A minimum viable product (MVP) is the first version of a new product that lets a team collect the maximum amount of validated learning about customers with the least effort. It is not a half-finished product but a deliberately small one, built to test whether an idea solves a real problem before larger investments are made.
MVP is a product philosophy before it is a deliverable, and that is where most misunderstandings start. This guide covers five things people often do not know about a good MVP: what the term really means, the strategy behind it, why it starts small, what makes it good, and which types exist. It then looks at a real example and at how AI and no-code tools are changing MVPs in 2026. For costs and timelines, see our 2026 guide to building an MVP.
TL;DR / Key Takeaways
- An MVP is the smallest version of a product that can test your riskiest assumption with real users; the goal is learning, not a smaller launch.
- Eric Ries, who popularised the term in The Lean Startup, defines it by validated learning with the least effort, not by the number of features.
- A good MVP follows the Build, Measure, Learn loop: build the minimum, measure what users do, learn and decide what comes next.
- There are several types, from landing pages and explainer videos to Wizard of Oz, concierge and single-feature MVPs; each tests a different question.
- AI-assisted prototyping and no-code tools make building faster, but they do not remove the need for a clear hypothesis and real user feedback.
What does MVP really mean?
MVP stands for minimum viable product. Eric Ries defines it as "that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort." In his early blog posts, later collected in The Lean Startup, it is a version of the final product, not the final product itself.
There are many definitions of MVP, almost as many as the teams that use it to bring new products or services to market. That is because the idea is not very prescriptive: it works as a light framework rather than a strict method. This flexibility is why it suits start-ups as well as established companies and even investors who want to test an opportunity.
Some practitioners suggest reading the acronym as market viable product instead: a first version that can already stand on the market next to its competitors, even if it is only minimally profitable. Whichever reading you prefer, the principle is the same. An MVP is a way of seizing an opportunity while reducing risk and learning as quickly as possible.
What strategy lies behind a good MVP?
The strategy behind an MVP is to replace one distant goal, "success", with smaller goals you can reach and test now. A test website, a landing page, a demo or a minimal product lets you probe the market, explore opportunities and find where the uncertainties lie, so that each next step is based on evidence.
Asking the right questions comes first: does your idea require teamwork, are there financial risks, which professionals do you need, and for how many hours? A good MVP also leaves you free to change course and question everything, because it has been built to learn rather than to prove you right.
Raffaele Gaito, one of Italy's best-known growth hacking experts, groups these questions into nine areas: assessing new opportunities, knowing your scenario, the value proposition, making decisions, being creative, analysing the business, innovation and change, customers, and personal success. Working through them before building saves time and money later.
(The Top 20 Reasons Startups Fail by CB Insights; the report has since been updated to twelve reasons)
Why should an MVP start small?
An MVP starts small because the most lasting products begin with only what is essential, then grow from what users actually do. Starting small keeps the cost of being wrong low, makes feedback faster to collect and turns the list of requirements into something derived from evidence rather than from assumptions made in a meeting room.
The requirements your developers work on should come from the feedback of your first users: people who sign up to your website, landing page, app or online shop, or who show interest in the functions you have assumed. This way of improving an initial solution step by step follows the Build, Measure, Learn loop described in The Lean Startup.
Software teams have an advantage here, because lean thinking and agile methods are already how they solve complex problems. WWG's web MVP service, for example, focuses on must-have features and integrates early user feedback into each release; the same approach applies to mobile MVPs.
A resilient MVP usually goes through five stages. Each one builds on the evidence gathered in the previous one:
- Ideas and research: develop the idea around the uses and needs it can meet.
- User-centred design: design from the user's perspective and the benefits they are really looking for.
- Few features: include only the highest-priority features, those that satisfy the first users and can evolve later.
- Building the product: even a minimum product must have the quality its users need and deliver real value.
- Validation and improvement: collect feedback and improve the product to retain users and attract new ones.
(Scrum Values from the Gunther Verheyen website)
What makes an MVP good rather than just small?
A good MVP delivers enough value that people actually want to use it, is built with minimal resources, and is designed so that its next version is shaped by user feedback. Smallness alone is not the goal: an MVP that nobody uses teaches you nothing, and one that took a year to build is no longer minimal.
MVP is the first stage of product management in lean and hybrid environments. In simple terms, it is a first version developed with basic features that meet the needs of early users. Once product managers get feedback from those users, they can test the feature set and decide whether and how to build the full product.
The advantage is that it reduces design and production costs and the risk of investing in a product the market does not accept. A well-designed MVP should:
- offer minimum value, but with features people are willing to use;
- retain users through its benefits, most of which will be realised later;
- base its future development on user feedback;
- test and validate the idea with minimal resources;
- speed up the learning of everyone involved;
- reduce development, design and release hours;
- reach the first customers in a short time;
- serve as a base for the products that follow;
- show what the team can think, build and launch;
- give a quick estimate of the potential of a new brand.
(Image from the book Delivering Happiness by Tony Hsieh)
What types of MVP are there?
