we can decide if they can keep it going."

And there are constant threats to the turnaround. IBM's new low-end ValuePoints were quickly repriced to match Compaq's Oct. 20 cut. ValuePoints also challenge Compaq on technology with an IBM-built chip that runs faster than the standard Intel Corp. chips used in ProLineas. In notebooks, long a Compaq stronghold, IBM is becoming more of a threat, too. Its new ThinkPad 700C, for example, has the largest color screen on the market. "The new ThinkPad is so good it's scary, and the ValuePoint will unlevel the playing field," says IDC's Brown. There are still intense pressures from the likes of Dell and Gateway. Compaq shipments may be up 40% a year, but Dell's are more than doubling.

While continuing hand-to-hand combat in the PC trenches, Pfeiffer is also plotting long-term strategy--Compaq's first. He turned to future planning as soon as he felt comfortable with the immediate situation. For the past five months, with help from the management gurus at McKinsey & Co., Pfeiffer has been putting in place a comprehensive long-range strategy. "The focus now is clearly on future strategy," says Pfeiffer. "This company is totally transformed."

Maybe not totally. The new "vision" is hardly radical. While other PC makers, notably Apple Computer Inc., are racing to build futuristic permutations of the personal computer such as handheld "personal digital assistants," Compaq is sticking to its knitting. True, the company does have a PDA prototype, but, says Chairman Benjamin M. Rosen, Compaq's growth for the next five years "is going to come from our present area." The plan is to continue pursuing the corporate PC market with traditional products and use ProLineas to build a new high-volume business among small businesses and consumers.

FEW BELIEVERS. What's radical is how aggressive Compaq has become on price. "We threw our old pricing policy overboard," Pfeiffer says. "We are now pricing to market." In the old days, Compaq jammed all sorts of bells and whistles into its machines and priced them at or above what IBM charged. When rivals started matching those features--at lower prices--the strategy unraveled.

When Pfeiffer took over, hardly anybody believed that Compaq could survive in the new, low-margin climate, much less thrive. When he briefed analysts on the "New Compaq" early this year, they applauded his cost cuts but discounted predictions that he'd match Dell on price. They were astounded--as were the competitors--when the ProLinea line emerged with an $899 entry-level machine, priced below Dell's cheapest.

It was a neat trick for a company that, after setting growth records for a startup in the 1980s, by 1990 was sitting smugly on its laurels. Cosseted in a lush office park in north Houston, Compaq sprawled into 22 glass-and-steel buildings with gleaming atriums, jogging paths, and sidewalks lined with lush flower beds. "A slickness came into Compaq," says a former employee. "All of a sudden, if you didn't have an MBA, red suspenders, and these little Gucci slippers, you were nobody." In engineering, overkill seemed to be the motto. Compaq was one of the few PC makers to build its own power supplies. It also built casings--designed to withstand a fall off a mountain. The factory dropped in eight layers of radio-emissions shielding where four would do. All this drove up costs and tied up engineering talent but did noticeably little for performance.

"Success fueled good pay raises, good positions. Compaq people were spoiled by that," says Victor Allison, a former engineering manager. "There was so much momentum from this culture that when the market changed very quickly, Compaq could not adjust very quickly."

START SLASHING. Finally, there was no choice. By 1991, the growth company was shrinking. In a 14-hour board meeting on Oct. 24, 1991, Canion was fired, even though he had just announced plans for layoffs and other cuts. Rosen, an original investor, was convinced that he needed a new CEO. Working secretly with a group of Compaq engineers, the venture capitalist had started to map out a strategy to get Compaq into cut-rate PCs. Canion balked, and Pfeiffer was assigned the task of making it happen. "It's hard starting a company, but it's a lot harder re-starting a multibillion dollar company," says Rosen.

Ten days after taking over, Pfeiffer told his executives the basic plan: slash costs by 35% to 50% and prepare Compaq to compete in every segment of the PC market--both on price and performance. "It was like a new election. New people were going to be running things. I couldn't have stayed around," says Steven Ullrich, an engineer who was one of Compaq's first five employees. At least five other top exeuctives turned in their ID badges right away, including co-founder Jim Harris. For months, the managers who did stay on were briefed every two weeks by Pfeiffer to make sure everybody got the message.

