Harvard Business Publishing

Last winter, when we all thought the economy had really taken a nosedive, I made the case that bad times could be the best of times to start a company. Well, over the last eight months, economic conditions have gone from bad to worse--but the startup boom shows few signs of slowing down.

Want proof? Spend time with Paul Graham, who's spent the last three years helping young entrepreneurs launch the companies of their dreams. Graham and his cofounders run a cool outfit called Y Combinator, with operations in two hotbeds of entrepreneurship: Silicon Valley and Cambridge, Massachusetts.

Graham and his colleagues provide both seed funding and hands-on advice to startups. Y Combinator invests a little money (rarely more than $20,000) and takes a small equity stake (an average of 6 percent or so.) And it funds its companies in "batches." Every summer, entrepreneurs from around the world apply for the chance to spend three months in Cambridge. Graham and his colleagues help people refine their ideas, critique their prototypes, and teach them how to present to venture capitalists and angels--all of which leads to a "Demo Day" for potential investors. Every winter, the same immersion experience takes place in Mountain View, California. There are two Demo Days in the Valley, given the sheer number of eager venture capitalists.

This summer, Graham's got 60 founders, representing 22 companies, at work on their ideas. All told, Y Combinator has invested in 102 startups--including a bunch that have attracted lots of money and attention (Loopt, for example) or been acquired by big-time players (Reddit, for example). In other words, Graham has seen it all when it comes to startups (including starting and selling a company of his own back in the first Internet boom.)

So how would Paul Graham advise young entrepreneurs who want to be "practically radical" in their approach to startups? He offers a few simple (and proven) principles that determine which startups work and which fail:

First, he says, "make something people want." It sounds obvious, but young entrepreneurs often fall in love with what technology can do as opposed to what customers need. One of the questions on the application to Y Combinator makes the point well: "What are people forced to do now because what you plan to do doesn't exist yet?"

Second, entrepreneurs have to "be willing to let their ideas change." This sounds strange: Aren't great startups built around a great idea? Yes, but a great idea isn't always the original idea. Countless times during the Y Combinator experience, startups have made dramatic changes not just in strategies and tactics, but in the very essence of what they do. Amazingly, Graham told me, the founders of Reddit came to Y Combinator with a plan to help people order fast food on their cell phones. When everyone agreed it wasn't exactly a killer idea, the desperate search for a new idea led to the launch of their successful company.

Third, and this also sounds strange, Graham tells startups: "Don't worry too much about money." That's what's really different about this second wave of Web-enabled startups: It's become so cheap to buy equipment, reach customers, and generate buzz on the Web, that the power of investors and venture capitalists is on the wane. "It's so much easier to get the money you need than to make something great," he says. Or, as Graham points out, today, unlike back in the mid-90s, you've got the MBAs working for the technologists, rather than the other way around.

Finally, Graham urges his company founders to always "be benevolent" in terms of how they do business--to act in the long-term best interests of customers, as opposed to the short-term best interests of themselves. In other words, the most important rule for starting a company is the Golden Rule--or, in the updated version in Google's strategic bible, "Don't be evil."

Why is benevolence powerful? First, it keeps morale and energy high. In an age of constant disruption and realignment, employees want to be the "good guys" in their field--so it makes sense for companies to act that way.

Second, more than ever, successful companies require the active participation of customers, suppliers, and industry enthusiasts. "If you're benevolent," Graham says, "people will rally around you" with ideas, improvements, and word-of-mouth marketing.

Finally, benevolence helps founders to be more decisive. If you make every decision based on "doing whatever is best for your users," it's that much easier to make decisions.

The most important decision, of course, is starting a company in the first place. And once you do, be sure to use these two resources: Paul's must-see talk to a gathering of entrepreneurs at Stanford called Startup School, and his must read-essay "How to Start a Startup."

Here's hoping you decide to get started--and that you succeed.