The Lean Startup
Eric Ries
The Lean Startup argues that a startup is not a small company but a search, under extreme uncertainty, for a business that works, and that the only progress that counts is validated learning from real customers. Eric Ries gives the search a method: state your riskiest assumption, test it with the smallest possible product, measure by cohort, and decide on schedule whether to pivot or persevere.
You have a plan, a roadmap and a product half built, and none of it has touched a customer. The Lean Startup's diagnosis is that the plan is the problem: it assumes you know who the customer is and what they want, which is precisely what a startup does not yet know. Ries replaces the plan with a loop, and the loop is why some founders learn in a month what others discover after a year and all their money.
The book in essence
Ries defines a startup as a human institution designed to create a new product or service under conditions of extreme uncertainty. Nothing in the definition says small or technology; what it says is uncertainty, and uncertainty is why the tools of the established company fail founders. So the unit of progress changes from revenue and features to validated learning, and the method for producing it is the Build-Measure-Learn loop, planned backwards from what must be learned. The minimum viable product is the smallest experiment that runs one turn of the loop; innovation accounting is how you read the results when revenue is too small to mean anything; and the pivot-or-persevere meeting is where the results become a decision.
Ries wrote the book after co-founding IMVU, where the team watched total registrations climb for months while every new cohort behaved exactly like the last, and after seeing the same waste across the companies he advised. The book's tone is that of an engineer describing a process, not a guru describing a mindset, which is why its vocabulary, MVP, pivot, vanity metrics, became the working language of startups worldwide.
Who it's for
- First-time founders with a plan, a spreadsheet and no customer contact yet
- Product people running a new bet inside an established company, where the old planning tools are the default
- Anyone who has spent months building something and is afraid to find out whether anyone wants it
Who it's not for
- — Founders in businesses with no real uncertainty, such as a franchise or a well-understood local service, will find the method heavier than the problem
- — Readers looking for a book about fundraising or company-building after product-market fit; this is about finding it
Key lessons
01
Validated learning is the only progress
Shipping a feature nobody uses is motion, not progress. Ries insists that a startup's progress is the demonstrated change in what it knows about customers, proven by an experiment. The practical consequence is that every week should end with a sentence beginning with what we learned from a customer that changed what we do next; a week that cannot produce that sentence was activity.
A team in Izmir spent four months on a meal-planning app and launched to silence. Their second attempt began with one belief, that parents would pay for weekly plans, tested with twenty families and hand-written plans. Eleven paid, and that number was the company's first real progress.
02
The MVP is an experiment, not a small product
Founders hear minimum and build a cheaper version of the full product, which tests nothing in particular. Ries's MVP is chosen backwards from the riskiest assumption: what is the least work that puts that belief in front of a customer? A video, a landing page, a service delivered by hand for ten customers all qualify, and the embarrassment of shipping them is the price of learning early.
A laundry pick-up startup in Eskişehir launched with a Google Form, a WhatsApp number and a friend's car. Forty orders in two weeks proved demand; the complaints about timing told them what the app, built two months later, had to get right.
03
Numbers that only go up are not measuring you
Total sign-ups, downloads and page views rise as long as you keep spending, so they cannot tell you whether the last change worked. Ries's test is whether a number could fall because of a decision you made. Cohorts make metrics actionable: compare what this week's users did with what last week's did before the change, and let the trend, not the total, decide.
A Turkish language app celebrated 200,000 downloads. A cohort chart showed four of every hundred installers still active a month later, unchanged across eleven releases. The downloads were real; the product had not improved.
04
Schedule the pivot so you do not drift into it
Founders delay pivots because the numbers are ambiguous enough to permit hope and their identity is tied to the idea. Ries's cure is a regular pivot-or-persevere meeting with cohort data and one written decision. A pivot keeps one foot in what has been validated and changes one thing; his catalogue of pivot types turns a frightening decision into a menu.
Two co-founders in Izmir met on the first Monday of each month under the heading Pivot or persevere. In the sixth month the retention cohort had not moved in three meetings, and the agenda forced the conversation they had been avoiding. The pivot took two weeks.
Try this today
Write your business as a numbered list of beliefs about customers, circle the one that kills the company if it is wrong, and design the smallest experiment that could prove it wrong this week.
Quick check
According to Ries, what is a startup's unit of progress?
What makes a metric actionable rather than vanity?
Selected quotes
“The only way to win is to learn faster than anyone else.”
“Success is not delivering a feature; success is learning how to solve the customer's problem.”
“If you cannot fail, you cannot learn.”