Lean startup

The lean startup methodology emerged from Eric Ries's experiences as a software engineer and entrepreneur, particularly his work on the failed startup IMVU. Ries drew on lean manufacturing techniques, especially the Toyota Production System, and adapted them to the context of innovation under uncertainty. The central idea is to treat a startup as an experiment rather than a static plan, using a build-measure-learn feedback loop to test hypotheses about the product and market.

Key concepts include the minimum viable product (MVP), which is the smallest version of a product that allows the team to collect validated learning with the least effort, and the pivot, a structured course correction to test a new fundamental hypothesis. The methodology emphasizes avoiding waste by not building features that customers do not want, and it encourages frequent releases and direct customer feedback.

In practice, lean startup involves defining a set of hypotheses about the business model, often using a tool like the business model canvas, and then designing experiments to test those hypotheses. The build-measure-learn loop is the engine of this process: teams build an MVP, measure how customers respond using actionable metrics (such as cohort analysis and split testing), and then decide whether to persevere or pivot.

Innovation accounting is another pillar, requiring startups to track progress using a framework of learning milestones rather than traditional financial metrics. This approach is particularly suited to high-uncertainty environments where customer needs are not well understood. The methodology has been widely adopted in software startups, but it has also been applied to corporate innovation, government projects, and social enterprises.

Critics argue that the lean startup methodology may not be applicable to all types of ventures, especially those requiring significant upfront investment in research and development, such as biotechnology or hardware. The emphasis on rapid iteration and customer feedback can be at odds with long-term vision or breakthrough innovation that requires patience and secrecy.

Some scholars have noted that the concept of the MVP is often misunderstood, leading to the release of low-quality products that damage brand reputation. Additionally, the methodology assumes that customer feedback is reliable and that customers can articulate their needs, which is not always the case. Despite these criticisms, the lean startup has become a dominant paradigm in entrepreneurship education and practice, influencing how many organizations approach innovation.

Beyond the well-known concepts, the lean startup has several lesser-known dimensions. For instance, Eric Ries has applied the methodology to government through the U.S. Department of Defense's Defense Innovation Board, where he served as an advisor. The methodology has also been adapted for use in large corporations under the term 'lean innovation' or 'corporate entrepreneurship.'

A notable edge case is the application of lean startup principles to social ventures, where the 'customer' may be a beneficiary rather than a paying user, complicating the measurement of value. Additionally, the concept of 'validated learning' has been critiqued for being difficult to operationalize, leading to debates about what constitutes valid evidence. The lean startup has also inspired related movements such as 'lean analytics' and 'lean UX,' which apply similar principles to data analysis and user experience design.

MVP

Minimum Viable Product: the smallest version of a product that allows a team to collect validated learning with the least effort.

Pivot

A structured course correction to test a new fundamental hypothesis about the product or business model.

Build-Measure-Learn

The core feedback loop of the lean startup methodology, involving building an MVP, measuring customer response, and learning whether to persevere or pivot.

Innovation accounting

A framework for tracking progress in a startup using learning milestones rather than traditional financial metrics.