(FORTUNE Magazine) – At about 9 P.M. on Oct. 15, Cheryl Ways' office phone rang with a call from her husband: "What are you still doing there? Come home," he begged. Ways, a 30-year-old IT worker at tech giant Agilent Technologies, told him she'd leave soon. Then she went back to work. A half-hour later, she shut off her computer and walked down Agilent's darkened corridors. It was hours after most of her co-workers had left, hours later than she typically stayed, and the last time she'd do it in her career.

Three weeks earlier, Ways had been laid off. In that, she was just another statistic in a truly dreadful year for the paycheck nation. According to outplacement firm Challenger Gray & Christmas, corporate America announced nearly two million job cuts in 2001, the most since Challenger started its survey in 1993. Agilent, an $8.3 billion spinoff of Hewlett-Packard, had been hammered not only by the downturn in the economy but also by the downfall of the telecoms that were once big buyers of its chips, electronic components, and testing and measurement devices. So Ways' getting caught in a downsizing isn't surprising. What is surprising is the manner in which Ways left: not by spending her lame-duck days updating her resume on company time or prying the A's off office keyboards, but by working harder, longer, and with an intensity that rivaled the most productive periods of her five years at Agilent and HP.

The answer to why she did it--and how Agilent was able to wring such performance and devotion from tens of thousands of employees working beneath an ax--provides a lesson for all companies struggling to slim down in these post-boom years. In many ways, Agilent had a head start: It had worked hard on the way up to gain its employees' trust. But on the way down, Agilent has also made a series of smart moves involving good management, good planning, and most of all, empathy. In interviews with dozens of current, former, and soon-to-be former Agilent employees, almost no one had a bad word to say about the company. Which is how Agilent was able to slash, cut, and downsize its way to No. 31 on this year's list of the Best Companies to Work For.

Let's get one thing out of the way: This isn't a story about money. In all of FORTUNE's interviews, Agilent's severance--a package that involved at the minimum three months' pay and career-placement help--barely came up. Instead a host of other actions were raised: the Hail Mary steps the company took to avoid downsizing; the barrage of e-mails and face-to-face meetings with top management down; even the tired sound in the CEO's voice as he delivered news of mass layoffs. Together, these created an atmosphere in which people like Ways--three months after being axed--could say of her bosses, "I felt horrible they had to do this," and of her former co-workers, "This was my gift to them: to leave my job in the best way possible."

Letting employees go in a humane fashion isn't just about finding a way for executives to sleep at night. Bad downsizing, which management experts argue is the norm, is bad business. A recently completed study of 18 years' worth of downsizing data by Wayne Cascio, a business school professor at the University of Colorado at Denver, found that while large layoffs can boost stock prices for a couple of years, they don't lead to greater profits. One problem is that while expenses drop, revenues tend to drop too, as the remaining workers cope with what organizational behaviorists call survivor syndrome: the anger, fear, anxiety, frustration, and decreased risk taking that can follow mass layoffs. "Just when you need employees to take risks to turn the organization around, they take to the trenches," says David Noer, an employment consultant in Greensboro, N.C. "You end up with a double loss."

Agilent couldn't afford survivor syndrome. When HP spun off the company in November 1999 to focus on PCs and computer equipment, Agilent was left with technologies that general consumers not only never saw but would be hard-pressed to understand if they did. Agilent's customers were engineers and scientists who demanded the finest high-resolution spectrometers, the best Ethernet-over-SONET mapper chips, the most advanced mouse cDNA microarrays--products that win or lose in the market thanks only to their technological edge. To get out of this recession, Agilent needed employees motivated enough to invent great products.

Agilent is run by three people, one of whom is alive. At the earthly level, there's CEO Ned Barnholt. Barnholt, a 35-year veteran of HP, sports big gold-framed glasses that droop slightly at the sides. He wears his gray hair in a comb-over. He is completely without pretense. When Barnholt, 58, chats with employees--something he does frequently enough that in the past six months he's given speeches to and taken questions from some 20,000 workers--they call him Ned; calling him "Mr. Barnholt" elicits giggles from employees. But while Barnholt manages the company, he seems to consider himself only a caretaker for HP founders Dave Packard and Bill Hewlett. All big decisions are vetted with the question, What would Dave and Bill do?

