For most of us, the phrase "management science" conjures up a decidedly non-entrepreneurial image, and for good reason. The preeminent management science, for much of the twentieth century was general management, pioneered by twentieth century giants like Peter Drucker and Alfred P. Sloan. As an entrepreneur, I often saw the best practices of general management fail startups. Applied out of context, they were dangerous.

This conflict has played out in many companies that I've worked with. My most recent startup created a marketplace for customers to buy and sell virtual goods for their 3D avatar. So you can imagine how I expected some skepticism when pitching ideas about technology innovation to, say, the U.S. Army. This was my first surprise: they understood that innovation needs to be understood at the level of principles, not just tactics.

Some tactics, like Continuous Deployment, are controversial, even among people with my background. But the really vexing questions came from entrepreneurs who inundated me with questions from other backgrounds. Could these techniques be used by enterprise software companies? Hardware manufacturers? Biotech startups?

Answering those questions requires understanding the principles of entrepreneurship in a rigorous way. I began to call this new methodology "the lean startup." It is a rigorous application of lean principles to the problem of new product innovation.

Inspired by the lean manufacturing revolution (and excellent books like Lean Thinking), I started with a first fundamental question: in a startup, what activities are value-creating and which are waste? Usually, new projects are measured and held accountable to milestones and deadlines. When a project is on track, on time, and on budget, our intuition is that it is being well managed. This intuition is dead wrong.

Most startups fail because they are building something that nobody wants. Enamored with a new technology or a radical new product, many entrepreneurs never find a set of customers who will buy it. Each new feature added to such a product is actually wasted effort, even if it's done on-time and on-budget. In product that nobody wants, all the features get thrown away.

This problem vexes venture-backed founders and general managers alike. Anyone who is tasked with creating disruptive innovation will find it familiar. In short, for an innovation on the wrong course, executing well doesn't increase the odds of success. What's needed, I realized, is a new definition of progress in a startup, one that recognizes that a startup's primary mission is to penetrate the fog of the unknown and find out what their customers ultimately will want and accept. I call this unit of progress "validated learning about customers." Because learning is the goal, the progress that startups make requires a very different kind of management.

This led me to a second surprise. As I was traveling the country talking to people about the lean startup, I kept meeting people in the audience who seemed out of place. They weren't what we usually think of as entrepreneurs. They were general managers, most working in very large companies, who were tasked with creating new product innovations. They were smart, well-read, and understood Clay Christenson's Innovator's Solution cold.

They were adept at organizational politics: they knew how to form autonomous divisions with separate P&Ls, and could shield controversial or disruptive teams from meddling. And — my biggest surprise yet — they were visionaries. Like the startup founders I have worked with for years, they could see the future of their industries, and were prepared to take bold risks to seek out new and innovative solutions to the problems their companies would face.

They found the lean startup helpful. In fact, they considered it obvious that these principles — rapid iteration, a focus on validated learning, and reducing cycle time from idea to learning — applied to their situation. Here's why. In my quest to put the practice of entrepreneurship on a rigorous footing, I had begun with this definition of a startup:

A startup is a human institution designed to create a new product or service under conditions of extreme uncertainty.

Originally, I used this definition to help entrepreneurs understand why the "best practices" of general management so often failed them. They can't thrive when transplanted into the soil of extreme uncertainty that is the province of startups. But this definition is surprising: it doesn't say anything about the size of the company, sector of the economy, or industry. Looking past my old preconceptions, I realized that entrepreneurs are everywhere: in government, in the occasional garage, and yes, in the traditional organization.

All entrepreneurs face the same fundamental challenges:

• How do we know if we're making progress?
• How do we know if customers will want the product we are building?
• And, if they do, how do we know what kind of value we can create with it?

But because every startup also strives to become an institution, answering these questions requires more than just disciplined thinking at the whiteboard. It requires the coordination of many different people, working in concert to answer them. In other words, it requires management.

The management challenges presented by entrepreneurship are different. For example, how do we hold an entrepreneurial team accountable for making progress? This is a core challenge faced by venture capitalists and corporate CFOs alike. A team might be making progress against their plan, might be hitting milestones, might even be earning revenue. But none of those artifacts constitute progress. Learning is progress. In short, it's demonstrating the ability to use what we've learned to change customer behavior, with key metrics as the common source of evidence.

To measure learning requires a shift of attention from the output of models to their inputs. I'll show you how to apply this to entrepreneurial situations in my next post.