Lean startup

What it means

Lean startup is an approach to building a new business based on rapid experiments. Instead of lengthy planning and developing a perfect product, the company creates a minimum viable product as early as possible, measures customer response and then refines it or changes direction.

The approach was popularised by Eric Ries in his 2011 book The Lean Startup, which drew on lean manufacturing and Steve Blank's work on customer development. Its core is the build-measure-learn loop: every version of the product is an experiment designed to test a specific hypothesis.

The aim is to find out as cheaply and quickly as possible whether the product has a market. Most new businesses fail not because they cannot build the product but because nobody wants to buy it. Lean startup tries to uncover that truth before all the money runs out.

How it works in practice

First, the key hypotheses are stated: who the customer is, what problem they have and how much they would pay. Each is tested in the simplest possible way, for example with a landing page offer, interviews or a manually delivered service, before investing in automation.

Results are measured with metrics that reflect real behaviour, not downloads or likes. Cohort analysis shows whether customers come back, and A/B testing compares variants of an offer or price.

If a hypothesis fails, the company makes a pivot, a change of direction: a different customer segment, problem or business model. If it holds, the company continues and gradually scales. The goal is to reach product-market fit while consuming as little capital as possible.

What to watch out for

Lean startup does not mean doing things cheaply and carelessly. The minimum product must be good enough to test the hypothesis, otherwise a good idea may be rejected because of poor execution. In regulated fields such as finance or healthcare, even early versions must comply with the law.

Experiments must be evaluated honestly. It is easy for a team to look for confirmation of its idea in the data and ignore unwelcome results. It helps to decide in advance, before launching the experiment, which result will count as success and which as failure.