Software Growth

Lean startup

The lean startup method, from Eric Ries, treats a new company as a series of experiments: build, measure and learn fast to find a model that works.

Lean startup is a way of building a new company under uncertainty. Eric Ries popularized it in his 2011 book The Lean Startup. The core claim is that a startup is not a smaller version of a big company. It is an experiment to find a business model that works, so it should be managed by learning, not by following a plan written on day one.

The five principles

The official Lean Startup site lists five:

  1. Entrepreneurs are everywhere.
  2. Entrepreneurship is management.
  3. Validated learning: test each part of your vision with experiments.
  4. Innovation accounting: track progress with metrics fit for startups.
  5. Build-measure-learn: the loop that turns ideas into products as fast as possible.

Build, measure, learn

You build the smallest thing that tests an idea, a minimum viable product. You measure how real customers respond. You learn whether to keep going or change direction. The goal is to finish the loop quickly, because the faster you learn, the less you spend on ideas that fail.

Ries ties this to the decision to pivot or persevere. If the data shows your assumption was wrong, you make a structural course correction. If it shows progress, you carry on.

Where it came from

Ries built on Steve Blank's work on customer development, which told founders to get out of the building and test their business model with customers. Lean startup combines that with agile engineering and gave the whole thing a name and a vocabulary.

Example

You think freelancers will pay $15 per month for a tool that chases late invoices. Before building it, you put up a landing page with a price and a "start trial" button, then send 300 visitors from a freelancer forum. If 12 click through to a payment step, that is a 4 percent rate. 12 divided by 300 equals 0.04. That is a sign worth building for. If only 1 clicks, the idea needs work. You spent a weekend, not six months.

Lean startup when building is cheap

AI coding tools have made the build step of the loop much cheaper. Jon Yongfook, founder of Bannerbear, says features that once took him a week now take an afternoon. That doesn't make lean startup obsolete. It means the slow, expensive part is learning whether anyone will pay. Rob Walling's new book has the subtitle Stop Building SaaS Nobody Wants in the Age of AI and focuses on validating before you write code.

Walling also points out that a long sales cycle slows the loop. In episode 847 (from 26:33), he advises a founder with a new product to prefer a market where deals close sooner, because you want to "optimize your learning speed." A 12-month enterprise cycle gives you one data point a year.

Treat the method itself as an experiment

There is context inside which the framework works, and outside which it doesn’t work.

Jason Cohen / “You didn’t do it right” (October 11, 2026)

Cohen's essay is about OKRs, Scrum and jobs-to-be-done. The same point applies to lean startup: if the build-measure-learn ritual isn't producing decisions, "you didn't do it right" is not the only possible explanation. He suggests trying a framework with one team for a quarter, keeping what works and adjusting the rest.

Getting attached to an outcome doesn’t increase its likelihood.

Seth Godin / The half-life of disappointment (October 27, 2026)

Godin's short post applies to the pivot-or-persevere decision. Each failed experiment is easier to act on if you were testing an assumption, not defending a plan.

Common mistakes

  • Calling anything half-built an MVP. It should test one specific question.
  • Measuring vanity numbers. Signups and page views feel good. Paid conversions and retention tell the truth.
  • Skipping customers. Experiments without conversations produce numbers with no explanation.
  • Never deciding. The loop only works if you act on what you learn.

For bootstrapped founders

Lean thinking suits bootstrappers naturally, since you cannot afford big bets. Run cheap experiments, charge early and reach product-market fit before you scale anything. Treat your cash as the experiment budget.

Sources

Back to the SaaS glossary