Indeed, the entire company functions on the premise that it is the real heir of Packard and Hewlett. When HP's house archivist had to choose between staying at HP or going with Barnholt, she went with Agilent. The company's Palo Alto headquarters is home to Hewlett and Packard's original 200A Audio Oscillator, which launched HP, and its labs still churn out testing and measurement devices. But even more, Agilent considers itself the true keeper of the HP Way--the management objectives devised by Hewlett and Packard that spelled out how to treat customers, shareholders, and most of all employees. The Way's key precept is that workers will give their best if they're treated honestly and listened to. In his 1995 book The HP Way, Packard asserted that making people feel that they are working toward a common purpose or solving a common problem creates "participatory management." To get there, all managers had to keep to a strict open-door policy and practice MBWA--management by walking around.

But by the late '90s, MBWA had turned into management by consensus. As the company's growth slowed, both former CEO Lew Platt and HP's new CEO, Carly Fiorina (who came aboard in the summer of 1999, as Agilent was spinning off), struggled with the dilemma of rejuvenating the business without tossing the Way. At Agilent, Barnholt's solution was to adopt it, but with three new "values" that he mentions at every turn: speed, accountability, and focus. To make them more than corporate buzzwords, he instituted a pay-for-performance plan for managers, for example, and retrained all 6,000 to get them making decisions faster and better.

For Agilent's first year, it seemed Barnholt had found the right formula. When the company announced in March 2000 that it had developed a photonic switch--a device it dubbed the "vital missing link" for the future all-optical network--its shares shot up 39% in one day, giving the company a $40 billion market cap. Hiring boomed as well. By November 2000, Agilent employed 47,000 people, a 12% boost in just one year.

Then came 2001: the year the revolution failed. Suddenly Barnholt had to figure out if the HP Way would work as well in reverse.

Early last March, Barnholt called a series of meetings with his top management, almost all of whom had been with HP for decades. They had heard their telecom customers report that business was bad and getting worse. And they had seen the cancellation rate on orders creeping up. Barnholt wanted to find a way to cut costs temporarily. What would Dave and Bill do? Easy. They would do the same thing they had done when HP's business soured in the '70s: Cut expenses and then, if necessary, cut salaries. So Barnholt froze hiring and slashed about 5,000 temps. He then asked all employees to cut discretionary spending--without explaining where, how, or how much.

Agilent peppered employees with information about why the cuts were necessary and how savings would help. Employees heard about it in e-mails, in Agilent's twice-a-week newsletter called InfoSparks, and in "coffee talks": regular news and brainstorming meetings that managers are required to hold with their people. Then Barnholt delivered his own take on the situation. As he does every quarter on the day Agilent releases its numbers to Wall Street, he stepped into a conference room down the hall from his office and recorded a speech to employees on the state of affairs. A few minutes later it was delivered, elementary school-style, over public-address systems companywide. He urged them to keep the fight going.

For Agilent workers, cost cutting became a calling. Employees built Websites to house data that once were printed out; they crashed with friends when traveling on business; they brought bags of chips to recruiting events instead of calling in catering. Without specific guidelines, no one knew how much they had to cut, so they tried to cut as much as possible. For Juan Yamuni, a 33-year-old analyst in Agilent's international treasury department, that meant curtailing approved business trips to South America, even though his managers never told him to cancel. "Top management was good about guiding you instead of getting a direct order"--he says, making a karate chop in the air--"from upstairs."

By late summer, the company had cut its travel expenses 50% and its printer and PC purchases 70%. But that wasn't enough to counteract the slowdown in business from Agilent's floundering customers. The company's revenue per employee fell from $71,000 a quarter at the end of October 2000 to $54,000 a quarter in April 2001 (excluding a 5,000-person division sold soon after to Philips Electronics). Even uglier, net income per employee fell from $6,500 to $1,900 in the same period. Instead of eliminating staff, Barnholt announced a temporary 10% across-the-board salary cut to wangle $280 million in annual savings. Employees cheered the move. "This was a matter of saving employees," says Stacy Yu, 25, who handles marketing for fiber-optic products. "It sounds hokey, but it's like a family. Everyone knows that we have to chip in to make sure that everyone else is okay."

But everything was far from okay. Sales were plummeting, with customers canceling almost $250 million in orders in Agilent's third quarter. For the first time, the company was facing a quarterly loss. Barnholt was forced to do something that Hewlett and Packard had never had to: cut staff. In mid-June he gathered his management team and challenged them to find a way to slash 4,000 people--9% of the company--and to do it without making them or the survivors feel mistreated. Hard enough. Then he laid out tough ground rules: Employees were to be told they were being let go only by their direct managers, and no across-the-board cuts were to take place. Barnholt wanted everything handled division by division, looking at each program and each employee. And he wanted all names ready in one month. "I was very worried," he says. "We were in new territory here. Our people don't have experience with this."

