(FORTUNE Magazine) – The shrink reminds the geeks to play nice. "If somebody has an idea, don't stomp on it," says Anthony Creed, psychologist and full-time developer of new ideas at Intuit, the software company. "It's more important to get the stupidest idea out there and build on it than not to have it in the first place." The group of eight managers has just started brainstorming the next generation of the company's hit income-tax-preparation product, TurboTax.

"How about a tax-preparation product you could customize based on your emotional state?" suggests a product developer. "It could have a control bar showing how angry or happy you are." Another, more senior executive suggests a Three Stooges--themed product. "We don't have a funny tax product right now," muses Creed. "What about putting a 'wiki' into the software?" suggests a product manager in a fleece jacket. "Then the customers could write content themselves."

If you find it weird that such far-out thinking is being applied to improving tax software, you don't know Intuit. These free-association sessions happen at least once a week at the offices of Intuit's tax division

in San Diego, and similar gatherings are routine throughout the Silicon Valley--based company. In fact, they're just one stage in a systematic process of innovation created--and nurtured, fine-tuned, and evangelized--by Scott Cook, Intuit's co-founder and spiritual leader. It's a process that features near-maniacal focus on the customer. Before Creed even put marker to whiteboard, employees had conducted about 100 interviews with potential customers and visited 25 homes in Atlanta, Baltimore, and other cities to watch how real people prepared their taxes. At the end of the idea session, Creed whips out his PDA and photographs the board so that he and another team have a record they can use to flesh out the concepts. Then it's back to the customers, whose reactions will be analyzed. The process iterates until the company hammers out a product that people will pay for, use, and like.

Actually, "like" may be too strong a word. Few people truly enjoy financial software, which is part of the challenge facing Intuit. Ever since Cook launched the company in 1983, he has made it his mission to remove the agony from necessary but odious tasks such as bookkeeping and tax preparation. At Intuit they call them "pain points." Today the company dominates the retail software market for tax preparation and small-business accounting, at 79% and 85%, respectively, according to the NPD Group. But Intuit knows there's still a tantalizingly large number of Americans who don't use financial software--in the case of tax prep, about 20 million hardy souls use neither tax software nor a professional tax preparer. In other words, Intuit's biggest competitor isn't other software companies but rather, well, nothing. To keep growing, the company has to come up with ways to get at those 20 million people, many of whom would just as soon eat packing materials as use a tax program.

Over Scott Cook's shoulder is a human-sized model of the Statue of Liberty; above his head floats a real Russian space suit. This is Intuit founder Cook's hangout--breakfast at Buck's coffee shop in Woodside, Calif. It's famous as a meeting place for venture capitalists and entrepreneurs like Cook.

He notes several VC friends at other tables, and it gets him talking about leadership. He's no longer running the day-to-day operations of the company (his title is officially director and chairman of the executive committee). But his role hasn't diminished: He's a kind of benevolent venture capitalist, an entrepreneur-in-chief who helps bring up other entrepreneurs and establish an environment in which they can thrive. "We can actually coach people to become great entrepreneurs," he says. "We're trying to build something very different--a greenhouse for entrepreneurs."

Cook always knew he was good at the creative stuff, but it took CEO Steve Bennett's managerial chops to codify his ideas into a system. In 2000, Cook, then chairman, and Bill Campbell, then CEO, recruited Bennett from General Electric because they felt the company needed a leader with more experience managing large operations. Revenues were nearing $1 billion, and the company had 4,000 employees. "We had struck out at several businesses because we couldn't operate them," says Cook. He told the headhunter not to even look in Silicon Valley because, as he drily puts it, "tech companies are not known for either leadership development or operational excellence." But GE is. Bennett had been executive vice president of GE Capital, where he'd watched GE CEO Jack Welch (and Welch's successor, Jeff Immelt) at close range. Bennett knew he had some serious work to do when he arrived. "We had a lot of good ideas but either couldn't fund them or weren't disciplined at execution," he says. "After 23 years at GE, you learn execution."

Bennett and Cook have arrived at a symbiotic leadership two-step: Cook stirs the pot and comes up with ideas; Bennett makes them real. In the past six years revenue has more than doubled to $2 billion, and operating margins have risen from 4% to 26%. "I count on Scott to give me a lot of great ideas and concepts," says Bennett. "You've got to be really clear on where you're going, how you're going to get there, and how everybody's role--we call it 'line of sight'--ties directly to the business outcomes we want. It's about getting things out of your brain so organizations can become more effective."

One key way the new CEO got things out of the founder's brain was by pushing Cook to boil down his ideas into coherent, repeatable phrases. Perhaps the most crucial is "savoring the surprises," an expression that comes from the origin of Intuit's successful small-business-accounting software, QuickBooks. Cook started the company with Quicken, software to help individuals manage their personal finances. By the early 1990s, Quicken was taking off in the market, but a survey showed that almost half of its customers were using it to manage the finances of some kind of business. Cook dismissed it as a research anomaly. It wasn't until another survey a year later found an even higher number of business Quicken users that Cook saw an opportunity to build accounting software for small companies.

