Entrepreneurship does not just guide the process of having an idea and then building a business. A true entrepreneur has to predict the future. The idea that a businessperson is great because he had "built a company from the ground up" is missing a major component-- the entrepreneur's original hypothesis. Without a well-defined theory and end goal, it's difficult to test a market and therefore difficult to start an enormously successful business. I think the original idea, the hypothesis, is sometimes given too little credit. Especially in the era of Web 2.0.
In scientific research, an experimenter develops a hypothesis with a suspected set of ideas and then builds an experiment to support the hypothesis. I think this is a great way to think about building businesses as well. Find a niche that is just beginning to show promise, develop a hypothesis about where that niche will go in the future, and then build a business to test that hypothesis. Working toward a philosophy rather than a company's success keeps you humble. You're not devoting your soul to the business; because it's an experiment, you're devoting your life and your time to supporting your hypothesis, or your philosophy.
Sometimes experiments fail, forcing you to step back and reevaluate your hypothesis. Or change your business slightly. The most successful businesses start with theories for the future, especially in rapidly evolving fields like the technology industry, and they're not afraid to rapidly iterate when they find the company isn't working.
In the early 1980's, Mitch Kapor created the first real visual user interface for computer software. It was all text-based on the Apple II, of course, but he knew the way people interacted with computers was going to totally change; the experience of the computer, he felt, would have to vastly improve for laypeople to be able to use them. This was his theory. When he set out to build his first piece of software, he started with the philosophy that computers should be easy to use. In 1983, he built a word processor and spreadsheet application called Lotus 1-2-3. Within the first the year, Lotus had $53 million dollars in revenue. The next year, Lotus made $156 million.
Kapor had predicted the future of computing, built a piece of software to match his ideal of the future, and he created a ridiculously successful company. But his example is not the only one. I cannot think of an enormously successful company that did not start with a theory, or a philosophy, for the future of its market.
So two things.
I think to build a great company you need to have a well defined hypothesis based on a theory for a market's evolved future.
And I think the most effective way to enter that market is to build a company like a scientist testing the theory. As an experiment.