Which MVP Test Should You Run First? A founder’s decision flow for manual, no-code, landing-page, and paid-ad experiments

Start by scoring your idea on four dimensions: cost to run the experiment, time-to-signal, data precision (how directly the test predicts real customer behavior), and engineering/hiring risk (how much the experiment requires building long-lived tech). Use these dimensions to decide which test reduces your riskiest assumption fastest.

Quick definitions for the tests we compare: Manual/Concierge MVP = you deliver value personally or via manual processes; No-code MVP = a functioning prototype built with tools like Bubble or Webflow; Landing-page (smoke test) = marketing page that collects interest or pre-orders without a working product; Paid-ad experiment = drive targeted traffic to a landing page or sign-up flow to measure willingness-to-click or pay.

Rule of thumb: when your riskiest assumption is about problem–solution fit and messaging, start with a landing page or interviews. When the assumption is about workflow or value delivery that can be simulated, start with a concierge/manual MVP. When you need interaction data or product behavior, prefer no-code. Paid ads are best used when you already have a clear user profile and messaging to test scalable demand.

This framework avoids the common trap of choosing a test because it’s “cool” (no-code) or because you want a neat product. Choose the cheapest, fastest test that answers the riskiest question.

Step 1 — identify the riskiest assumption. Is it (A) no one cares about the problem, (B) messaging/messaging-to-conversion, (C) whether you can deliver value operationally, or (D) that people will use an interactive product feature? Each maps to a different first experiment.

Step 2 — apply the 4-dimension filter. If the riskiest assumption is A (demand), run a landing-page smoke test with clear CTAs and an email/‘pre-order’ capture. If it’s B (messaging), iterate landing pages or ad copy quickly and measure conversion. If it’s C (delivery), run a concierge/manual MVP: onboard customers yourself and deliver manually so you learn the workflow. If it’s D (product interaction), build a small no-code prototype that reproduces the core interaction.

Step 3 — include escalation rules. If a landing page gets traffic but zero conversions, stop and revisit messaging and target profile. If a concierge MVP scales to several paying customers but is too time-consuming, transition to a no-code prototype. If paid ads perform well (positive CAC signal), plan a product-focused test that collects retention/use metrics.

This flow prioritizes minimal cost and maximum learning: measure what matters and only build when the data says the problem and solution are real.

Landing pages: lowest engineering cost and fastest to signal. A well-targeted landing page can test willingness-to-click or pre-order with minimal build time, but clicks are a noisy proxy for real payments — use a clear CTA (paid booking or a refundable voucher) to increase signal quality.

Concierge/manual MVP: higher founder time cost but excellent data precision about real willingness-to-pay and how the product must behave operationally. It exposes hidden implementation costs and customer support work you’d otherwise discover late.

No-code MVP: sits between manual and built product for both cost and signal. No-code can produce usable interactions quickly, but vendor lock-in, scaling limits, and hidden integration work are real trade-offs. Use no-code when the core behavior requires an interface.

Paid ads: fastest way to scale traffic and to test demand at volume, but expensive if your targeting or messaging is immature. Ads are best used after you’ve polished messaging or when you need to test market size quickly; otherwise you risk paying for low-quality signals.

Scenario A — B2B workflow automation for SMBs with unknown buyers: Start with founder-led discovery calls plus a concierge MVP to deliver the service manually. B2B buying cycles and complex workflows make direct interviews and hands-on delivery the fastest way to learn whether you can meet the customer’s needs.

Scenario B — Consumer marketplace where demand is the open question: Launch a landing page that explains the marketplace and captures emails or deposits, then run small paid-ad tests to validate traffic channels only after messaging is dialed. Marketplaces need both demand and supply signals — use separate landing pages for each side.

Scenario C — Consumer app with a new interactive UX (e.g., novel editor or workflow): Build a focused no-code prototype that recreates the interaction and pair it with usability sessions. If the prototype shows early engagement, follow up with a small paid-ad test to measure scale potential.

Scenario D — Niche SaaS where pricing and packaging are the unknowns: Combine a landing page with pricing options and A/B test copy; drive a few sales using founder outreach or concierge fulfillment to validate willingness-to-pay before building a full product.

The worksheet is a 6-row checklist you can complete in under 10 minutes. Fill it out and add scores 1–5 (1 = low, 5 = high): 1) Problem clarity (do you know the user and pain?), 2) Need for interaction (does value require an interactive product?), 3) Willingness-to-pay uncertainty, 4) Channel clarity (do you know how to reach users?), 5) Operational delivery risk (can you deliver manually?), 6) Budget for ads/build.

How to interpret scores: if Problem clarity ≤3 and Channel clarity ≤3 → start with interviews and a landing page. If Need for interaction ≥4 → prioritize a no-code prototype. If Operational delivery risk ≥4 → run a concierge MVP. If Budget for ads ≥4 and messaging is known → run paid-ad experiments to measure scalable demand.

Quick checklist to use after scoring: pick the test that addresses the highest-scoring uncertainty; set a time/budget cap (e.g., 2 weeks and $500) and one primary metric (conversion to paid, retention after 7 days, demo booked). Stop if the metric is below your predefined threshold and iterate the riskiest assumption next.

Use this worksheet as your guardrail: it forces you to choose the experiment that minimizes cost while maximizing the chance to falsify the riskiest assumption.