The primary goal of creating an MVP is not to build something, but to learn something.
You have learned what a Minimum Viable Product (MVP) is — a minimal version of a product designed to test your core assumptions with the least effort. The actual job is to treat the MVP not as a product but as an ongoing experiment that guides the development of the final product.
Running an MVP experiment is a disciplined, step-by-step process. It ensures you avoid wasted effort on building features nobody wants and instead learn quickly what your customers truly value.
Start by listening to the voice of the customer
The first and most crucial step in running an MVP experiment is to listen carefully to the voice of the customer. This is the source of truth that directs you to the real problem — and eventually to the right solution.
As a Product Manager, you should ask yourself:
- Who will buy this product?
- Why will they buy it?
- How will they buy it?
Consider the example of Uber. The founder noticed that people wanted to rent premium black cars from local taxi companies but found these options prohibitively expensive. That insight came directly from observing customer behavior and needs.
Listening to your customers is not a one-time activity — it is continuous throughout the MVP experiment.
Identify and prioritize your assumptions
Once you understand the problem and brainstorm potential solutions, you have inevitably made many assumptions. These assumptions are guesses about your customers, their needs, and how they will behave.
It is critical to understand your target audience deeply. One way to do this is by creating user personas — archetypes that represent different segments of your customers. Without understanding your customers, your assumptions will likely be inaccurate.
But even with user personas, you will have multiple assumptions. You cannot test them all at once. You must prioritize and start with the riskiest assumptions — those that, if proven wrong, would have the most devastating impact on your product’s success.
For example, Uber assumed that timeliness was the most important attribute for a certain group of customers, such as office goers. This was a testable assumption.
Build testable hypotheses around your assumptions
Assumptions are guesses. Hypotheses are testable statements derived from those assumptions. Developing a product is no less than a scientific experiment.
Unlike assumptions, hypotheses are measurable, actionable, and have defined outcomes. For example, an assumption might say, “Timeliness is important to office goers.” A hypothesis would say, “If our service reduces wait time to under 5 minutes, then 70% of office goers will use our app at least twice a week.”
The hypothesis is precise and measurable. You can design your MVP experiment to validate or invalidate it.
Set minimum criteria for success with metrics
To determine whether a hypothesis is valid, you must set clear metrics — minimum criteria for success. These metrics represent the break-even point that validates your assumption through the hypothesis.
For example, if your hypothesis is about customer satisfaction, your minimum criteria could be a Net Promoter Score (NPS) above 40 or a Customer Satisfaction Score (CSAT) above 80%.
These metrics become your guiding north star throughout the experiment.
The real purpose of an MVP is to test worthiness, not technical feasibility
The main idea behind building an MVP is not to prove that you can build the product technically. It is to find out if it is even worth the effort to build it.
Many teams fall into the trap of building elaborate prototypes that demonstrate technical capability but do not answer the core question: will customers actually use and pay for this product?
An MVP experiment focuses on learning whether the product solves a real problem and whether your assumptions hold true.
Plan and execute the MVP experiment with diligent data collection
After you have done your due diligence — listening to customers, prioritizing assumptions, building hypotheses, and setting success metrics — you need to plan and execute your MVP experiment.
Planning involves setting up a data collection plan for each minimum criterion for success. For example, if customer satisfaction is your metric, you must decide:
- What data will you collect?
- Where will the data come from?
- How much data do you need to collect for statistically meaningful results?
This plan ensures that when you launch the MVP, you gather the right evidence to validate or invalidate your hypotheses.
Collect feedback continuously and iterate
Once your MVP is in the market, early adopters will begin providing feedback. The best way to collect feedback is through surveys using tools like Qualtrics, SurveyMonkey, or Google Forms.
Other methods include live chat sessions or customer calls, but traditional calling is often less effective at scale.
As feedback and data flow in, you will validate assumptions and discover new problems and opportunities. Product development is a continuous learning process — each iteration surfaces new insights that refine your product and hypotheses.
The MVP experiment timeline and expectations
MVP experiments typically run between two and twelve weeks, depending on the scale and sample size.
Most experiments take four to eight weeks to yield meaningful data. You should not expect immediate results.
Early iterations may not go as planned. Data collection methods can be imperfect, and assumptions may be wrong. But after a few cycles, your MVP experiments will become more reliable and informative.
Example: How Zappos validated online shoe buying
Zappos started as a simple MVP. Nick Swinmurn took photos of shoes from local stores and uploaded them online. When customers placed orders, he manually bought the shoes from the stores and shipped them.
This MVP allowed Zappos to validate the assumption that people would buy shoes online without investing in inventory or complex infrastructure upfront.
Only after validating this demand did Zappos invest in building a full-fledged e-commerce platform.
Example: MVP as a basic building in a land-shortage town
Imagine a town where residents complain of land shortage. You suggest building high-rise buildings with multiple apartments but do not know if people will buy these apartments.
You build a simple building with basic amenities — an MVP. Early adopters move in, validating the assumption that residents are open to apartment living.
Over time, residents provide feedback requesting additional features like a gym or swimming pool. You add these features iteratively, moving from MVP to a final product.
The scientific mindset: experiment, learn, iterate
Running an MVP experiment is about managing risk through hypothesis testing. It requires a scientific mindset:
- Identify assumptions and frame testable hypotheses.
- Define measurable success criteria.
- Design and run experiments to collect relevant data.
- Analyze results to validate or invalidate hypotheses.
- Iterate based on learnings.
This approach helps you avoid costly mistakes and build products that customers truly want.
FieldExercise title="Plan your MVP experiment" time="20 min"
Pick a product idea you are interested in or currently working on. Follow these steps:
- Write down the top three assumptions you have about your customers and their needs.
- For each assumption, write a testable hypothesis with a measurable outcome.
- Define the minimum success criteria (metrics) for each hypothesis.
- Outline how you will collect data to validate these metrics.
- Estimate the timeline for your MVP experiment (2-12 weeks).
Use this plan to guide your MVP experiment execution.
Test yourself: MVP experiment at a Bangalore early-stage startup
You are the PM at a seed-stage SaaS startup in Bangalore building a productivity app for small businesses. You have identified that your riskiest assumption is that small business owners will pay for a task management tool integrated with WhatsApp. You have 8 weeks and a small engineering team to run an MVP experiment.
The call: How do you design your MVP experiment? What hypotheses do you test, what metrics do you set, and how do you collect data?
Your reasoning:
Where to go next
- If you want to learn how to frame strong hypotheses: Hypothesis-Driven Product Discovery
- If you want to master user research techniques: User Research Methods
- If you want to understand metrics and analytics for experiments: Metrics and KPIs
- If you want to learn about continuous product discovery: Continuous Discovery Habits
- If you want to practice prioritization for MVP scope: Prioritization Frameworks