I am listening to Rework, the new book by Jason Fried and David Heinemeier Hansson of 37 Signals. I agree with much of what they say about startups: focus on the idea and the product, not marketing or building an infrastructure; don’t write a long business plan, or even one at all; avoid expensive consultants – do the PR yourself; minimize hiring and other expenses and try to not take financing early on; don’t chase an image of what you think a successful company should be like; no matter how different it seems, figure out what you really need and focus on just that. A friend of mine started what eventually became a large company. His customers were large enterprises, where image can seem important. In the early years, though, he didn’t even print business cards. He just said he didn’t have any on him whenever he made a sales call. Didn’t hurt in the slightest. What mattered was he had a compelling message about his service and team and then followed through and provided value at a reasonable price.

I am still in the early part of the book, but one thing that bothers me is the bias towards small, private, bootstrapped companies. Because 37 Signals’ approach is similar to ours in many ways, I found much to identify with. But I don’t think there is anything inherently superior about such companies. I congratulate Jason and David for figuring out what they wanted and finding a narrow path to get there. For others, a very different kind of company will be appropriate. My advice for starting a company is first figure out your life goals and think about how your company can support them. Then decide what business you want to be in. Then find a path to get there. If you want to build a new car company, or a new type of computer hardware, or start a biotech to cure a disease, or come up with a better CRM package, the models will be very, very different than for many software companies. Virtual or not, capital or not and many other questions can only be answered once you understand your goals, your market and your product.

I’ll write more later about how this relates to starting Ipswitch.