Startup Idea Validation: 4 Tests to Run Before You Build Anything

Startup idea validation content online usually stops at “build a landing page and see who signs up.” That’s a start, but it only measures interest. Not willingness to pay, not real usage behavior, and not whether the problem is urgent enough for someone to change what they’re currently doing. A more complete framework runs through four escalating tests, each one requiring more commitment from the prospective customer than the last: a smoke test, a concierge MVP, a Wizard-of-Oz MVP, and finally a scored go-or-kill decision. Run all four before you commit real build resources to a single feature.

Step 1: The Smoke Test

Build a landing page describing the finished product as if it already exists. Add one clear call to action—join the waitlist, pre-order, or request access. Keep it tight: one headline naming the outcome, three to five benefit bullets, and a single CTA. Skip the feature dump entirely; focus on the result the buyer wants most.

Industry benchmarks put average cold-traffic landing page conversion at 2–5%. A 5%+ email signup rate on cold traffic is a genuinely strong signal. Under 2% usually means the positioning, audience, or messaging is off. Sample size matters here: fewer than 100 targeted visitors makes any conversion number statistically unreliable, so don’t conclude a handful of clicks.

Step 2: The Concierge MVP

Before automating anything, manually deliver the outcome yourself to your first 5–10 customers. If you’re building an AI-powered scheduling tool, manually do the scheduling behind the scenes. If you’re building a marketplace, manually match buyers and sellers by hand. This step validates that people want the result, not just the pitch, and it surfaces the operational edge cases your eventual product will need to handle, which no landing page or survey can reveal.

This is also where you get your clearest read on pricing. Asking someone if they’d pay $50 a month for a hypothetical product produces weak, unreliable data. Actually charging them $50 for a manually delivered version of the same outcome produces real data, because now there’s money on the table.

Step 3: The Wizard-of-Oz MVP

Let the product appear automated to the customer, while a human executes the process manually behind the scenes. This tests willingness to pay and real usage patterns at a fraction of the cost of building the actual automation. It’s the step most founders skip entirely, jumping straight from an idea to full development because it feels more “real.”

The Wizard of Oz step matters because it reveals usage frequency and stickiness under closer-to-real conditions, without the sunk cost of a finished engineering build. If customers stop using the manually powered version after a week, that’s a signal worth having before you spend months building the automated version of the same thing.

Step 4: Score and Decide

Set your pass/fail threshold before you look at the results, not after. This single habit avoids the most common validation failure of all: rationalizing weak data because the team is already emotionally invested in the idea. It’s far easier to hold a threshold you set in a neutral moment than to invent one after you already know the numbers disappoint you.

A structured research phase exists specifically to force this evidence-gathering ahead of any capital commitment. IGF’s process starts with a case study and market research before MVP scope is even defined, which is the same sequencing this four-step framework is built around.

Common Mistakes That Undermine Startup Idea Validation

A few patterns quietly invalidate otherwise well-run tests. Counting cheap clicks as demand is one. Changing the offer, audience, or messaging mid-test is another; it makes the results impossible to interpret cleanly. And relying on friends, family, or fellow founders for feedback is a third. They’re supportive by default, and that supportiveness isn’t the same signal as a stranger’s money.