What Is an MVP? Testing Assumptions Before Expanding a Product

An MVP, or Minimum Viable Product, helps a team investigate an uncertain customer or business assumption through a limited offering and observation. Define the learning objective first, then choose the features, participants, and evidence needed. The purpose is to inform the next investment decision before expanding the product.

Start with the learning objective

Eric Ries’s explanation of an MVP emphasizes learning about customers with limited effort. Reducing the number of features or screens is not sufficient when the result cannot answer the team’s question. Connect a specific customer, proposed value, uncertain assumption, and observable behavior.

For example, “a team member uses a prepared weekly report in a meeting” and “we can automate report preparation economically” are separate assumptions. One experiment may not answer both. Decide whether customer value or technical feasibility needs to be investigated first, considering which uncertainty could invalidate the business.

Choose one important uncertainty

Replace a broad ambition such as “build a convenient service” with a statement about who will do what under which conditions. Investigate the customer’s current workaround and the conditions under which they might change it. This makes it easier to avoid spending heavily on a polished demonstration that does not address the main risk.

A positive answer to “Would you want this?” is a clue, not the same evidence as usage or payment. Separate enthusiasm for a feature, willingness to try something free, and willingness to rely on it repeatedly. Selecting the evidence clarifies what to ask and what to provide.

Distinguish MVPs, prototypes, proofs of concept, and betas

Organizations use these terms differently, and they are not necessarily exclusive stages. A customer-facing prototype can contribute to an MVP experiment. Describe the question and limits of a particular test rather than assuming its label explains everything.

Prototypes are not limited to internal teams. External users can try them while researchers observe understanding and operation. Conversely, a feature-rich beta may yield little decision-relevant learning if the team has not specified whom it serves and what it is testing.

Select the smallest offering that answers the question

In the fictional reporting example, a person could organize an authorized sample of documents and return a report to participants. Observing whether they use it in a meeting and what they change can provide evidence about value. Explain the manual work, trial scope, and expected participation so people do not mistake the service for completed automation.

Successful manual delivery does not demonstrate that software can achieve the same accuracy or cost. Record the work and checks involved, then investigate automation and economics separately. Otherwise, a positive customer reaction can conceal an expensive or technically fragile delivery process.

A signup page can also be useful, but it measures response to the explanation and offer shown. Registrations do not establish actual product value, retention, and payment at once. Record both what was observed and which questions remain unanswered.

Preserve minimum safety, quality, and clarity

A small experiment still needs appropriate safeguards. Check the purpose and permissions for personal data, retention, payment and contract explanations, and incident handling. Where failure could cause significant harm, begin with simulated or non-production data, or change the test method instead of exposing participants to an avoidable risk.

• Explain what is being tested and what is actually available.
• State the delivery period, supported scope, and contact route.
• Collect only necessary evidence and assign retention and deletion responsibilities.
• Define stopping conditions, remedies where needed, and a fallback for participants.

Speed alone is an incomplete measure of experiment quality. List the minimum learning scope separately from the conditions that must be protected. This allows the team to reduce unnecessary work without confusing a limited experiment with permission to omit essential responsibilities.

Use results to continue, change direction, or stop

The Lean Startup principles connect building, measuring, and learning. In practice, ask what the present evidence supports and what uncertainty remains. A small test does not conclusively establish the truth of a business assumption across all customers and conditions.

If the report is not used, inspect the target group, meeting context, content, and delivery timing before deciding why. If it is used, the next question might concern another group, repeated use, or payment. Feature expansion is one possible next action; the learning should determine which action deserves attention.

For the next step, use a completed MVP experiment plan; the first-use onboarding process; practical examples of creating value with customers.