Minimum Viable Product (MVP)
Build the smallest thing that lets you start learning from real customers
The Minimum Viable Product is the version of a new product that allows a team to collect the maximum amount of validated learning about customers with the least effort. It is not simply a smaller or cheaper product; it is specifically designed to test fundamental business hypotheses. The MVP may lack many features that will prove essential later, and it may seem embarrassingly simple, but its purpose is learning, not revenue.
MVPs come in many forms depending on what needs to be tested. A video MVP (like Dropbox's explainer video) tests demand without building the product. A concierge MVP delivers the product experience manually to a tiny number of customers to learn what they value. A Wizard of Oz MVP presents an automated facade while humans perform the work behind the scenes. A smoke test MVP uses a landing page or fake door to measure interest before building anything.
The critical principle is that any additional work beyond what is required to start learning is waste. Entrepreneurs consistently overestimate how many features are needed to start gathering customer feedback, and they underestimate how much can be learned from a rough early version. The MVP challenges entrepreneurs to question every assumption about what customers need by putting something real in front of them as fast as possible.
- Any work beyond the minimum needed to start learning is waste.
- The right form of your MVP depends entirely on which hypothesis most urgently needs testing.
- Entrepreneurs consistently overestimate how many features are required before customers can give meaningful feedback.
- Learning comes from putting something real in front of customers, not from debating internally.
- A rough early version in the hands of real users produces more useful data than a polished version in your head.
- Identify the Riskiest Assumption to TestDetermine which leap-of-faith assumption, if wrong, would invalidate the entire business. This is typically either the value hypothesis (do customers want this?) or the growth hypothesis (can we acquire customers efficiently?). Your MVP should target this specific assumption.
- Choose the Right MVP TypeSelect the lightest-weight form of MVP that can test your hypothesis. Options include video demonstrations, landing page smoke tests, concierge delivery to individual customers, Wizard of Oz prototypes, or single-feature products. Match the MVP type to the assumption you need to test.
- Strip Away Everything Non-EssentialRemove every feature, design element, and piece of infrastructure that does not directly contribute to testing your hypothesis. If you are unsure whether something is necessary, leave it out. You can always add it in the next iteration if the data shows it matters.
- Launch to Early Adopters, Not the MainstreamTarget customers who feel the problem most acutely and are willing to try imperfect solutions. Early adopters can imagine what the product could become and will provide more useful feedback than mainstream customers who expect a polished experience.
- Measure Customer Behavior, Not OpinionsTrack what customers actually do with your MVP rather than what they say they would do. Measure sign-up rates, activation rates, retention, and willingness to pay. These behavioral metrics reveal the truth about your hypothesis far more reliably than surveys or focus groups.
Dropbox founder Drew Houston faced a challenge: the product required complex technical infrastructure to build, but he needed to validate demand first. Instead of building the full product, he created a three-minute video demonstrating what the product would do. The video was targeted at technology early adopters and showed the planned user experience.
Food on the Table CEO Manuel Rosso started his meal planning service with a single customer. Rather than building an app, he personally went to the grocery store, found deals, and created customized meal plans for one family. He then manually repeated this process, adding one customer at a time, learning exactly what they valued before automating anything.
The Minimum Viable Product is the version of a new product that allows a team to collect the maximum amount of validated learning about customers with the least effort. It is not simply a smaller or cheaper product; it is specifically designed to test fundamental business hypotheses. The MVP may lack many features that will prove essential later, and it may seem embarrassingly simple, but its purpose is learning, not revenue.
MVPs come in many forms depending on what needs to be tested. A video