Roger Gilbertson, electronics whiz-kid, inventor and robotics maven was relatively happy working at a San Francisco Bay Area-based electronic security systems firm that supplied first-class hotels with reprogrammable guest room keys. Then one day he learned that the company was relocating to Chicago. It was a push that set the 25-year-old Gilbertson to quickly calculate his savings: barely enough for half a year’s living expenses.

But as entrepreneurial luck would have it, the push was followed by a pull. One rainy day soon thereafter, Gilbertson visited Futureworld Expo. He saw a unique Japanese technology called BioMetal—very fine wires that shortened when an electric current was passed through them. "My brain fell out right there! I bought some on the spot." An entire vision of how he could use them in tiny robots and retail displays danced in his mind. Three months later, Gilbertson made the leap from the corporate tower.

According to a Gallup Poll, about 13 percent of the people who leave corporate life every year do so to launch their own careers as self-employed professionals and entrepreneurs. Whether they succeed depends to a great extent on how well they plot their escapes and whether they make a good decision as to what business to be in.

Gilbertson’s epiphany at the trade show set him off on the right track. Rather than starting over as an independent in a new field, he took his skills and built a business around them, learning to leverage his expertise and existing network.

With the wires in hand, Gilbertson began discussing his dream with Christopher Paine, a former high school teacher who had become a lawyer. In exchange for stock, Gilbertson walked away with enough cash to incorporate. There was no time to think of a company name, so in the spirit of science-fiction comics, "I said ‘Mondo-tronics’ as a joke!" It stuck.

Gilbertson’s first product was a lot of little butterflies—$10,000 worth. Paine sold them to the high-end Bonwit-Teller department store in New York as an animated Christmas window display.

Gilbertson glided on butterfly wings for three months until he and some friends met to create a business plan that included developing and manufacturing mass market retail products. "We funded Mondo-tronics by asking friends and family to put in money that we knew they could afford to lose," he recalls modestly.

Ten years later, still home-based in San Anselmo, California, Mondo-tronics generated $1 million in 1997 sales. Gilbertson distributes 125,000 copies of The Robot Store, an annual mail-order catalog filled with inexpensive "build your own robot" kits. Most are based on those wires he saw at Futureworld Expo, re-named Muscle Wires.

Orders come in mostly by telephone (Gilbertson has contracted out 24-hour telephone order-taking) and via his Web site. True to the robotics theme, his business is primarily run by three computers, using Quicken and Applescript, augmented by two live assistants during the day.

Gilbertson still does original, custom designs for visual merchandising. He recently moved the warehousing and shipping operations out of his apartment and into a facility in Northern California’s wine country.

It’s daunting enough to transform a hobby into a viable enterprise such as Mondo-tronics. For real terror, try turning your back on a seven-year investment in corporate ladder-climbing, just after you reach the top. That’s what Kim Polese did less than two years ago when she made a high-profile dive from Sun Microsystems.

Polese had already been with Sun for four years when promoted to product manager for a programming language designed to allow users to send software applications around networks, no matter what kind of devices were sending and receiving them. The product, Java, was launched three years later and became an instant success. That was in 1995. But it was a success that didn’t scratch one particular itch: Polese had always wanted to be an entrepreneur.

Now with Java to her credit, she had global visibility. "The stars were aligned as never before," Polese says. "Silicon Valley was awash in venture capital. I had a great team of three guys I’d worked with on the most talked-about product in years. No mortgage, no husband or kids."

But before bailing out after seven years of fast-tracking, Polese and her team spent eight hours a day for 10 days during Christmas season 1995 brainstorming, planning and challenging the idea of striking out on their own with a company devoted to Java-based applications. They painstakingly tested every scenario. How would they exit Sun gracefully? What products would they launch? Who would buy them, and how would they finance the venture?

On day 10 they decided, Polese remembers, "It’s now or never." Each kicked in $15,000 of seed money to bootstrap the project for six months, and the endeavor was underway.

