It turns out that the lean startup consists of three components; one is business model design, two is customer development, and three is agile engineering. And these three components help startups become incredibly efficient, reducing the amount of time to get to first product and reducing the amount of cash necessary to build that product. So in a lean startup we really know something quite different than a large company. In a lean startup we know that you just don’t execute per plan, you’re actually going to be learning and discovering from customers outside your office. We kind of know that there are no facts inside your building, so you’re going to get the heck outside. We also know that you’re not smarter than the collective intelligence of your potential customers.

So what we’re going to do is put together a process that will allow you to rapidly learn by getting out of the building and testing some of your key ideas and key features of the product as quickly and as rapidly as we can. And what you’re going to find is a good number of your initial hypotheses are just simply wrong. And the way we’re going to do this is we’re going to start with writing down all the hypotheses. And the common framework we use is Alexander Osterwalder’s Business Model Canvas. This nine‑box diagram puts all the hypotheses about your business, who are the customers, what are you building for them, what’s the channel, how are you going to reach them, how are you going to make money, and puts them in a single page so it’s going to provide a framework and then a score card for us to see how we’re doing. But unlike just doing a strategy in a large company you’re going to take these hypotheses and go outside the building and actually test them. Not test them randomly, but test them with a formal process called customer development.
What we’re going to do is design experiments, run the experiments outside the building, get the data and look for insights. And in fact, the way we’re going to actually get insights is by showing customers either powerpoints, wire frames, clay models, prototypes, at any stage we can, to actually gather as much information as we can about their feedback.

And one of the things we’re going to find out is that there are two unique things that we do in a lean startup. One is when our hypotheses turn out to be incorrect instead of firing people we actually get to change those hypotheses. And these are called pivots. A pivot is defined as a substantive change to one or more of the business model canvas components. If it’s a minor change we call it an iteration. And so in a normal company, or an old‑fashioned startup we fire the VP of Sales if their assumptions were wrong. Here what we’re doing is firing the plan. Not only does it work for startups it actually works for large companies as well.

We work for General Electric, one of the smartest, big companies in the world. And GE was about to commercialize some of their high temperature sulfur battery technology. And they wanted to offer these batteries to other companies for industrial applications. And they had done a great market research report and had pages of here’s where the product should go. And the CEO was going to give the General Manager, Prescott Logan, $100 million to just go build the factories and execute and hire the staff and whatever. We have the technology and here’s the report. But Prescott luckily said, you know, this doesn’t feel right. This doesn’t feel like an existing business, this feels like a startup and we ought to check our hypotheses. And so Prescott Logan and his customer development team flew around to almost every continent, in fact, every continent in the world, including Antarctica and discovered that the requirements actually differed by industry segments. Substantially enough that if they would have had a standard factory making standard product they would have sold none of them because different customers required different packaging, different cooling, different connectors. And none of this was visible from inside the building. After 18 months they figured out who the customers were, what they wanted and then built the factory. And now a couple years later they have backlogs for the next three years.

And so what we now know is that no business plan survives first contact with customers. That means we wouldn’t know whether we were right or wrong until after we shipped the product and customers gave us feedback. And by then we’ve burned lots of time, lots of money, and potentially you’re out of business. But imagine if you had a process that allowed you to quickly iterate and incrementally build the product a piece at a time for each increment, learning and discovering whether you’re on the right track. What’s the right track? Do you have the right customers for the right features, which we call product market fit? Do you the right distribution channel? Do you have the right revenue model? Right pricing? Partners? Do you understand your costs? We have a process now to test each one of these incrementally, cheaply and quickly. And that’s what the whole Lean Startup Model is about.