{"id":47759,"date":"2023-04-06T15:00:37","date_gmt":"2023-04-06T15:00:37","guid":{"rendered":"https:\/\/www.arimetrics.com\/glosario-digital\/mvp"},"modified":"2026-10-01T16:05:45","modified_gmt":"2026-10-01T16:05:45","slug":"mvp","status":"publish","type":"encyclopedia","link":"https:\/\/www.arimetrics.com\/en\/digital-glossary\/mvp","title":{"rendered":"MVP"},"content":{"rendered":"

Definition\"MVP\"<\/strong><\/p>\n

An MVP<\/strong>, or Minimum Viable Product, is a version of a product used to test a business hypothesis with real users and obtain useful learning with the least reasonable development. It must deliver enough value for observed behavior to be interpretable, but it is not intended to represent the final product.<\/p>\n

The term does not mean any unfinished product or a minimum list of features. Its scope depends on the specific hypothesis<\/strong> being tested, the selected audience, and the evidence required to decide whether to retain, change, or abandon a proposition.<\/p>\n\n

What an MVP is for<\/h2>\n

An MVP reduces initial uncertainty<\/strong> before more resources are committed. It can reveal whether a problem exists, whether a solution is understandable, whether particular users adopt it, or whether they are willing to take a relevant action such as registering, requesting information, or paying.<\/p>\n

Within the Lean Startup<\/a> approach, its primary function is to generate validated learning<\/strong>. The team states an assumption, builds a proportionate test, measures the response, and uses the findings to choose the next step. Speed is valuable only when it shortens this cycle without invalidating the test.<\/p>\n

An MVP also helps specify what success means. Before release, the segment, expected behavior, metric<\/a>, observation period, and decision rule<\/strong> should be defined. Without these conditions, any outcome can later be reinterpreted to support the original idea.<\/p>\n

How to design the test<\/h2>\n

Design begins with the critical hypothesis<\/strong>, not with an accumulation of features. The selected artifact should let users experience the value proposition or display behavior sufficiently close to the behavior that matters.<\/p>\n

A basic cycle<\/strong> can be organized as follows:<\/p>\n

    \n
  1. Hypothesis:<\/strong> state what is believed about the problem, user, solution, or business model.<\/li>\n
  2. Segment:<\/strong> define who will participate and why they represent the audience under study.<\/li>\n
  3. Experiment:<\/strong> select the simplest version capable of producing relevant evidence.<\/li>\n
  4. Measurement:<\/strong> set sources, events, thresholds, and possible confounding factors before release.<\/li>\n
  5. Decision:<\/strong> agree which outcome means persevere, change the approach, or stop the initiative.<\/li>\n<\/ol>\n

    The MVP may be a functional digital product<\/a>, but it does not always require complete automation<\/strong>. A manually delivered service behind an interface, a service restricted to a few customers, or a landing page<\/a> can work if it reproduces the decision being studied and does not invite conclusions that the test cannot support.<\/p>\n

    Types of MVP<\/h2>\n

    The appropriate form depends on the question and risk. There is no universal format<\/strong>, and several tests may precede an operational version.<\/p>\n

    Common MVP formats<\/strong> include:<\/p>\n