Polese says that venture capital groups began offering millions of dollars right away, but she turned them down so "we could bring the startup to the highest possible valuation before letting in such capital." They made sure that basic test versions were widely used, and that a revenue flow was being generated. "That way," Polese notes, "we didn’t have to give away so much when we eventually sold shares."

The new team first set up a bare bones office in a warehouse with equipment borrowed from friends and from Sun Microsystem’s loaner programs. "We also made sure to speak with Sun CEO and founder, Scott McNealy, to let him know we weren’t going out to compete with him," emphasizes Polese. Despite these cautious steps, things immediately got deeper than Polese’s technical know-how alone could handle.

On the first day out, for example, Wired magazine showed up to interview her. "We were being interviewed months before we had even named the company Marimba."

On the marketing side, Polese and her team—having created their first product—started looking at competitors’ pricing models to determine whether to charge a royalty, go for a one-time fee or settle on licensing streams. She answered this and other early startup questions by "talking to smart people, and staying plugged into my industry network. I love to have agenda-free lunches or dinners with people who have opinions, and let what I hear cross-fertilize in my mind."

(In 1997, Polese was named one of America’s 25 most influential people, along with Tiger Woods and Madeline Albright. And Newsweek named her one of the 113 most important people for the 21st Century.)

Some of Polese’s pricing and marketing advice came from her former colleague at Sun Microsystems, thirtyish Dean Ritz, who currently runs a small consultancy with two partners from his home in Washington. Ritz had specialized in product pricing and licensing services, as well as strategic marketing, during his earlier years at Sun and other firms in California’s Silicon Valley. Four years ago Ritz realized that even as a fairly well-salaried employee, he was stuck. He had no intellectual resources left for himself, nor the time to enjoy them.

His resolution came one day while he reviewed his situation in a "cookie-cutter apartment I was renting with a roommate in Palo Alto," Ritz recalls. "Commuting for more than an hour to the art and music and cultural richness of San Francisco wasn’t satisfying. I also realized I had no control over how much a project cost me in personal time and attention. I saw that as a salaried employee, I was selling all my time. All of it."

Ritz decided to move to Seattle, a city he frequently visited, where he found an affordable two-bedroom house only 10 minutes away from the live theater, museums, restaurants and music halls he enjoyed. The high-tech firms upon which his livelihood depends were close at hand as well.

Using some of the savings he had accumulated during several years of living unpretentiously, Ritz set up the office for his firm, Client Servers, in the bedroom with the best view and the most light. "After all," he says, "I spend days there with my eyes open. Why keep the best view for when I’m sleeping?"

In essence, Ritz launched a business doing exactly what he had done on someone else’s time clock. "The only pro-active marketing I did was to design and print business cards. My main marketing investment was telephones, and the time spent using them."

Ritz wasn’t shy about cultivating clients. Using existing business associates, he networked back into the companies where he had worked, or where he had trusted professional friends and acquaintances. The results: Sybase and Sun, plus a few smaller clients, now provide him and his two partners with enough business to maintain his previous income level, while working little more than half the hours.

When it came to establishing pricing, he relied on his previous experience as a corporate buyer for his kind of services. And Ritz did something "I’d never advise my clients to do. I raised the price with each job until someone balked. Then I knew I had reached the appropriate level." His free time is now spent playing classic guitar and riding motorcycles, with a liberal dash of perpetual house maintenance.

Like Ritz and Polese, Nick Cameron was just one step from the pinnacle of a corporate peak. He’d worked at Allied Chemical (now AlliedSignal) directing acquisitions and divestitures, and was one of the initiators of the firm’s quality management movement. After 35 years in this multibillion dollar company, Cameron was clearly the heir apparent to the chairman. But he realized that after endless years of commuting by air all over the country four days a week from his rural New Jersey home, he was never home. "With the chairmanship as the only place left to go," he says, "I decided to venture out from the familiar world and consult with startups."