One of Pfeiffer's most important decisions was his approach to the new low-end product. He organized an independent business unit to tackle the project and by last December was being briefed on a machine that looked nothing like a traditional Compaq and didn't even bear the Compaq label. There was even talk of having it built by an outside subcontractor. Shortly after the team presented its garish blue prototype, Pfeiffer hopped a flight to Munich to spend Christmas with his family. "Something just didn't feel right," he recalls. From the airport, he phoned Houston and told them to start over. "It finally sunk in after all these meetings that there was really only one way to go." If Compaq was going to build a low-price machine, it had to be a real Compaq.

Good move. Sticking with the Compaq logo and in-house manufacturing have been key to the success of ProLinea and a companion line of budget-priced Contura laptops. Pfeiffer had been struck by something that cropped up during reviews of advertising agencies last year: "They all came back and told us what an incredible name and confidence Compaq brought to this industry, even during a period of trials and tribulations." A cheap clone--even under a different marque--would devalue the Compaq name, Pfeiffer realized.

So he pushed his designers and manufacturing managers to out-Dell Dell, which cuts costs by farming out almost all manufacturing. To get ProLinea's price down, Compaq jettisoned that beloved gold-plated technology found in more expensive Compaqs. Engineers took costs out by, for example, designing custom circuits to reduce the number of chips needed in each ProLinea. An added bonus: Fewer chips means less chance of faulty connections. Computer magazines have given the machines rave reviews. "The ProLinea still has what it takes in the important areas," say Byte Magazine reviewers Steve Apiki and Tom Thompson. "Performance is good, and its high level of integration should make it a reliable system." In other words, they conclude, "a Compaq is still a Compaq, even for $1,000."

A BANG. The new ProLinea and Contura lines, sold through mass merchandisers as well as through computer stores, started off with a bang. "By the last week in June, everybody knew we had a tremendous success on our hands," says Cooley. "We were dancing in the halls." By July, Compaq was outshipping both IBM and Apple in the dealer channel. "Demand has been awesome," says Edward R. Anderson, president of ComputerLand USA. Compaq's share of machines sold through the chain doubled from a year ago, to 26%. The backlog is so huge that Compaq missed its goal of catching up by September, despite adding a third production line in Houston and Singapore. Now, Compaq says it's building 200,000 ProLineas a month. It expects to work off the backlog byDecember.

To get manufacturing costs down at the plants in Houston, Singapore, and Scotland, Compaq had to rethink its processes. It started building an entire system on a single assembly line instead of making the motherboard in one building and the chassis in another. It stopped testing every subassembly and now tests a sample. All finished systems are still fully tested. Most important, Compaq leaned on suppliers to cut prices, which is why, when longtime partner Connor Peripherals Inc.'s prices were higher than Quantum Corp.'s, Compaq made Quantum the main disk-drive supplier and Connor only secondary. Compaq later sold its 20% stake in Conner for an $86 million gain.

Even competitors are impressed with Compaq's resurgence. "Since June, they're really back in there, being competitive for the first time in two years," says Martin Slagter, vice-president for European marketing for archrival Dell. "I'd put them among my toughest competitors now."

And the contest isn't just being fought with the new ProLineas and Conturas, says Pfeiffer. He admits he was worried that customers who once paid $3,000 or $5,000 for a Deskpro would simply opt for a $1,000 ProLinea. They haven't. "Our volume of all existing products is up dramatically," says Pfeiffer. "There's been practically no cannibalization. This comes as a very, very pleasant surprise." Generous price cuts on Deskpros helped. Donald W. Matto, manager of technical services at Chevron Corp., says he always liked Deskpro quality but had switched because of price. "Now that the prices are more competitive, we're buying again," he says.

The ProLineas, meanwhile, are picking up the customers that Compaq had left behind. By going into mass merchandisers, the company reaches consumers and small businesses. Compaq just tripled such outlets, to 600, signing CompUSA, Circuit City, and Office Depot. "We want to have our products where our customers want to buy them," says Gian Carlo Bisone, head of North American marketing. "We have to play on all the tables."

The latest move is into direct response. In November, six Compaq dealers will start, and within 60 days, Cooley says, Compaqs will also be sold in catalogues and via telemarketing--Dell territory. "I think Dell is very vulnerable," says Peter H. Jackson, president of Granite Computer Products Inc., a direct marketer that expects to sell $100 million worth of Compaq PCs next year.

