The Lean Startup

"The only way to win is to learn faster than anyone else."

— Eric Ries

Eric Ries wrote the book after living through a failure and a success. In his first startup he spent years building a perfect product nobody used. At IMVU, the second, he launched in weeks a version full of bugs, put it in front of real customers and discovered that almost everything the team assumed about them was wrong. From there comes his definition: a startup is an institution designed to create a new product under conditions of extreme uncertainty, and its main job is not to execute a plan, but to learn which plan is worth executing.

On learning, his lessons are three:

• Validated learning is the unit of progress. What counts is not lines of code or features shipped, but the hypotheses about the customer you confirmed or discarded with real data.
• Every business idea boils down to two bets. The value hypothesis, that the customer wants this, and the growth hypothesis, that there is a repeatable way to attract more customers. The value one is tested first.
• Build, measure, learn. You build the minimum that allows measurement, measure the customer's reaction and learn in order to decide the next cycle. The speed that matters is that of completing the cycle, not of building.

The central tool is the minimum viable product, which Ries defines as the simplest version that allows you to complete a cycle with the least effort. It can be a demo video, like the one Dropbox made before the product existed, which multiplied its waiting list overnight, or a concierge MVP, in which the founder serves each customer by hand and does what the software will later do. The MVP is not a bad product, but the cheapest question you can ask the market.

On measurement, the book separates two kinds of metrics:

• Vanity. Registered users, downloads or visits, which go up even when the business doesn't improve.
• Actionable. They compare cohorts, groups of customers who came in during different weeks, and show whether each cohort behaves better than the previous one.

With them you decide the hardest thing: pivot or persevere. The pivot is a structured change of strategy that keeps what was learned, not a lurch: focusing on a feature that works, changing customer segment, channel or revenue model. Ries proposes a periodic meeting with that single question, because the most common mistake isn't pivoting badly, but not pivoting out of attachment. And to get to the root cause of any problem, he recommends Toyota's five-whys technique.

The destination of all these turns of the cycle is product-market fit, a concept Ries takes from Marc Andreessen: the moment when the product clearly solves a market's need and customers start asking for it on their own. Before getting there, accelerating growth only multiplies waste. Afterward, Ries proposes choosing one of three engines of growth:

• Sticky. Customers stay and buy again; what matters is that fewer leave than arrive.
• Viral. Each customer brings others by using the product.
• Paid. Each customer leaves more than it costs to attract them (LTV exceeds CAC), and the difference is reinvested in attracting more.

A concept to master is validated learning, because it redefines what moving forward means. A team that built three features without learning anything didn't move forward; one that discarded its main idea in two weeks, with twenty interviews, moved forward with great efficiency.

How to apply it? Write your value hypothesis in one sentence that can turn out true or false. Define the cheapest experiment that tests it: a video, a sign-up page or twenty calls. Set in advance what result would make you persevere and which would make you pivot. And on your dashboard, swap the most visible vanity metric for a cohort one.

As a founder it will help you invest your time and capital in learning and not in building what nobody wants, and in launching any new line of business once your company is established, where the temptation to plan instead of learn is even greater. It will also help you understand and practically measure the industry's key metrics, like Customer Lifetime Value (LTV) and Customer Acquisition Cost (CAC).