A minimum viable product is the earliest version of a product that can be deployed to real users and generate useful feedback. The concept, popularized by Eric Ries in The Lean Startup, is a response to the pattern of companies spending months or years building products in isolation, only to discover that what they built does not match what users actually need.
The word "minimum" is frequently misunderstood. An MVP is not a prototype, a demo, or a half-built product. It is a working product that delivers a specific value to a specific user and does that one thing well enough that users choose to use it and return to it. Everything beyond that core value proposition is scope that can wait until there is evidence to justify it.
For technical founders and their teams, the challenge of an MVP is discipline: the ability to identify what is genuinely necessary for initial learning and to defer everything else. Common failure modes include building for scale before achieving product-market fit, adding features based on assumption rather than user behavior, and treating engineering completeness as a prerequisite for shipping.
For non-technical founders, the challenge is different: knowing what is and is not feasible within the constraints of time, budget, and team capability, and making architecture decisions in the MVP phase that do not create prohibitive technical debt when the product needs to grow. This is where experienced engineering leadership, even fractional, pays for itself quickly.