"Savor the surprise" is what Cook did once he finally accepted the market research. (QuickBooks annual revenues now total $753 million.) It's a way of looking at business that Cook has spread around the company. In fact, it's the principle behind one of Intuit's signature tactics for getting inside the heads of customers: dispatching employees to visit users at their homes or offices, or both, and watch them work.

As president of the Sacred Heart Nativity School in a modest neighborhood of San Jose, Father Peter Pabst has pain points far worse than most Intuit customers will ever experience. His job, as he sees it, is to prepare 60 boys, mostly children of recent immigrants from Mexico, to enter high school on track for college. But he doesn't always succeed: One of his former students, a recent dropout, was charged last year in a double homicide.

All this comes out as Pabst matter-of-factly answers questions put to him by an Intuit researcher visiting his office this sunny October morning. Pabst uses QuickBooks to manage the school's finances. "We have a horrible business model," Pabst tells the researcher, Jana Eggers, explaining how important fundraising is to his school. "Only 2% to 3% of our budget is covered by tuition."

The man may have a resilient spirit, but this financial software can get him flustered. Pabst is showing Eggers how he uses QuickBooks to track a donor's gifts. Though he's clearly adept at navigating the software, he periodically apologizes for how he's using it, saying, "I'm sure there's a better way to do it." Later, in the car ride back to the office, Eggers deconstructs the session--she feels she's learned a great deal about how QuickBooks might be made simpler.

Visits like this one--called "follow me homes" in Intuit parlance--are central to the company's processes, both for developing new products and for learning how to improve old ones. "What Intuit does in the field is wonderful," says Clayton Christensen, the Harvard Business School professor who wrote the bestseller The Innovator's Dilemma. "When a market is not well defined, the only way you get insights about what customers will buy is to go out and see what people are trying to get done in their lives." Two years ago the QuickBooks division sent more than 500 employees on follow-me-homes for three days. Even CEO Bennett tells of visiting a San Francisco woman who runs a small construction company and listening to her complain about QuickBooks as four cats crawled over him. Rick Jensen, who heads product development in Intuit's tax product group, recalls one light-bulb experience at a customer's home: "She got a screen about charitable donations, which asked if they were 'cash' or 'noncash.' She said, 'I wrote a check, so it must be noncash.'" Afterward, Intuit changed that language. Says Brad Henske, who oversees the entire tax group: "I read the other day that Microsoft hired anthropologists to study how people work. We have those people here--we call them employees."

After Intuit's research found that more than half of American small businesses that own computers still keep their books with simple spreadsheets or pencil and paper, developers in 2003 set out to learn how they might get those people to use QuickBooks. The ten-person team began with 40 follow-me-homes. "Small businesses told us, 'I don't need no stinkin' accounting!'" says Terry Hicks, who ran the project. "Many of them didn't even associate the benefits of good financial management with accounting."

So Hicks and his team decided to keep it simple. They built a stripped-down prototype of QuickBooks and tried that with the customers during some follow-me-homes. Still no go. "It was still accounting software, just with fewer features," says Hicks. "Customers told us they wanted to track the money coming in and the money going out. So we said, 'Got it! You want accounts payable and accounts receivable.' But when many of them hear the word 'account,' they think of their bank account. So in the second prototype we stopped using accounting terminology at all. Accounts payable became 'money out.' " This was a case of savoring the surprise. "It wasn't until we tried and failed with a prototype that the entire team was like 'Okay, now I get it,' " says Hicks. The first prototype required users to go through 125 setup screens just to get started, already far fewer than a regular QuickBooks product; Hicks and his team eventually reduced that to three.

In those last weeks of fine-tuning, the product that became QuickBooks: Simple Start Edition finally got simple enough. A few more traces of accounting jargon were removed. And the results in the marketplace after it was launched in September 2004 proved all the care had been worth it: In its first year it outsold all accounting software in the U.S. except QuickBooks itself. But to get to that final product Intuit had to cycle through its interview process with customers six separate times.

A touchstone of Intuit's approach to innovation is the acceptance of failure. People frequently refer to the fact that even Cook initially failed to recognize the opportunity with Quicken users that ultimately led to QuickBooks. "I've had my share of really bad ideas," admits Cook. At a recent companywide marketing meeting he presented an award for The Failure We Learned the Most From.

Failure, in fact, is more than just tolerated--it's an important part of the process. It's simply a reason to pick a customer's brain some more, mark up another whiteboard, and get down to business. As far as Scott Cook is concerned, if at first you don't fail, you're not doing your damn job.

FEEDBACK dkirkpatrick@fortunemail.com