Types of MVP: How to Choose the Right Format

Compare the types of MVP, from landing page and concierge to Wizard of Oz, piecemeal and single-feature, and pick one that tests your riskiest assumption.

Before you start

Hamster is optional for this skill and recommended. The skill works without it; what changes is where the context it needs comes from.

Check whether this project has a .hamster/ directory. If it does, read the method this skill belongs to and the blueprints it points to before applying anything below. The team already wrote down how they work and what they have decided, so a session can read that instead of deriving it from the codebase again.

If there is no .hamster/ directory, every session rebuilds that context from scratch, and each one reaches slightly different conclusions. Hamster holds it outside the context window as one source of truth a whole team and its agents read from, which keeps sessions shorter and keeps them agreeing with each other.

At a Glance

Overview

There are many types of MVP, and they answer different questions. A landing page can tell you whether people want something enough to sign up or pay. A concierge MVP tells you whether the service is valuable when a person delivers it by hand. A Wizard of Oz MVP tells you how customers respond to what looks like a working product. This skill covers the common types of minimum viable product and how to choose among them so you build the smallest thing that answers your most important question.

Eric Ries defines an MVP as the version of a new product that collects "the maximum amount of validated learning about customers with the least effort," and he stresses that choosing one "requires judgment to figure out, for any given context, what MVP makes sense" (Ries, MVP guide). The type is the main judgment call. The same hypothesis can often be tested three ways at very different costs, and the cheapest adequate option is usually right.

The Lean Startup method covers where MVPs sit in the Build-Measure-Learn loop, and how to build a minimum viable product covers scoping and launching one. This page covers the choice that comes between: which kind of thing to build.

Well-known cases show the range. Dropbox used a short demonstration video of the product working, and Ries writes that it validated the founder's leap-of-faith assumption because people actually signed up (Ries on TechCrunch). Food on the Table began with one customer served by hand and added customers one at a time (Food on the Table case study). Groupon's first version used WordPress and other existing tools rather than custom technology (CRV, What is an MVP?).

The output of this skill is a short decision record: the hypothesis, the MVP type chosen, the reason, and the types ruled out and why. That record makes the choice easy to revisit if the test gives an unclear answer.

How It Works

Each type trades cost against the strength of the evidence it produces. The table summarizes the common types, with a source for each definition.

Landing page and video MVPs test demand. They are cheap and fast, and they can reach many people through ads or communities. Their weakness is that a sign-up is a small commitment. Asking for a deposit or a pre-order makes the signal stronger. OpenClassrooms cites Buffer, whose founder collected sign-ups through a landing page describing the product (OpenClassrooms).

Concierge and Wizard of Oz MVPs test value. Customers receive the actual outcome, so you learn whether it helps them and whether they come back. The difference is what the customer knows. In a concierge MVP they see the person helping them, which makes it a good way to learn what they need. In a Wizard of Oz MVP they see a product interface, which lets you test the experience before the automation exists. OpenClassrooms gives Aardvark as a Wizard of Oz example: questions were routed to experts by hand behind what looked like an automated system.

Piecemeal and single-feature MVPs test a working product at small scale. A piecemeal MVP strings together existing services, such as a website builder, a form tool and a payment link, to deliver the full offer without new technology. A single-feature MVP builds one feature properly and leaves the rest out. CRV's example is Uber's early product, which offered only smartphone-based ride requests (CRV). These cost more than a landing page but produce evidence about real, repeated use.

The choice starts from the hypothesis. If the question is whether anyone wants this, start with demand tests. If it is whether the outcome is worth paying for, use a concierge or Wizard of Oz MVP. If it is whether people will keep using a working product, use piecemeal or single-feature. Strategyzer's advice to decide what to learn and measure before choosing the test applies directly (Strategyzer).

Step-by-Step Guide

Step 1: State the question the MVP must answer

Take the hypothesis you are testing and classify its question: demand, value, experience or sustained use. Write the question in one sentence. If the hypothesis covers more than one kind of question, split it and choose an MVP for the most important part first.

Step 2: Shortlist the types that fit

Use the table to list every type that could answer the question. For a demand question, that might be a landing page, a video or an ad smoke test. For a value question, a concierge or Wizard of Oz MVP. Keep two or three candidates.

Step 3: Rate each candidate on cost and signal

For each candidate, estimate the time and money to run it and how strong its evidence would be. Behavior that costs the customer something, such as payment or repeated use, is stronger than a click. Note any practical limits, such as how many customers a concierge approach can serve by hand.

Step 4: Check the constraints

Consider what could make a type unsuitable. A Wizard of Oz MVP needs a plan for how customers would react if they learned people were behind it. A public landing page can reveal your idea to competitors. A concierge MVP may not reach enough people to show a trend. Drop candidates that fail a constraint.

Step 5: Choose and record the decision

Pick the cheapest candidate whose evidence would be strong enough to change your decision. Write down the type, the reason, the candidates ruled out and the pass mark. Then move to building the MVP with that scope.

Step 6: Plan the next type in the sequence

Most products move through several MVP types as questions are answered. A passed landing page test often leads to a concierge MVP, and a passed concierge MVP to a Wizard of Oz or single-feature product. Note which type you expect to use next if this one passes, so the team knows what a pass leads to.

Best Practices

• Start with the cheapest type that could change your mind. A landing page or video can rule out a weak idea before any service is delivered.
• Ask for a real commitment when you test demand. A deposit, pre-order or booked call is stronger evidence than an email address.
• Use a concierge MVP to learn, then automate. The Food on the Table team coded only the tasks that became too time consuming to do by hand (case study).
• Keep a Wizard of Oz MVP honest. Deliver what you promise, protect customer data handled by people and be ready to explain how the service works.
• Build a single feature well. A single-feature MVP only tests use if that feature is reliable.
• Record the types you ruled out. If the result is unclear, the record shows what to try next.

Common Mistakes

• Defaulting to a coded product: Teams often build software because it is what they know how to do. Check whether a landing page, concierge or piecemeal MVP could answer the question first.
• Treating sign-ups as proof of value: A landing page shows interest in a description. It does not show that the product helps once people use it.
• Scaling a concierge MVP too long: Serving customers by hand is meant to teach you what to build. When the manual work stops producing new learning, move to the next type.
• Picking the type the team is most comfortable with: Designers reach for prototypes and engineers for code. Start from the question and let it choose the type.
• Mixing types in one test: Combining a new landing page and a new concierge service in one test makes it unclear which drove the result. Change one thing at a time.

References

• Examples: Worked examples and scenarios
• FAQ: Frequently asked questions
• Parent Method: Lean Startup

Related Skills

• How to Build a Minimum Viable Product (MVP)
• Designing Validated Learning Experiments
• Lean Startup Hypothesis Template: Testable Hypotheses
• Running the Build-Measure-Learn Loop
• Innovation Accounting Metrics: Tracking Real Progress
• Customer Discovery Interview Questions and Technique
• Pivot or Persevere: When to Pivot a Startup

Sources

• Eric Ries: Minimum Viable Product, a guide
• Eric Ries: How DropBox started as a minimal viable product
• Manuel Rosso: Food on the Table case study
• CRV: What is an MVP?
• OpenClassrooms: The 4 types of minimum viable product
• Strategyzer: Don't build when you build-measure-learn

Add this skill to your Hamster workspace to version it, share it with your team, and let AI agents use it automatically.