Eric Ries defined it in 2009: a 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.
An MVP is not a broken product. It is not a demo. It is not a prototype. It is a real product with just enough features to deliver value and learn from real customer behavior. The goal is learning, not launching.
Types of MVPs
Concierge MVP. You manually deliver the service. No automation, no software. You solve the problem for one customer by hand. If they pay, you have validation. Then you automate.
Wizard of Oz MVP. It looks automated to the user, but you do everything manually behind the scenes. The user thinks they are using software. You are doing it by hand.
Landing page MVP. A simple page describing your product. Drive traffic. Measure sign-ups. If people give you their email for a product that does not exist, demand exists.
Single-feature MVP. Build only the core feature. Nothing else. Test if that one feature delivers enough value to retain users.
Famous MVP Examples
Airbnb. The founders rented out their own apartment with air mattresses. No platform, no website. They validated that people would pay to stay in a stranger's home. Then they built the platform.
Dropbox. A video demonstrating the product before it existed. 75,000 sign-ups overnight. They validated demand without writing sync code.
Zappos. The founder took photos of shoes in local stores, posted them online, and bought them at retail price when orders came in. He validated that people would buy shoes online without building inventory.
Buffer. A landing page with pricing plans before the product existed. When people clicked "Sign Up," they saw a message: "You caught us before we are ready." They collected emails and validated demand.
How to Build Your MVP
Step 1: Identify the core hypothesis. "Customers will pay for X." Not "Customers might like X." A testable, falsifiable hypothesis.
Step 2: Define the minimum feature set. What is the smallest thing you can build to test that hypothesis? Cut everything else.
Step 3: Build it fast. Use no-code tools if possible. Bubble, Webflow, Glide, Softr. Most have free tiers. Ship in two weeks, not six months.
Step 4: Get it in front of real users. Post on Reddit, Twitter, Facebook Groups. Do not wait for it to be perfect. Ugly and working beats beautiful and unfinished.
Step 5: Measure and learn. Track activation rate, retention, and willingness to pay. If the data supports your hypothesis, keep building. If not, pivot.
Common MVP Mistakes
Building too much. You do not need 50 features. You need one feature that delivers value. Everything else is waste.
Waiting too long to ship. Your MVP will be wrong. Ship it anyway. You learn more from real users than from internal debates.
Ignoring the data. If users do not use your product, that is data. Do not convince yourself they will come back. Pivot or drop it.
Confusing an MVP with a bad product. An MVP is not an excuse for poor quality. It is the minimum viable version of a good product. If it does not work, it is not an MVP. It is a broken product.
MVP and Product Validation
An MVP is a product validation tool. It tests whether your specific solution solves the problem. Market validation tests whether the problem exists. You need both.
Start with market validation. Use GetNoBurn to check demand signals from YouTube, Reddit, Google Trends, and Product Hunt. If the market exists, build an MVP to test your solution.
Product validation without market validation is building a great product for a market that does not exist. Market validation without product validation is knowing the problem exists but not solving it. You need both.
The Bottom Line
An MVP is the fastest way to test your business idea with real customers. It is not about building something small. It is about learning something big: whether strangers will pay for what you are building.
Start with a landing page or a concierge service. Get real data. Then build the product.