I’ve been reading my old friend Eric Ries’ excellent new book, The Lean Startup. If you do – or want to do – anything entrepreneurial, in any industry, you’d be well-served reading it, too.

The hard truth is, most new companies fail. We write that failure off to bad luck, to bad timing, to bad leadership. But, really, we have no idea. Every new business faces a barrage of unexpected challenges, wades through strings of ambiguous victories and nebulous defeats. In the end, some companies thrive, while others disappear. That’s how entrepreneurship is, we say; it’s more art than science.

But, usually, when we say something is ‘more art than science’, it later turns out to be a science we just don’t yet fully understand.

Consider the world of medicine, where doctors diagnose patients. A uniquely human task. Governed by instinct and gut feeling. Another art!

Really? Or is it a science that depends on reserves of implicit knowledge that great diagnosticians possess but that they simply can’t explicitly articulate? I’d argue the latter. Which is why, over time, after decades of careful analysis and modeling and increases in computing power and information storage, we’ve begun to explicitly match that implicit knowledge, and computer-aided diagnosis systems have started to empirically best what physicians can do themselves.

So, too, in the startup world, where it turns out that a rigorous reliance on data, a unified customer-development process, and an adherence to agile development methods may take some of the ‘magic’ and ‘genius’ out of startup ideation and execution, but hugely up the odds of actual, lasting success.

Which is all to say: buy the book.