Validation guides
Minimum Viable Product: Test Your Riskiest Assumption
Most MVPs are too big, because founders treat them as a first release instead of a test. The right MVP tests one thing: the assumption that, if wrong, sinks the idea. Here is how to find it and build only what tests it.
A minimum viable product is the smallest thing you can put in front of real users to test your riskiest assumption. Eric Ries defines it as the version of a product that lets a team collect "the maximum amount of validated learning about customers with the least effort." Pick the assumption first, then build only what tests it, even if that is a landing page, a manual service or a prototype.
The definition matters because it changes what you build. An MVP is measured by what you learn, not by how complete it is. The term was coined by Frank Robinson in 2001 and popularised by Steve Blank and Eric Ries (Wikipedia).
Key takeaways
- An MVP is a test, not a first release.
- Start from the riskiest assumption: the one that sinks the idea if it is wrong.
- The smallest test is often not software: a landing page, a video or a manual service.
- Define success before you build, with a metric and a threshold.
- For Indian founders, SISFS grants of up to ₹20 lakh fund proof of concept and prototypes.
What is a minimum viable product?
The smallest thing that lets you learn whether your riskiest assumption is true, from real users, with the least effort.
Ries's full definition: "The minimum viable product is that version of a new product a team uses to collect the maximum amount of validated learning about customers with the least effort" (Wikipedia).
Two words do most of the work. Validated means the learning comes from what real users do, not what the team believes. Least means anything that does not help you learn that one thing is out of scope.
How do you find your riskiest assumption?
List what must be true for the idea to work, then rank each by how uncertain it is and how badly the idea breaks if it is wrong. The top of that list is what your MVP should test.
| Assumption | How uncertain? | Impact if wrong | Test first? |
|---|---|---|---|
| Clinic managers spend hours a week on appointment reminders | Low (confirmed in interviews) | High | No |
| They will pay ₹2,000 a month to remove it | High | High | Yes |
| They will connect it to their existing booking software | Medium | High | Next |
| They want a mobile app | High | Low | Later |
In this illustration the MVP should test willingness to pay, which may need no software at all: a paid pilot delivered by hand would answer it.
What are the types of MVP?
Anything from a landing page to a manually delivered service to a working prototype. Choose the cheapest type that tests your riskiest assumption with real users.
| Type | What it tests | Example |
|---|---|---|
| Landing page with a price | Interest at a price | Buffer added a pricing page before building, to see which plan people clicked (Buffer) |
| Explainer video | Interest from a specific audience | Dropbox's demo video took its beta waiting list "from 5,000 people to 75,000 people literally overnight" (TechCrunch) |
| Concierge or manual service | Whether the outcome is valuable | Delivering the result by hand to a few customers |
| Clickable prototype | Whether users understand and can use the solution | Screens users click through on their own task |
| Single-feature product | Whether the core feature keeps users coming back | One workflow, built properly |
How do you scope an MVP so it stays minimum?
Write the assumption, the metric and the threshold first. Then cut every feature that does not help you measure that metric.
A one-line MVP brief keeps scope honest:
We believe [assumption]. We will know we are right if [metric] reaches [threshold] within [time], measured with [MVP].
Anything not needed to produce that measurement waits. Our guide to MVP feature prioritization turns validation findings into a build list.
What is the difference between an MVP and a minimum sellable product?
An MVP is built to learn. A minimum sellable product (MSP) is the smallest version someone will actually pay for. Once your riskiest assumption is tested, the next goal is usually the MSP.
The MSP is what you put in front of paying customers after the learning phase. It needs to deliver the promised outcome reliably, which often means more polish than an MVP, but still no more features than the buyer needs. NELL's Launchpad plan includes an MSP definition with a clickable wireframe as part of the build plan.
How can founders in India fund an MVP?
Beyond savings and customer pilots, the Startup India Seed Fund Scheme offers grants for proof of concept and prototype development to eligible DPIIT-recognised startups.
The scheme provides "up to Rs. 20 Lakhs as grant" for "validating proof of concept, prototype development, or product trials", in milestone-based instalments, through participating incubators. Startups must be "recognized by DPIIT" and "incorporated not more than 2 years ago at the time of application" (India Science, Technology and Innovation portal). An MVP scoped around one riskiest assumption fits that kind of milestone well.
Once the MVP is scoped, NELL's Build It tab turns a validated idea into a specification you can hand to an AI coding tool such as Claude Code, Codex or Cursor. Start with a free Quick Validate to find the riskiest assumption.
Frequently asked questions
What is a minimum viable product in simple terms?
The smallest thing you can put in front of real users to learn whether your riskiest assumption is true, with the least effort.
Does an MVP have to be software?
No. A landing page with a price, an explainer video or a service delivered by hand can all be MVPs, if they test the riskiest assumption with real users.
How long should it take to build an MVP?
As little time as the test allows. If it takes months, it is probably testing more than one assumption, or testing the wrong one.
What is the difference between an MVP and a prototype?
A prototype shows how something could work. An MVP is used with real users to test an assumption and produce a measurable result. A prototype can be one type of MVP.
What comes after an MVP?
If the riskiest assumption holds, test the next riskiest, and build toward a minimum sellable product: the smallest version customers will pay for and keep using.
Start where you are
Find the assumption first. Then build only what tests it.
Sources
- Wikipedia, Minimum viable product
- Buffer, Idea to Paying Customers in 7 Weeks: How We Did It (2011)
- Eric Ries in TechCrunch, Dropbox Minimal Viable Product (2011)
- India Science, Technology and Innovation portal, Startup India Seed Fund Scheme (SISFS)
The clinic example and prices are illustrations. Scheme terms change; check the official Startup India portal before applying.