The Lean Startup

Vision: a startup is an experiment

Eric Ries built the method in this book from his own failures and from his work at IMVU, a company that made 3D avatars and instant messaging. He borrows from the Toyota Production System and lean manufacturing, and from Steve Blank's work on customer development. The book has three parts: Vision, Steer and Accelerate.

Ries defines a startup as a human institution that tries to create a new product or service under conditions of extreme uncertainty. By this definition, a team inside a large company can be a startup. So can a nonprofit or a government office. The size of the group does not matter. The uncertainty does.

Most startups fail, Ries argues, not because they build badly but because they build the wrong thing. Traditional management assumes you know the customer and the product. You write a plan and execute it. Under extreme uncertainty, that plan is a guess. Executing a guess perfectly is still a waste.

So Ries says the real unit of progress in a startup is validated learning. This is learning backed by real data from real customers, not opinions or forecasts. At IMVU, the team spent months building an add-on that let users bring their existing instant messaging friends into the product. Customers did not want it. They wanted to meet new people. The team could have learned this in days, with a crude test. Ries asks a hard question: which of our efforts created value, and which were waste? Anything that does not help you learn what customers want is waste.

Steer: the Build-Measure-Learn loop and the MVP

The core of the method is the Build-Measure-Learn feedback loop. You turn ideas into products, measure how customers respond, and learn whether to keep going or change course. The goal is to get through this loop as fast as possible.

Ries notes that you plan the loop in reverse. First decide what you need to learn. Then decide what you must measure to learn it. Then decide the smallest thing you can build to get that measurement.

The loop starts with leap-of-faith assumptions. These are the beliefs your whole plan depends on. Ries names the two most important:

• The value hypothesis: does the product actually deliver value to customers once they use it?
• The growth hypothesis: how will new customers discover the product?

Before you build, you should talk to customers directly. Ries uses the Toyota phrase genchi gembutsu, which means "go and see for yourself."

The minimum viable product (MVP) is the version of a product that lets a team run a full turn of the loop with the least effort. It is not a smaller version of the final product. It is a tool for learning. Ries gives several types:

• Video MVP. Drew Houston of Dropbox made a short video that showed how file syncing would work. The beta waiting list jumped from about 5,000 to 75,000 people almost overnight. The team learned there was demand before the hard technical work was done.
• Concierge MVP. The founders of Food on the Table, a meal-planning service, started with one customer. They did the planning and shopping lists by hand and delivered them in person. They automated only after they knew what customers valued.
• Wizard of Oz MVP. The customer believes they use a working product, but people do the work behind the scenes. Nick Swinmurn tested the idea for Zappos by taking photos of shoes in local stores and posting them online. When someone ordered, he bought the shoes at the store and shipped them.

Ries addresses the fears that stop teams from shipping an MVP. Teams worry about legal issues, about competitors stealing the idea and about damage to the brand. He argues that most ideas are not stolen, that you can launch under a different brand name, and that the real risk is building something nobody wants. If the MVP fails, do not give up. Treat it as data.

Measure: innovation accounting

To know whether you are making progress, you need a way to measure it. Ries calls this innovation accounting. It has three steps:

1. Use an MVP to establish a baseline: real numbers on where the company stands today.
2. Tune the engine: make changes that try to move the baseline toward the ideal.
3. Decide to pivot or persevere based on the results.

Most startups track the wrong numbers. Ries calls them vanity metrics: totals like gross registered users or total page views that go up and to the right and make people feel good, but do not show whether a change worked. He prefers actionable metrics, which show clear cause and effect. He uses tools such as cohort analysis, which tracks each group of new customers separately, and split testing, which shows different versions to different customers at the same time.

He also gives three tests for good reports, the three A's: actionable, accessible and auditable. People must be able to understand the numbers and trust where they come from.

Ries describes Grockit, an online test-prep company. Its team kept shipping features, but cohort data and split tests showed that a change it believed in did not alter customer behavior. That insight saved the team from building more of the wrong thing.

Pivot or persevere

At regular intervals, a startup must make the hardest decision: keep going on the current strategy, or pivot. A pivot is a structured change of course that tests a new fundamental hypothesis about the product, strategy or engine of growth. It keeps one foot in what the team has learned.

Ries lists the kinds of pivot:

• Zoom-in pivot: one feature becomes the whole product.
• Zoom-out pivot: the whole product becomes one feature of something bigger.
• Customer segment pivot: the product solves a real problem, but for a different customer.
• Customer need pivot: the customer has a more important problem to solve.
• Platform pivot: a change from an application to a platform, or the reverse.
• Business architecture pivot: a switch between high margin and low volume, and low margin and high volume.
• Value capture pivot: a change in how the company makes money.
• Engine of growth pivot: a change in the main way the company grows.
• Channel pivot: a change in how the product reaches customers.
• Technology pivot: a new technology to reach the same result.

Ries tells the story of Votizen, a civic platform, which went through several pivots before it found a model that worked. He also warns about the land of the living dead: startups that are neither growing nor failing. They move just enough to keep going, but not enough to succeed. Regular "pivot or persevere" meetings force the team to face the evidence.

Accelerate: small batches and engines of growth

The last part of the book is about how to move through the loop faster as the company grows.

Ries starts with small batches. He describes stuffing envelopes for a newsletter. Most people fold all the letters first, then stuff them all, then seal them. Working one envelope at a time is often faster overall, and it reveals problems early. In software, small batches mean shipping small changes often. Ries describes continuous deployment at IMVU, where code went to customers many times a day, with safeguards that caught bad changes.

Next come the engines of growth. Ries names three:

• The sticky engine: growth comes from keeping customers. Watch the churn rate. Growth happens when new customers arrive faster than old ones leave.
• The viral engine: growth comes from existing users bringing in new ones as a side effect of use. Watch the viral coefficient. If each user brings in more than one new user, the product grows on its own.
• The paid engine: growth comes from spending money to acquire customers. Watch the gap between what a customer is worth over time and what it costs to acquire one.

A startup should usually focus on one engine at a time.

Adapt and innovate

As a company grows, it must build habits that keep it learning. Ries uses the Five Whys, a technique from Toyota. When something goes wrong, ask "why" five times to find the root cause. Then make a fix at each level, with the investment in each fix proportional to the size of the problem. Small problems get small fixes. This keeps the team from overbuilding process.

Ries also explains how large companies can innovate. They should protect internal startup teams with an innovation sandbox: a space where teams can run experiments with clear limits. The teams need secured resources, independent authority to act, and a personal stake in the result. He cites Intuit's SnapTax, a mobile tax product built by a small internal team, as an example.

The book closes with a call to stop wasting people's time. Ries wants management to treat entrepreneurship as a discipline that can be taught, not as a matter of luck.

What the critics say

Some critics say the method fits software and web products better than hardware, health care or deep science, where an MVP can be costly or unsafe. Others say the term MVP has been misused to justify shipping low-quality products, and that a strong focus on metrics can push teams toward small, safe improvements instead of bold visions. Ries himself frames the method as a way to test a vision, not to replace one.

Who should read it

This book is for founders, product managers and anyone who builds new products, inside or outside a large company. It is most useful for teams that are about to spend months building before talking to a single customer.