Lean startup methodology: build-measure-learn and MVPs

The lean startup methodology is a way of building new products and businesses by testing ideas with real customers in small, fast experiments. Instead of following a long plan, a team builds a minimum viable product, measures how people respond and learns whether to keep going or change course.

This guide covers where the method came from, its five principles, the build-measure-learn loop, minimum viable products, validated learning, innovation accounting and the pivot. It ends with a simple way to run a first experiment.

Key takeaways

Lean startup treats a new product as a series of experiments, not a plan to execute.
The build-measure-learn loop turns ideas into products, products into data and data into decisions.
A minimum viable product is the smallest thing that tests your riskiest assumption.
Validated learning, backed by actionable metrics, is the measure of progress.
Each cycle ends with a choice: pivot to a new approach or persevere with the current one.

Where the lean startup came from

Eric Ries described the method in his 2011 book The Lean Startup, drawing on his own startups. He combined lean manufacturing ideas such as small batches and the five whys with Steve Blank's customer development process.

Blank's work, set out in The Four Steps to the Epiphany, argued that startups should search for a business model by talking to customers before scaling. Ries added a way to run that search with fast product experiments and clear metrics.

The five principles of lean startup

Ries sets out five principles in the book. They apply to new ventures inside large companies as much as to startups.

1
Entrepreneurs are everywhere
Anyone creating a new product under extreme uncertainty is doing entrepreneurial work, whatever the size of the organization.
2
Entrepreneurship is management
A startup is an institution, and it needs a form of management suited to uncertainty.
3
Validated learning
Startups exist to learn how to build a sustainable business, and that learning should be tested by experiments.
4
Build measure learn
The core activity is turning ideas into products, measuring how customers respond and learning whether to pivot or persevere.
5
Innovation accounting
Progress is measured with milestones and actionable metrics designed for uncertain, early work.

The build-measure-learn loop

The loop has three steps. Build turns an idea into something customers can react to, often an MVP. Measure collects data on what they actually do. Learn decides what the data means for the next step.

Ries suggests planning in reverse: start from what you need to learn, work out what to measure to learn it, then build only what is needed to get that measurement. The aim is to make each trip around the loop as short as possible.

Minimum viable product

A minimum viable product (MVP) is the version of a product that lets a team collect the most validated learning about customers with the least effort. It is an experiment, not a cut-down first release.

MVPs can be very small: a landing page that describes the product, a short video, a manual service run by hand behind a simple front end, or a single feature. The MVP tools guide lists software for building them, and A/B testing tools help measure the response.

Two documented cases show the range. Dropbox founder Drew Houston posted a demo video in 2008 that showed file syncing before the product was ready for the public, and Ries describes how sign-ups from its viewers helped confirm demand. Zappos began in 1999 with founder Nick Swinmurn photographing shoes in local stores and listing them online, then buying and shipping each pair by hand when an order came in. The minimum viable product guide covers both and more MVP types.

Validated learning and innovation accounting

Validated learning is learning that has been tested with real customer behavior, rather than opinions or surveys alone. It is how a lean startup shows progress before revenue arrives.

Innovation accounting sets a baseline with an MVP, tunes the product toward an ideal through experiments, then decides whether to pivot or persevere. It relies on actionable metrics, such as conversion rates by cohort, rather than vanity metrics such as total sign-ups, which rise whatever the product does.

What to validate: six kinds of assumption

Validation is not asking friends whether an idea sounds good. It is evidence from the people who would use, buy or approve the product, and the strongest evidence is behavior: sign-ups, payments, booked demos, repeat use and referrals. A business plan usually rests on six kinds of assumption, and each needs its own test.

1Problem

Do the target customers have the problem often enough, and does it hurt enough, for them to act?

2Customer

Is this the right group to serve first, or is another segment more urgent?

3Solution

Does the product actually help customers reach the outcome they want?

4Channel

Can the team reach customers through sales, marketing or partners it can afford?

5Revenue

Will customers pay, and does the price support the business?

6Retention

Do people come back after the first use, or does interest fade?

Test the one most likely to sink the idea first. If the business depends on a monthly subscription, test willingness to pay before building the full product. The lean canvas template puts these assumptions on one page.

Pivot or persevere

A pivot is a structured change of course to test a new hypothesis about the product, the customer or the business model, while keeping what the team has learned.

Ries describes ten types of pivot. Each one changes a single part of the plan and keeps the rest:

1
Zoom-in. One feature becomes the whole product, such as a reporting add-on that customers use more than the main tool.
2
Zoom-out. The product becomes one feature of a larger one, when a single feature is too narrow to stand alone.
3
Customer segment. The same product serves a different customer, such as moving from consumers to small firms.
4
Customer need. The team stays with its customers but solves a different, bigger problem it found while testing.
5
Platform. An application becomes a platform that others build on, or the other way around.
6
Business architecture. A switch between high-margin, low-volume sales and low-margin, high-volume sales.
7
Value capture. A change in how the business earns money, such as from licenses to subscriptions.
8
Engine of growth. A change in how the business grows, between viral, sticky and paid growth.
9
Channel. The same product reaches customers another way, such as online instead of through resellers.
10
Technology. The same result delivered with a different, often cheaper, technology.

Regular pivot-or-persevere meetings stop a team drifting on an idea that is not working.

Common lean startup misconceptions

Myth“Lean means cheap”

In practiceLean is about less waste and faster learning, not the lowest spend. A good test can cost real money if it answers the riskiest question.