So Agilent went on a downsizing campaign that was two parts communication, one part execution. Barnholt knew that he had to set the tone. On Aug. 20, the day Agilent would report a quarterly loss of $219 million--its first ever--Barnholt broke tradition and got on the P.A. before releasing the information to Wall Street. He wanted employees to hear about the downsizing from him, not from CNBC. Barnholt first thanked everyone for cutting costs and cutting salaries. Then he presented the deteriorating state of business. As staffers across the country stared slack-jawed at speakers in the ceiling, Barnholt said that downsizing was unavoidable. He detailed how many people would lose their jobs, where the number came from, and how the "painful" process would work. "This is the toughest decision of my career," he said, his voice ragged. "But we've run out of alternatives."

The second round of communication came from the managers doing the layoffs. Barnholt sent more than 3,000 managers through a series of daylong training sessions at outplacement firm Drake Beam Morin, where they role-played and listened to the right and wrong ways to let people go. Back at Agilent, managers were told to be as honest as possible, to keep the door even more open than usual, and to field every question lobbed. At the same time, they had to evaluate every program and position they oversaw to come up with the 4,000 total. Forms used to help in the analysis were posted on Agilent's intranet; employees could see the criteria by which they'd be measured.

Just as people were being told whether they'd stay or go, Sept. 11 hit. Any chance of an imminent recovery disappeared. Barnholt again met with his executive staff and announced in November that another 9% would have to be cut. In one year, then, Barnholt would be eliminating 8,000 full-time workers and almost 5,000 temp workers--27% of Agilent's staff at its peak. The August downsizing program was repeated, with Barnholt crafting a new message and managers submitting new names.

"I knew that this wasn't part of the HP Way, and it's not what Bill and Dave would have wanted," says Benjamin Steers, a 26-year-old IT worker who joined Agilent out of college. "But if they were faced with the same situation, they would have had to do the exact same thing. And even though all of us probably lost sleep worrying about our jobs and whether we'd have them or not, I know Ned probably lost a lot more having to get up there in front of everybody and make this announcement and have to let go people in his family."

Management couldn't have written a better response. Agilent had succeeded in turning the "us vs. them" of corporate downsizing into just "us."

For all Agilent's communication, empathy, and family bonding, the company is not out of the woods, either financially or culturally. Tech and telecom companies still aren't buying as they used to. And if finances don't improve, another round of downsizing is possible. The more Barnholt cleans house, the less likely the remaining staff are to view him as the benevolent head of their family.

Take the paradox of the booming semiconductor factory. Employees at Agilent's Newark, Calif., fab (as it's called in the industry) developed a groundbreaking way to make special cell phone filter chips called FBARs. After years of research, the fab by early summer was just beginning to turn out chips. Despite that, Agilent's top management figured that they could save millions by shutting it down and moving operations--fewer than 5% of Newark's 300 people--to a bigger fab in Colorado.

Dave Allen, the general manager of the division, delivered the news. Standing in the cafeteria in early September, he explained the cost savings and told the workers that almost all of them would be out of a job within a year. The fab, he said, would close by October 2002. But he also said that since the chips were in high demand and since the manufacturing process was so new, he needed everyone to stay focused and committed.

At first, "everything went to hell," as Allen reported to his boss. For ten days, production dropped. Then it went up. And up. Soon Newark was producing more chips at a faster rate than anyone had forecast. Why? "We were brutally honest with them about what we're doing--what drove the decision, what the timing is, what's going to happen," explains Allen. "That honesty and integrity up-front is critical. If you don't have it, you lose their hearts and their minds, and they won't be productive."

That's the management point of view. But on the ground, more basic human instincts are at play, like hope and the desire to hold on to a job. A group of the fab's workers sat in a break room recently trying to explain why they've been so driven--even while being driven out. At first they couldn't come up with any definite reasons. But then one woman started talking. "Well," says Mary Dominguez, who has been working at the fab for 16 years, "maybe Fort Collins won't work. And maybe they'll let us stay." Everyone nodded.

FEEDBACK: droth@fortunemail.com