The instinct when starting something new is to protect the idea until it's ready — build the full, polished version in private, then reveal it once it's something to be proud of. In his 2011 book The Lean Startup, Eric Ries argued for close to the opposite: build the smallest possible version that tests the riskiest assumption first, show it to real customers immediately, and treat the whole business as a series of experiments rather than a plan being executed on faith.
The build-measure-learn loop
Ries's core cycle has three steps. Build a minimum viable product — not a low-quality product, but specifically the version that generates the maximum validated learning about customers for the least possible effort. Measure how people actually behave when they use it, not what they say they'd do in a survey. Then use that data to decide whether to persevere with the current strategy or pivot to a meaningfully different one. Repeat the loop as fast as the business can manage, on the theory that a company's early survival depends less on getting the plan right on paper than on how quickly it can find out where the plan is wrong.
A minimum viable product is a question, not a discount
The most common misreading of the term treats "minimum viable" as shorthand for "cheap" or "unfinished, shipped early to save money." That's not what it's built to do. An MVP is deliberately built to test one specific hypothesis about what customers actually want — sometimes something as minimal as a single landing page describing a feature that doesn't exist yet, or a "concierge" version where a human quietly does by hand what the eventual product would automate, just to see if anyone signs up at all. The point isn't shipping a lesser copy of the real product. It's finding out, as cheaply and quickly as possible, whether the real product is even worth building.
What we're still unsure about
The Lean Startup's central claim — that fast, cheap experimentation reliably beats extensive upfront planning — grew out of software startup case studies, where building and discarding an experiment really can be quick and inexpensive. Outside that context, the picture is genuinely more contested: in capital-intensive fields like hardware, biotech, or infrastructure, even a "minimal" testable version can take months and real money to build, and critics argue the methodology doesn't transfer as cleanly to those settings as its software-world success stories suggest. How far "build fast, test cheap" generalises beyond the domain it was developed in is still an open, actively argued question in entrepreneurship research, not a universally settled method.
This sits inside Lean Startup & Minimum Viable Product, one of seven topics in Entrepreneurship, one of four domains in Business, one of seventeen subjects the app can quiz you on.