MVP (Minimum Viable Product)

An MVP is the smallest thing that tests the riskiest assumption standing between you and knowing whether to invest. Eric Ries coined the term in The Lean Startup (2011): “the version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort.” The key word is learning, not shipping. In a PM interview, the fastest signal of junior thinking is defining MVP as a stripped-down launch; the fastest signal of senior thinking is naming the assumption being tested and the metric that confirms or kills it.

The common misread

“Minimum” gets read as “low quality” or “fewer features.” It means minimum scope to learn the one thing you most need to know. The goal is a predefined learning signal, not a product milestone. A correct MVP has three elements: a specific assumption, a form of evidence chosen because it is the cheapest way to test that assumption, and a success metric that triggers the next investment decision.

Real examples and what each tested

Each of these is famous because the founders chose the cheapest possible test of the one assumption that mattered.

• Dropbox: a 3-minute demo video, no product built. Collected 75,000 email signups overnight. Tested whether people wanted cloud file sync before a single line of backend code existed.
• Airbnb: three air mattresses in the founders’ apartment, listed on a basic site. Tested whether strangers would pay to stay in someone’s home.
• Uber: San Francisco only, SMS dispatch, no app. Tested whether people would hail a car via smartphone.
• Amazon: bookstore only. Books were easy to ship and catalog. Tested whether people would buy physical goods online.

None of these were shipped as complete products. Each targeted the one assumption that, if wrong, would void the whole bet.

MVP vs RAT vs MLP vs MMP

Senior interviewers distinguish these and penalize candidates who conflate them. Knowing the vocabulary signals that you think about learning, not just shipping.

RAT (Riskiest Assumption Test): narrower than MVP. A RAT is a single experiment, often not a shippable product: a fake-door landing page, a Wizard of Oz backend, a five-person concierge trial. Prefer RAT vocabulary when you want to signal precision: you are testing one thing, not building a product.

MLP (Minimum Lovable Product): applies when demand is already validated and the question is whether users will switch and stay. Requires higher craft than a traditional MVP because users compare you to polished incumbents and churn immediately if the experience is rough. “Minimum” still means minimum scope, but the quality bar is set by the market, not by what is cheapest to build.

MMP (Minimum Marketable Product): the narrowest thing you can sell commercially. More relevant in enterprise B2B where procurement and compliance requirements would disqualify a startup MVP from consideration.

The 2026 reframe

AI coding tools (Cursor, Replit Agent, v0, Bolt) have reduced the cost of building an MVP-grade prototype to a weekend. The old tradeoff (“what can we build cheaply?”) has mostly collapsed. The bottleneck is now hypothesis clarity: what is the riskiest assumption, and what is the cheapest form of evidence? MVP becomes a thinking tool first and a build decision second. In crowded consumer markets, the lovable floor has also risen: users compare your MVP to polished incumbents on day one, so minimum quality is set by the competitive baseline, not by what you can ship fast. See feasibility is free and proving viability.

Context changes the answer

The right MVP answer depends on where you are building and for whom.

• Early-stage startup: test demand before building anything. A fake-door or a video is a valid MVP. The riskiest assumption is almost always “do people want this at all.”
• Competitive consumer market: the lovable floor is set by incumbents. “Minimum” means minimum scope, not minimum craft. An MLP frame is often more accurate than an MVP frame here.
• Enterprise B2B: procurement, security review, and integration requirements push you toward MMP. An MVP that does not clear compliance baselines will not get piloted, regardless of how well it tests the core assumption.
• Apple: the MVP frame is a poor fit for Apple’s culture. Apple ships products that are expected to exceed existing standards, not clear a minimum bar. If you are interviewing at Apple, use language about deliberate scope and quality gates rather than MVP.

Weak vs strong interview answer

Applying this in a product-sense question

When a question asks for your MVP (for example: “design a product for commuters”), do not jump to a feature list. Work the assumption stack first.

A commuter product could serve dozens of different needs. The riskiest assumption determines the shape of the MVP. If you are betting on stress reduction during transit, the riskiest assumption is whether commuters will engage with a calming experience on a phone when they are already distracted and often underground. A RAT for that assumption is not an app: it is a curated playlist recommendation sent via push notification and a one-tap mood check-in. The success metric is completion rate and opt-out rate over two weeks, not DAU.

Once you state the assumption and the metric, the MVP scope becomes obvious. That is the move interviewers are looking for. For more on scoping a product-sense answer at this level, see how to prioritize a backlog and design a product for commuters.