Cameron realized a market was waiting for him the moment he put out the word that he was leaving Allied. Friends and colleagues immediately began asking him to consult with them. The only investment in his consultancy was intellectual capital. His extensive, long-term relationships among venture capitalists and other industry innovators provided a natural segue for his new independence. In his solo venture, Cameron develops business plans, helps raise venture capital, and fine-tunes human resources planning.

Unlike Polese, however, his own experience tells Cameron that starting up with venture capital money may be the wisest move. "When you get VC, you’re getting a lot more than money. You’re getting real help and management expertise as well."

Cameron believes it’s better to have five to 10 percent of a big success down the line, than to have 113 percent control of a small enterprise. "Too many people who go out on their own think they know everything and try to do it all," he warns. "They may have a dream, but that can also give them tunnel vision. Those starting out should know that there is a lot they don’t know, and get experts on board right away."

An expert might even be able to tell you when you’ve got a bright idea—by the wrong end. Just ask John Dillon (not his real name), who spent more than 20 years in high-tech industry sales. One day in mid-career he found that his typical corporate work schedule conflicted with his interest in starting a second family with his new wife. Dillon bailed out of his six-figure job and got into the packaged housing business in Northern California.

Dillon believed he could springboard on his background as a salesman, even though the packaged housing product was radically different from the technology products he knew. He financed the purchase of a dealership by taking a second mortgage on the new showplace home and office he purchased from the company for which he was now a dealer. Using demographics from the long-established company, Dillon placed local ads and displayed at home shows.

For four years he bumped along from deal to deal, project to project, "making just enough, but not exactly getting ahead." As the father of a new baby, Dillon found that an erratic income caused more anxiety for him and his family than his old workaholic life. "Frankly, I was getting buried in tax stuff, state, federal and local laws, regulations and forms—and missing the hustle and water-cooler politics of the corporate wars."

By the time his son was three, Dillon was ready to get back into commute traffic and pull down a steady salary in the industry he knew best. "I learned that whatever I found unsatisfying during my previous years in corporate life was far less frustrating than going it alone."

In retrospect, Dillon says his biggest mistake was choosing a whole new field. "I was mostly learning a living for four years, not really earning one."

Just ask Gilbertson, Ritz, Polese and Cameron: Leaping from the corporate tower can be as thrilling as any extreme sport. But if you stray far from your natural skills and talents, you may wind up viewing it like Dillon did: a disaster. It all starts with that operative question: "What do I want to be when I grow up, and how best do I make the jump?

Must Do’s:

- Dedicate a room to the exclusive use of your new enterprise.

- Set up a comfortable desk, chair and atmosphere, safe from interruptions.

- Dedicate a computer, phone and fax to your exclusive use.

- Name your venture and create stationery and business cards.

Hire experts to help if you aren’t already savvy about these vital steps.

- Get a business license, if required by your city.

- Obtain a Federal Tax ID number.

- Throw a launch party for your most trusted colleagues.

A Few Things That Can Wait:

- Cold calling. Don’t line up business before you’re entirely ready to start filling your agenda.

- A brochure or press release about your new endeavor and further marketing projects. Reminder: hire experts so you can focus on what you do best.

- Raise the capital to start your business, and accumulate at least six months of living expenses—the absolute minimum. Many ventures take three years to turn a profit.

- Be sure you have the support of your spouse or partner to stop receiving a predictable income.

- Don’t burn your bridges. Make a graceful exit, including honoring any non-disclosure or non-competition agreements you have made with your current employer.

- Maintain your network, so you’ll be welcomed back if you later decide to run for cover.

- Research any arrangements you should or must make with neighbors, city zoning and government regulations–especially tax laws.

Excerpted with permission from Small Business Success, Volume XI, produced by Pacific Bell Directory in partnership with the U.S. Small Business Administration.