PATIENT BUYERS. To be sure, Compaq's longtime dealers aren't happy about the new competition, especially when there are shortages. "They've opened up other retailers while still not having product for their traditional dealers," complains Douglas Perry, vice-president of marketing for InaCom Inc., a chain of 1,200 dealers. "We've been trying to shift customers to other lines, but they've been surprisingly patient." Despite the shortage, he adds, his Compaq shipments soared 80% in the third quarter.

Computer dealers, and the large corporations they sell to, remain the biggest chunk of Compaq's business, and Pfeiffer hasn't forgotten that. The plan for shoring up this part of the business revolves around networking, mainly with servers, the high-end PCs that are hubs for networks of desktop PCs. Compaq made an ambitious play for this market with the SystemPro line, starting in 1989. But SystemPro was a symptom of Compaq overengineering. Price tags ran as high as $26,000, and a proprietary multiprocessor design required new software that didn't exist. The result: Compaq has less than 10% of the server market.

The new ProSignia servers are meant to change that. The base model uses a single Intel i486 processor and, at $2,700, costs 55% less than the lowest-priced SystemPro. "We will shake up the server market like we shook up the PC market a couple of months ago," says Gary Stimac, general manager of Compaq's systems division. Analysts expect servers to contribute some $375 million to Compaq's sales next year.

Servers are only part of the strategy in networks. Pfeiffer's theory is that if the hub of the network is a Compaq, then it's a lot easier to sell the spokes--the computers and peripherals that connect to it. That's why Compaq jumped into laser printers, with a model specifically designed for networks. Again, it looks like Pfeiffer has a hit on his hands. The printers, in development for three years, run up to twice as fast as similar models from market leader Hewlett-Packard Co. but are priced as much as $1,000 less. Analyst Barry F. Willman of Sanford C. Bernstein & Co. says they could add $90 million in sales next year.

Compaq has the money to develop all these new products, despite its shrunken earnings, because of another hard-nosed decision by Pfeiffer. In 1990, sensing an opportunity to be a leader in a whole new phase of the industry, Canion created the ACE consortium. The group intended to use a speedy RISC (reduced instruction-set computing) chip built by MIPS Computer Systems Inc. instead of Intel chips. Compaq bought a 13% stake in Silicon Graphics Inc., a workstation maker, and the two began joint development of MIPS computers for Compaq. But last April, Pfeiffer pulled the plug on RISC development, withdrew from ACE, and sold back its Silicon Graphics stake for $150 million.

BALANCING ACT. By scuttling its ACE machine, Compaq freed millions in research-and-development money, Pfeiffer says. He's also back in bed with Intel, which regarded ACE as a threat to its dominance in PCs. Pfeiffer hints that improved relations with Intel will help Compaq lead the market with the next generation of Intel microprocessors.

The danger is that the balancing act that Pfeiffer is now executing so gracefully could come crashing down. Pressure to cut prices and costs is not about to ease, but he's limited in how much he can cut before he starts sacrificing Compaq's advantage. To ensure top quality and innovation, the company wants to do as much of the manufacturing as possible. But that means a lot more bricks and bodies than such rivals as Dell need. Compaq's fixed costs are about 6.3% of sales, compared with 1.5% for Dell. That limits Compaq's ability to be the low-cost supplier. Even today, its average selling price is some 30% higher than that of Gateway, about the lowest-priced national brand. Dell Chairman and CEO Michael S. Dell insists there's no way Pfeiffer can be tops in both manufacturing and distribution for long: "I'm convinced that it's hard to play both sides of this equation," he says.

Just watch, counters Pfeiffer. By owning its own source of supplies for important parts, he says, Compaq can control costs and quality better than any other PC maker--and get innovations to market faster. Even now, Compaq's research budget is $180 million a year, about 5% of sales. Granted, that's down from $197 million last year, but it blows Dell's budget--$33 million on $890 million in sales last year--out of the water.

"We are predestined to continue the leadership role in this industry because we have the resources," Pfeiffer says. Still, he has to prove over time that he can keep hammering rivals on price and his factories on cost to give Compaq the wherewithal to build in something extra. If he can't, trailblazing Compaq winds up being just another cheap clonemaker.

Before it's here, it's on the Bloomberg Terminal.