MVP vs prototype vs landing page: choose the test your decision needs

By Dmitry Volopianskii — product entrepreneur, consultant and founder of Lectoroom. Prepared October 2, 2026.

MVP vs prototype is a choice about evidence. A landing page tests response to an offer. A prototype lets you observe how people understand and navigate a proposed experience. An operational minimum viable product, or MVP, lets intended users complete a real task. Choose the format that can answer the question holding up your next investment.

For a B2B founder, that question often involves more than one person. The employee using the product, the manager approving a pilot and the buyer authorizing continued spending may have different reasons to participate. A successful demonstration to one role does not settle the others' decisions.

This guide compares the formats through those roles. If you cannot yet identify a specific problem and audience, first validate the startup idea before building. The worked scenario below is hypothetical, not a customer result or a report of a completed experiment.

On this page: comparison · landing-page evidence · prototype tasks · operational MVP · B2B example · decision worksheet · FAQ.

MVP vs prototype: what does each format actually test?

A prototype represents a proposed product or workflow. It may be a paper sketch, linked screens or realistic code with simulated behavior. A landing page presents an offer and asks for an action. For the comparison below, an operational MVP is a bounded product or service that delivers a real outcome, possibly with human assistance.

These are working definitions, not an industry standard. Some teams call an early landing-page experiment their MVP. State the operating boundary in the brief so that a shared label does not conceal different expectations.

FormatUseful questionObservable evidenceImportant limit
Landing pageWill this audience take a next step toward this offer?Qualified inquiries, requested conversations, responses to an explicit proposalDoes not demonstrate successful use of the future product
PrototypeCan intended users make sense of this workflow?Task attempts, mistakes, expectations and assistance requiredSimulated success does not establish real delivery
Operational MVPWill users complete a real task under agreed conditions?Outcomes, return use when the need recurs, support effort and commercial decisionsA small assisted pilot does not establish scalable acquisition or economics

You can combine the formats. A landing page can recruit prototype participants, and a prototype can clarify a workflow during an active pilot. The artifact is useful when its observation changes a decision; completing a sequence of artifacts is not itself progress toward evidence.

What can landing page validation tell a B2B founder?

Landing page validation can show how a reachable audience responds to a specific promise and next step. It becomes harder to interpret when the audience, offer and requested commitment are vague. “Join the future of operations” and “Discuss a pilot for collecting supplier documents” invite very different responses.

Keep three records together: where the visitor came from, what they saw and what happened after they responded. If a page invites pilot applications, record whether each respondent fits the intended role and can describe a relevant task. A list of email addresses without that context leaves the original assumption largely unresolved.

What does the Buffer example actually show?

In Buffer's first-person launch account, Joel Gascoigne describes testing the idea with a small website, then adding a pricing step before building a functioning product. Visitors could select a plan and leave an email address. The account distinguishes those early actions from customers paying after launch. Buffer's original account.

The useful lesson here is the distinction between observations. Clicking a paid plan is evidence about response to a presented offer. It is not a completed transaction or proof that the product will retain customers. Buffer's historical outcome is not a conversion benchmark for another business.

For a B2B pilot, a respondent might be an enthusiastic employee with no permission to introduce software. That interest can help you understand a workflow, but it does not establish purchasing approval. Follow up with a concrete question: what would have to happen inside the company for this pilot to start?

Describe availability honestly. A pilot application should say that it is an application. If the product is not ready, avoid a page that implies immediate access and only reveals the waitlist afterward. Clear expectations also make the evidence easier to interpret.

How should you test a prototype with real users?

Give an intended user a realistic task, then observe how they try to complete it. A prototype session is especially useful when the uncertainty concerns interpretation: what a status means, which document belongs to which request, or what happens after a submission.

Nielsen Norman Group recommends task scenarios that are realistic, encourage action and avoid giving away interface instructions. That makes task design part of the test rather than an afterthought. Task Scenarios for Usability Testing.

For a supplier-document tool, compare these prompts:

  • Leading prompt: “Click Replace Document and choose the new file.”
  • Task prompt: “The certificate you sent is out of date. You have its replacement. Make sure the customer can review the current version.”

The second prompt lets you observe whether the user understands replacement, version history and review status. If the participant cannot proceed because the relevant part of the prototype is missing, record a prototype limitation. If they choose the wrong action despite the intended path being available, record the observed usability problem.

Keep prompted and independent completions separate. A facilitator who rescues every task can make an unclear workflow look usable. Assistance may be appropriate, but the report should preserve what the participant did before and after it.

A realistic coded prototype also needs a clear boundary. GOV.UK warns against treating prototype code as production-ready merely because it looks like the real service. For this decision, the implication is to review operational requirements separately from the demonstration. GOV.UK: Making prototypes.

When should you build an operational MVP?

Build an operational MVP when the next decision depends on real delivery: whether a document reaches the correct reviewer, whether a request is completed, or whether a team returns when another supplier needs onboarding. A simulation cannot establish those outcomes.

The pilot can be narrow. It does not need every integration or a self-service administration console if those capabilities are outside the question being tested. It does need enough reliability and support to fulfill the promise made to participating customers. Use those obligations to define your MVP scope.

Human assistance is compatible with a real service. Someone may set up an account or classify incoming documents manually. Record that work because it affects what you can conclude. A completed request with intensive founder intervention establishes a different operating condition from a completed request with no intervention.

