What an MVP Actually Is
The minimum viable product concept — introduced most influentially by Eric Ries in The Lean Startup — is the version of a new product that allows the team to collect the maximum amount of validated learning about customers with the minimum effort. The two operative words that most define the concept and that are most frequently misunderstood in practice: minimum (the least amount of product required to generate the specified learning, not the least polish or the least care, and not a full product with minimal features) and viable (sufficient to deliver the core value proposition that generates genuine customer engagement, not a broken prototype that no one would actually use).
The MVP misconception that most commonly produces the poorly designed experiments that generate misleading conclusions: the equation of MVP with cheap, half-finished product that the team is embarrassed to show customers. The MVP is not a product the team is embarrassed by — it is a deliberately scoped experiment designed to test a specific hypothesis about customer behaviour. The landing page that describes the product and measures sign-up intent is an MVP for the hypothesis that customers will seek the product out; the concierge service that manually delivers the product’s promised outcome is an MVP for the hypothesis that customers value the outcome enough to pay for it; the low-fidelity prototype that tests the user experience flow is an MVP for the hypothesis that the proposed interface enables the intended behaviour. Each MVP is designed for a specific question, not for a specific quality level.
Identifying the Riskiest Assumptions to Test
The assumption identification process that most efficiently directs MVP design toward the learning that most matters: the explicit mapping of the assumptions underlying the business model — all the things that must be true for the business to work as envisioned — followed by the ranking of those assumptions by their combination of importance (how much the business depends on this assumption being true) and uncertainty (how confident the team is that it is true). The assumption that is both critically important and highly uncertain is the assumption that most needs to be tested before significant investment is made; the assumption that is certain to be true or whose falsity would only marginally affect the business outcome is the assumption that does not need to be tested before building.
The riskiest assumption categories that most early-stage startups share regardless of their specific market or product: the problem assumption (do the target customers actually experience the problem the startup is solving, with the frequency and severity that make them willing to pay for a solution?), the solution assumption (does the specific solution the startup has designed actually solve the problem better than what customers currently do?), and the willingness-to-pay assumption (will customers pay the price that the business model requires, rather than expecting the solution to be free or priced lower than the model requires?). The startup that has not tested all three of these assumptions has not validated the fundamental viability of its business model, regardless of how much product has been built.
MVP Types and When to Use Each
The MVP format selection that most efficiently tests each type of assumption: the landing page MVP (a webpage that describes the product and invites visitors to sign up for early access, measuring the sign-up rate as a proxy for purchase intent) for testing the problem-solution fit assumption before building anything, the concierge MVP (manually delivering the product’s promised outcome to early customers without any automation) for testing the value of the outcome and the customer’s willingness to pay before investing in the technology that would deliver the outcome at scale, the Wizard of Oz MVP (the product that appears to be automated but that is actually operated manually behind the scenes) for testing the customer experience of a product whose underlying technology is not yet built, and the single-feature MVP (the product that implements one core feature rather than the full feature set) for testing whether the core value proposition justifies a purchase without the distraction of additional features.
The MVP design error that most commonly generates the misleading validation that gives founders false confidence: the MVP that is tested with the wrong audience. The product tested with friends, family, and professional contacts who provide encouraging feedback out of social obligation rather than genuine interest has not been tested with the target customer — and the enthusiasm of the supportive audience produces the misleading signal that has caused many founders to invest significantly in building a product that the actual target market does not want. The MVP that is tested with strangers who have no social obligation to be encouraging, who found the MVP independently rather than being recruited by the founder, and who represent the specific demographic and psychographic profile of the target customer generates the genuine signal that determines whether the assumption is valid.
Learning From the MVP
The MVP learning extraction process that most effectively converts the MVP results into the specific, actionable conclusions that should guide the next iteration: the structured analysis that compares the actual customer behaviour to the specific predictions that the assumption implied, rather than the general impression of whether the MVP went well or poorly. The hypothesis that thirty percent of website visitors will sign up for the waiting list is validated when the actual sign-up rate meets or exceeds thirty percent and invalidated when it falls significantly short — the specific comparison against a pre-stated prediction is what makes the result meaningful rather than subject to the post-hoc rationalisation that interprets any outcome as validation.
The pivot decision that most correctly responds to MVP results that do not validate the initial assumptions: the specific, hypothesis-driven pivot that changes the element of the business model that the MVP evidence most directly challenges, rather than the panicked pivot that changes everything or the denial pivot that rationalises the results as anomalies and continues without change. The MVP that showed low conversion from the target customer segment but high enthusiasm from an unexpected adjacent segment is providing a specific signal about the customer pivot that the evidence supports; the founder who changes their target customer based on this evidence has made an evidence-based pivot that may redirect the business toward a more promising market.
From MVP to Product
The product development sequencing that most effectively builds on MVP learning to create the product that succeeds in the market: the iterative development cycle that builds only the next increment of product that the current evidence supports, tests that increment with real customers, extracts the learning, and uses that learning to determine the next increment — rather than the big build that attempts to build the full envisioned product based on the assumptions that the MVP was supposed to test but that was never allowed to test them properly.
The MVP graduation criteria that most clearly indicate when the startup has learned enough from the MVP phase to invest in building the more complete product: the evidence of repeatable, customer-initiated demand that does not depend on founder heroics to generate, the positive unit economics on the MVP transactions that indicate the business model works at the scale the MVP tests, and the specific understanding of which product elements most drive the customer value that motivates purchase and retention. The startup that graduates from MVP to product build with this evidence is investing in a product whose market has been validated; the one that begins the full product build before this evidence exists is investing based on assumptions that the MVP phase was supposed to resolve.