
Prototype or MVP: 3 Formats to Validate Your Idea
The most expensive question a founder can skip is: “does this actually need to be built?” Before hiring a development team, it’s worth deciding between a prototype or MVP (minimum viable product) — and within MVP, there’s more than one path to choose from.
The short answer: pick the simplest format that can test the hypothesis actually at stake. If the open question is about the problem or the usage flow, a prototype is enough. If the question is whether someone will pay for the offering, a concierge MVP — with a manual process behind the scenes — settles it without writing a single line of production code. Only when technology itself is part of the value proposition does a functional MVP make sense.

Prototype or MVP: what’s the difference between the three formats
An MVP is defined as an early, functional version of a product capable of demonstrating the value proposition, testing critical hypotheses, and being evaluated in a real or representative environment (Finep, 2026). That definition already points to the most common mistake: treating an MVP as a synonym for “a simplified version of the final product,” when it’s really a testing instrument.
There are at least three formats, each serving a different purpose:
- Prototype: a clickable or visual representation of the product, with no real functionality behind it. It’s used to test whether the problem is relevant and whether the intended usage flow makes sense to the people who’ll actually use it.
- Concierge MVP: the delivery happens, but manually, with the team doing the work that would later be automated. Running it manually keeps the team close to users and clarifies what actually needs to become a system (BNB, 2019).
- Functional MVP: a version with real technology running, even if scope is limited. This is the format Finep recognizes as the entry point into the validation stage, as opposed to the traction stage, when a startup is already selling with consistent growth (Finep, 2026).
In our view, the choice between prototype or MVP shouldn’t be based on how much the founder has already decided about the product, but on which question is still unanswered. Skipping steps — for example, building a functional MVP when the real question is “does this problem even exist?” — is the most common way to waste a development budget.

5 criteria for choosing between prototype, concierge, and functional MVP
These criteria help you decide which format tests the right hypothesis with the smallest possible investment.
- What’s the critical hypothesis right now? If it’s still unclear who the customer is and whether the problem is relevant, the discovery phase should come before any code, investigating these questions directly with your audience (MCTI, 2025). A prototype or a structured interview handles this.
- Can the delivery be done manually? If so, a concierge MVP tests whether people will actually pay or use the product, without the team needing to build automation. Example: before coding a recommendation engine, the team can curate recommendations by hand for the first customers.
- Is the technology part of the value proposition? If the product’s edge is precisely its automation, speed, or technical integration, only a functional MVP can test that honestly. A clickable prototype won’t convince anyone that an instant payment system actually works.
- Is there a way to measure real behavior? Interviews, A/B tests, and usability tests all belong in a validation roadmap, but observed behavior still needs interpretation before becoming a product requirement (BNB, 2019). A landing page with a “buy” button that tracks clicks is already a valid test, with no need to build the checkout.
- How much can the team sustain manually? Concierge MVPs have a ceiling: they work well with dozens of users, not thousands. If validation requires scale from day one, the manual format isn’t viable, and a functional MVP should come sooner.
Concierge MVP: how it works in practice
The concierge MVP is often misunderstood as the “lazy MVP.” In practice, it’s the most disciplined format, because it forces the team to define precisely what’s being tested and what counts as success — without technology as an excuse.
Example scenario: a startup wants to validate a personalized product curation service. Instead of building a recommendation algorithm, the team serves the first ten customers manually, chatting by message and putting together selections by hand. If customers pay and recommend the service, the value hypothesis is validated — only then does automation become a conversation.
An MVP should allow for structured feedback collection, iterative adjustments, and early evidence of market validation; an isolated concept isn’t enough on its own (Finep, 2026). The concierge format delivers on that without requiring a systems development budget before it’s warranted.
Common mistakes with this format: treating the manual process as “temporary” without defining when to stop, or using concierge to validate a technology hypothesis (in that case, it simply doesn’t apply).
When a functional MVP is necessary
There are situations where jumping straight to a functional MVP is the right call, even though it’s the most expensive of the three formats.
- The value proposition depends on automation or technical integration. Example: a product that promises to reconcile data across multiple systems in real time can’t be convincingly simulated by hand.
- Problem and interest have already been validated. If interviews and prototypes have already confirmed demand, building the functional version is the logical next step, not a shortcut.
- The manual operation has become unsustainable. Once a concierge MVP has proven demand but the team can no longer keep up manually, it’s time to automate what’s been learned.
A project like ZapiFiq, a system developed by Kairen, illustrates this kind of milestone: when the product logic demands a real system running, there’s no visual shortcut that replaces it. The same applies to apps and web apps — see the Garage 665 case — where the real-world experience is part of what’s being tested, which also makes good UI/UX design relevant from the prototype stage onward.

When it makes sense to validate with a prototype or MVP — and when it doesn’t
Not every idea needs an MVP right away — and not every prototype actually qualifies as a test.
It makes sense when:
- The main hypothesis is still about the problem, audience, or value proposition — in that case, start with interviews, a clickable prototype, or a landing page.
- The solution can be delivered manually and the team needs to observe the process before automating it — use a concierge MVP with clear success criteria defined upfront.
It doesn’t make sense when:
- The prototype will only be shown internally, with no contact with real users or potential customers — that’s pitch material, not validation.
- The team wants to build a full platform before defining the hypothesis, the audience, a success metric, and a criterion for continuing, adjusting, or stopping the project.
On cost: there’s no reliable universal price for a prototype, concierge MVP, or functional MVP, because it depends on the industry, the hypothesis, and the chosen validation method. Any fixed number given without that context should be treated with skepticism.
Conclusion: the next step depends on the hypothesis, not on ambition
Choosing between prototype or MVP isn’t a matter of available budget, but of which question still has no answer. Start with the simplest format that forces a real test with people outside the team — and only move to full technology once it’s genuinely part of what you’re validating.
If your team already knows which hypothesis needs testing and it’s time to build the functional version, talk to Kairen at /en/contact to scope the right approach for this stage of your product.
Frequently asked questions
Prototype or MVP: where should an early-stage startup start?
Almost always with a prototype or interviews, since the initial doubt is usually about the problem and the audience, not the technology itself.
Does a concierge MVP count as an MVP for formal validation purposes?
Yes, as long as it delivers real value and allows for structured feedback collection — what matters is market evidence, not how much automation sits behind the delivery.
When should I move from a concierge MVP to a functional MVP?
When the manual operation becomes unsustainable at volume, or when technology itself becomes the part of the value proposition still left to test.
Is it worth skipping straight to a full system?
Rarely. Before investing in app development or full systems, it’s worth checking whether the critical hypothesis has already been tested with a simpler format — see also our post on when to build an MVP.