MVPs differ in what they test and how much they cost to build. A landing page or an explainer video tests demand before any product exists; Wizard of Oz and concierge MVPs test the service by doing the work manually; single-feature and piecemeal MVPs test a working product; crowdfunding and pitch decks test willingness to pay or invest.
| Type | How it works | What it tests | Relative effort |
|---|---|---|---|
| Landing page | A page describes the product and asks visitors to sign up or pre-order | Whether the problem and the promise attract interest | Very low |
| Explainer video or demo | A video or clickable demo shows how the product would work | Demand for a product that does not exist yet | Low |
| Wizard of Oz | Users see an automated product; behind the scenes the work is done by hand | Whether the service is valuable before automating it | Low to medium |
| Concierge | The service is delivered manually and users know a person is doing it | What users really need, step by step | Low to medium |
| Single-feature | One core feature built properly, nothing else | Whether the core problem is worth solving | Medium |
| Piecemeal or no-code | The product is assembled from existing tools and services | The whole flow, without custom development | Low to medium |
| AI-assisted prototype | A working prototype generated or accelerated with AI coding tools | Usability and flow with real users | Low to medium |
| Crowdfunding or pitch deck | The idea is presented to backers or investors before it is built | Willingness to pay or to invest | Medium |
Choose the type by the question you need answered first. If you are unsure people want the product, start with a landing page or a video. If you are unsure you can deliver the value, run it as a concierge or Wizard of Oz service. If both are proven, a single-feature MVP is the next step.
The cost of an MVP depends mainly on the number of functions, the complexity of each one and the team you involve: internal, outsourced, freelance or agency. Our 2026 MVP guide explains how to estimate cost and timeline in detail.
What is a real example of a successful MVP?
Dropbox is one of the best-documented MVP examples. Before the product was ready, founder Drew Houston published a three-minute video demonstrating how file synchronisation was meant to work, aimed at a community of early adopters. According to Houston, the beta waiting list went from 5,000 to 75,000 people overnight.
The video was an explainer MVP: it tested demand for a technically difficult product without building it first. As Eric Ries tells the story in a guest post for TechCrunch, it gave the team evidence that people wanted the product before the hardest engineering work was done.
The lesson is not that every MVP should be a video. It is that the cheapest test that answers your riskiest question is usually the right MVP, even if it contains no code at all.
How are AI and no-code tools changing MVPs in 2026?
AI coding assistants and no-code platforms make it much faster to build a working first version, so the cost of the "build" step has fallen. What they have not changed is the rest of the loop: you still need a clear hypothesis, real users and honest measurement. A faster build only helps if you learn from it.
The risk is building more than you need because building has become easy. A prototype generated with AI tools can look finished while hiding weak security, poor data handling or code that is hard to maintain. That is fine for a test with a few users, but it should be reviewed before real customers or real data depend on it.
The practical rule is to use AI and no-code to reach users sooner, then decide deliberately what to rebuild properly once the idea is validated. The Lean Startup approach inside established companies works the same way: small, fast experiments first, larger investments only when the evidence supports them.
To talk through the right type of MVP for your idea, write to us at business@wwg.it.
Sources
- Eric Ries, Minimum Viable Product: a guide, Startup Lessons Learned (2009) — https://www.startuplessonslearned.com/2009/08/minimum-viable-product-guide.html
- The Lean Startup, Principles — https://theleanstartup.com/principles
- TechCrunch, Eric Ries, Dropbox minimal viable product (2011) — https://techcrunch.com/2011/10/19/dropbox-minimal-viable-product/
- Raffaele Gaito, Domande per crescere — https://www.raffaelegaito.com/domande-crescere/
- CB Insights, The Top Reasons Startups Fail (archived December 2025) — https://web.archive.org/web/20251222225750/https://www.cbinsights.com/research/report/startup-failure-reasons-top/
- Gunther Verheyen, The Scrum Values — https://guntherverheyen.com/library/the-scrum-values/
FAQ
Frequently Asked Questions
A prototype shows how a product could look or work and is usually tested internally or with a few users; it does not need to deliver real value. An MVP is released to real users to test whether the product is valuable and whether they will use it. A prototype answers design questions; an MVP answers market questions.
It depends on the type of MVP, the number and complexity of its features and the team that builds it. A landing page or concierge test costs very little, while a single-feature software product needs a development budget. Our 2026 MVP guide explains how to estimate the budget.
It depends on the type. A landing page, video or concierge service can be set up in days, because it needs little or no code. A single-feature web or mobile product takes longer, since it must be designed, built and tested well enough for real users. Defining the riskiest assumption first keeps the timeline short.
No. The MVP approach suits start-ups, but it works just as well for established companies launching a new product, service or internal tool, and for investors who want evidence before committing funds. Any organisation that faces uncertainty about what customers want can use an MVP to learn quickly and reduce the risk of a larger investment.
Measure how users behave, collect their feedback and compare the results with the hypothesis you set out to test. Then decide whether to continue, change direction or stop. If the idea is validated, plan the next release around the features users actually need, and review any shortcuts taken to build the first version quickly.
