Definition (canonical)
An MVP (Minimum Viable Product) is the smallest thing you can build to test whether customers actually want — and will pay for — what you plan to make. It is a learning instrument, not a small product: the deliverable is validated knowledge, and code is optional.
The test defines the MVP, not the feature list
Work backwards from the riskiest assumption. If the risk is demand ("will anyone want this?"), a landing page with pricing and a buy button measuring real clicks answers it — no product needed. If the risk is willingness to pay, pre-orders or signed LOIs. If the risk is the solution itself, a concierge MVP (deliver the service manually behind a facade) or Wizard-of-Oz (interface real, humans backstage) tests it at 1% of build cost. Only when demand and payment are evidenced does building software become the rational next test. The recurring pattern of dead startups: eighteen months of building answered a question nobody asked.
Minimal but viable — both words are load-bearing
"Minimum" fails when the product is so broken it tests patience, not demand — a food-delivery MVP that delivers cold food teaches nothing about the model. "Viable" fails when scope creeps until the MVP is just... the product, slow and expensive. The working balance: ONE core journey done well enough that the target user completes it and reveals a decision (pays, returns, invites) — decisions are data; compliments are noise. Metrics before launch, or the result will be interpreted by mood.
From MVP to iteration
The MVP is turn one of a loop: hypothesis → smallest test → measured behavior → decision (persevere, pivot, kill). Teams that ship an MVP and then switch to feature-factory mode skip the entire point. Kill criteria set in advance ("if <X% activate, we stop this line") protect against the sunk-cost drift that keeps zombie projects alive — and the discipline compounds toward product-market fit, which is the loop's exit condition.
FAQ
Is an MVP allowed to be ugly? Ugly yes, broken no — and in trust-sensitive categories (payments, health) the bar for "viable" includes looking credible.
MVP vs prototype vs beta? Prototype tests feasibility internally; MVP tests demand on real customers; beta hardens a validated product. Different questions, different tools.
Should we charge from day one? Whenever the hypothesis involves payment — a price tag is the strongest truth serum available to founders.
Related: Startup · Product-Market Fit · Business Model Reviewed: 2026-07-15 · Sources: lean startup canon