Myth“An MVP can be unfinished”

In practiceAn MVP is small, but it still has to deliver enough value for the feedback to mean something.

Myth“Customer feedback is validation”

In practiceFeedback helps, but validation is stronger when it rests on what people do, not what they say.

Myth“A pivot means failure”

In practiceA pivot often means the team learned fast and moved toward a better opportunity.

Myth“Lean startup replaces strategy”

In practiceIt makes strategy evidence-based. The team still needs a vision and a clear question for each test.

Lean startup tools

The Lean Canvas, created by Ash Maurya, puts a business model on one page so its riskiest assumptions are easy to see. The lean canvas template guide shows a filled-in example box by box. Customer discovery tools help run and analyze interviews.

The lean startup tools hub covers software for each step of the loop, and actionable metrics and validated learning are explained in the glossary.

How to run your first experiment

Write down the assumption that would sink the idea if it were wrong. Usually it is whether customers have the problem at all, or whether they would pay to solve it.

Decide what result would prove or disprove it, and set the threshold before you start. Then build the smallest test that produces that result, run it with real prospects and compare the outcome with your threshold.

Record what you learned, decide to pivot or persevere, and start the next loop.

A first experiment from start to finish

Here is an illustrative experiment with made-up numbers. A team wants to sell automatic receipt matching to small accounting firms for $29 a month.

Experiment log (illustrative)

Riskiest assumption
Small accounting firms care enough about receipt matching to pay for it. If they don't, nothing else in the plan matters.
Threshold, set before the test
At least 4% of landing page visitors leave an email to join a paid pilot. The team picks 500 visitors as the sample, all from ads aimed at accounting firm owners.
Test
A one-page site that describes the product, shows the price and has a "Join the pilot" form. There is no product behind it yet.
Result
520 visitors and 11 emails, a rate of 2.1%. That is about half the threshold, so the assumption fails as written.
Learn
The team calls 8 of the 11 who signed up. Most say their bank feed already matches receipts for simple clients. The interest comes from firms with retail clients, who handle hundreds of small receipts a month.
Decision
A customer segment pivot. The next test targets bookkeepers who serve shops and restaurants, with the same 4% threshold. The team writes down the new assumption and starts the loop again.

When lean startup fits

Use lean startup when there is real uncertainty about the customer, the problem, the product, the channel or the business model. A simple rule: the higher the cost of being wrong, the more a test is worth. If building the wrong feature would waste months, test it first. If pricing is unclear, run a pricing test before assuming what customers will pay.

The method is not only for software startups. These illustrative uses show the range.

Software

A landing page and a waitlist test demand before the platform is built.

Services

A manual pilot shows which package customers value most.

Ecommerce

A small product drop tests demand before a large inventory order.

Inside a company

A prototype of a new internal workflow is tried with one team before a company-wide rollout.

Nonprofits

A program runs with one community group before it expands to more locations.

It matters less when the work is predictable or demand is already proven, though parts of it, such as testing messages or prototyping a workflow, still help.

Limits and criticisms

Lean startup works best when experiments are cheap and quick. It is harder to apply in four common cases.

Hardware and physical products

A prototype can cost months and real money, so each loop is slow. Teams lean more on interviews, pre-orders and landing pages before they build.

Regulated products

Medical devices, financial services and similar products can't ship a rough version to the public. Experiments move to the parts that are safe to test, such as demand, pricing and the sales process.

Long B2B sales cycles

When a deal takes nine months, a team can't wait for revenue to learn. It needs earlier signals it trusts, such as signed letters of intent or paid pilots.

Brand risk

An established company can harm its name with a poor public MVP. Many test under a separate brand or with a small invited group.

Wider criticism

Some critics also argue that fast experiments favor small changes over bold ideas, and that a team can keep pivoting without ever committing. A clear goal and a limit on the number of loops help with both.

Frequently asked questions

What is the lean startup methodology in simple terms?

A way to build a new product under uncertainty by testing the riskiest assumptions with small experiments, learning from real customer behavior and then deciding to persevere or pivot.

Who created the lean startup?

Eric Ries, in his 2011 book The Lean Startup. He combined ideas from lean manufacturing with Steve Blank's customer development process.

What is a minimum viable product?

The smallest version of a product that lets a team test a key assumption with real customers. It can be a landing page, a prototype or a manual service. The MVP guide covers the common types.

What is a pivot in lean startup?

A change in strategy based on what experiments showed, such as a new customer segment or a different problem to solve, while keeping what the team has learned.

Does lean startup work in large companies?

Yes, for new products and markets where the answer is not known yet. Ries covers this in The Startup Way (2017). For improving an existing, well-understood process, lean or Six Sigma usually fit better.

Guides on this topic

Sources

Books and papers show their publication year. Web pages show the month we last checked them.

1.Eric Ries, The Lean Startup (Crown Business, 2011)
2.Steve Blank, The Four Steps to the Epiphany (Cafepress, 2005)
3.Ash Maurya, Running Lean: Iterate from Plan A to a Plan That Works, 2nd edition (O'Reilly, 2012)
4.Alexander Osterwalder and Yves Pigneur, Business Model Generation (Wiley, 2010)
5.Jeff Gothelf and Josh Seiden, Lean UX: Applying Lean Principles to Improve User Experience (O'Reilly, 2013)

Lean startup methodology: principles, build-measure-learn and MVPs