If the hypothesis is specifically that the customer can finish without staff assistance, manual rescue cannot count as evidence that it passed. If the hypothesis concerns whether the delivered result is useful, assistance may be acceptable for this round. State which question you are answering before the pilot begins.

Does a technical proof of concept come first?

Only when a technical uncertainty blocks the relevant decision. A proof of concept, or PoC, can test whether a necessary integration or capability works under specified conditions. It does not automatically establish usability or customer demand. Conversely, a well-received prototype does not settle whether the underlying capability can work.

For an AI feature, a polished screen with prepared outputs cannot establish performance on the intended inputs. A separate evaluation of that capability may be the useful next step even while customer research continues. Keep its result separate from enthusiasm for the product concept.

How do user, sponsor and buyer signals change the choice?

Consider a hypothetical service that helps purchasing teams collect documents from suppliers. This is an illustrative planning exercise, with no measured results or implied client relationship.

There are several participants: a supplier submitting a document, a purchasing coordinator checking it and a manager deciding whether to sponsor continued use. A single “positive feedback” field would hide which person's uncertainty has been reduced.

Current uncertaintyFirst useful testObservation to captureDecision still unresolved
Coordinators recognize the problem and will discuss a changeFocused offer with a pilot applicationRelevant workflow, reason for inquiry and practical next stepSupplier participation and managerial approval
Suppliers understand how to replace an outdated documentPrototype task using sample materialCorrect version selected, expected status and prompts neededActual delivery and review in the customer's environment
The service helps coordinators finish a live requestBounded operational pilotRequest outcome, elapsed workflow time and every interventionWhether another team will adopt it under different conditions
A manager supports continued useExplicit continuation discussion with scope and termsDecision-maker, conditions, unresolved objections and actual decisionA repeatable sales process across other companies

Do not turn these rows into an automatic funnel. If an existing customer has already supplied a real workflow and authorized a pilot, another public landing page may add little. If suppliers cannot understand the proposed task, improving that interaction may matter more than collecting additional applications.

A rejected continuation proposal also deserves a precise interpretation. The problem may be valuable while the purchase is blocked by timing, deployment conditions or ownership. Record the reason and evidence rather than relabeling every objection as either “no demand” or “sales friction.”

For recurring workflows, define the observation window around the next genuine opportunity to use the service. Daily return rate would be a poor fit if the participant only needs to onboard a supplier occasionally. Record whether the need recurred before interpreting absence of use.

What should your test brief contain?

Strategyzer's Test Card asks teams to specify the assumption, test, measurement and success criterion. The worksheet below applies that discipline to a B2B build decision and adds explicit role and inference boundaries. Strategyzer: Validate your ideas with the Test Card.

FieldWrite down before building
Pending decisionWhat will you fund, change or stop after this test?
AssumptionWhat must be true for that decision to make sense?
Participant roleUser, coordinator, sponsor or buyer; why this role matters
ArtifactLanding page, prototype, operational pilot or technical test
Requested actionThe specific behavior, including any disclosed assistance
EvidenceEvents, observations and context you will preserve
Decision conditionsWhat would support proceeding, revising or stopping?
Alternative explanationWhat else could produce the same observation?
Unanswered questionWhat this test cannot establish
Owner and limitResponsible person, review point and resource boundary

Choose decision conditions before seeing the outcome. This does not require inventing a universal success percentage. An exploratory study may support a revised workflow rather than a market-wide estimate; say so in the brief. Report counts with their denominators and recruitment context when you have actual data.

Compare total experiment effort, including participant recruitment, preparation, support and analysis. A page built quickly can still be a poor test if you cannot reach the intended audience. An assisted pilot can be worthwhile if an existing customer already supplies the context and the result answers the next decision.

Frequently asked questions

Can a landing page be an MVP?

Yes, under a broad definition of MVP as an experiment for learning. A narrower operational definition reserves the term for a service that delivers a real outcome. Agree on the definition, then document what participants actually receive and what their actions establish. The label does not turn signups into evidence of successful product use.

Should every startup build a prototype before an MVP?

No. A prototype helps when the proposed interaction remains uncertain. If the workflow is already understood and the open question requires real delivery, a bounded pilot may be more useful. If a necessary capability is unproven, a technical test may take priority. There is no mandatory sequence that fits every product.

Does an MVP have to take payments?

Only if payment is necessary for the question you are testing. A working internal tool may first test task completion and operational fit. If the decision concerns willingness to pay, record a real commercial decision under clear terms. A pricing click, an informal expression of interest and a completed purchase are distinct observations.

How many signups validate an idea?

There is no universal count. Interpretation depends on participant fit, acquisition source, the offer and what happened after signup. A small exploratory sample can help choose another experiment without establishing market demand. Define the decision conditions in advance and retain uncertainty when the evidence cannot support a stronger conclusion.

What should you build next?

Resolve the MVP vs prototype decision by naming the observation you need. Use a landing page for response to an offer, a prototype for a proposed interaction and an operational MVP for a real outcome. In B2B work, record whose behavior you observed and whose decision remains open before committing to a larger build.

About the author. Dmitry Volopianskii is the founder and Product Lead of Lectoroom, which he has been building since 2013. His work includes product strategy, monetization and user journeys; see the Lectoroom product case study. The supplier-document example is independent editorial material, not a claimed Lectoroom result.

Sources and editorial notes

Sources checked October 2, 2026. Source-based methodology is linked at the relevant claim. The role comparison, decision worksheet and hypothetical scenario are editorial applications, not reported customer research. Future revisions should recheck sources and date any material corrections.