Idea Validation: How to Test a Business Idea Before You Build

Most founders think they've done idea validation when they've asked ten friends what they think. That's not validation. That's a popularity contest, and everyone you know is going to be nice to you.

Real idea validation looks different. It means a stranger gives you money, blocks time on their calendar, or hands over a real commitment before your product exists. If nobody's willing to do that, you don't have a business yet. You have an opinion with a landing page.

This guide walks through what idea validation actually is, five experiments ranked by how much signal they generate, and how to run each one cheaply this week. By the end, you'll have a decision rule for whether to build, adjust, or walk away.

What idea validation actually means

Idea validation is the process of testing whether people will take a real action for your idea before you spend months building it. Not "would you use this?" Not a survey. An action: a payment, a deposit, a signed-up time slot, a waitlist with skin in the game.

The distinction that matters is opinions versus commitments.

An opinion costs the other person nothing. "That sounds cool" costs nothing. A LinkedIn comment costs nothing. A friend telling you they'd "definitely use that" costs nothing, and it tells you nothing either, because they're speaking hypothetically about a future where the product magically exists and works perfectly.

A commitment costs something. Money is the strongest commitment signal. Time is close behind it (a scheduled call, a stated deadline they're waiting on). A public statement tied to their reputation counts too, though it's weaker than the first two.

Rob Fitzpatrick's The Mom Test built an entire framework around this gap. His core argument: people are bad at predicting their own future behavior, and worse, they'll lie to you about it because they want to be encouraging. The fix isn't better survey questions. It's asking about their past behavior and their actual life, then watching what they do next, not what they say they'd do.

If you've already narrowed down which idea is worth your time, you've done the filtering step. Idea validation is the next one: proving that specific idea gets a real commitment before you write a line of code.

The 5 validation experiments, ranked by signal strength

Not all validation methods are created equal. Some tell you a little. Some tell you almost everything. Here they are ranked from foundational to strongest, along with how to run each one in the next 7 days.

1. Customer-discovery interviews

What it is: structured conversations with 10-15 people who actually experience the problem you think you're solving, run before you mention your solution at all.

This is the floor, not the ceiling. Interviews alone won't prove anyone will pay you. What they will do is tell you whether the problem is real, how people currently cope with it, and what language they use to describe their pain. That vocabulary becomes your future landing page copy.

How to run it this week: find 10 people who match your target customer (not friends, not family). Ask about a recent specific instance of the problem, not a hypothetical future. "Walk me through the last time this happened" beats "would this be useful to you?" every time. Former Y Combinator CEO Michael Seibel's advice on how to get and test ideas makes the same point: talk to potential users before you build anything, and don't fall in love with a solution before you understand the problem.

Signal strength: low to moderate. Good for direction. Not proof of demand.

2. Landing-page smoke test

What it is: a simple page describing your offer, with a clear call to action (join the waitlist, get early access, pre-order), pointed at real traffic.

This is where opinions start turning into measurable behavior. You're not asking if people like the idea. You're seeing what percentage of visitors take an action when they don't know you and have nothing to gain from being polite.

How to run it this week: build a one-page site with Carrd describing the problem and your solution in plain language. Add one button. Drive a small, targeted batch of traffic to it (a relevant subreddit, a niche newsletter, a handful of cold outreach messages to your interview list). Track visitor-to-signup rate, not just raw signups.

Signal strength: moderate. A page that converts at 10%+ with cold traffic is a real signal. A page that converts at 1% with traffic you personally warmed up is not.

3. Fake-door or pre-order test

What it is: an offer that looks fully real, with a payment or deposit step, that stops the moment someone tries to complete it (fake door) or actually collects money for a product that doesn't exist yet (pre-order).

This is where the signal gets sharp. Someone entering payment information, even if you refund it immediately or the page tells them "coming soon," is behaving the way a real customer behaves. That's a different category of data than a survey response.

How to run it this week: set up a simple pre-order page with Gumroad and a specific price. Tell early sign-ups exactly what they're buying and when they'll get it. If someone pays, you have validation. If ten people abandon at the payment step after clicking "buy," you've learned your price point or your offer clarity is off, which is still useful information.

Signal strength: high. Money on the table is the closest thing to proof you can get before building.

4. Concierge or manual MVP

What it is: you personally deliver the outcome your product promises, by hand, for a small number of paying customers, with no software behind it.

This is the highest-fidelity test available before you build. You learn exactly what the customer values, what they'll pay, and where the actual work is hardest, because you're the one doing it. Eric Ries's Lean Startup methodology frames this as validated learning: the goal isn't to ship a feature, it's to run a real experiment and measure what customers do with it.

How to run it this week: find 3-5 people from your interview list who showed the strongest signal. Offer to deliver the result manually (a report, a matching service, a curated list, a done-for-you task) for a fixed fee. Do the work yourself, even if it's messy and doesn't scale. You're testing the value, not the delivery mechanism.

Signal strength: very high. Repeat customers at this stage are close to proof you have a business.

5. Paid pilot

What it is: a short, priced engagement with a real customer, using a rough version of your actual product or service, with a defined scope and a real invoice.

This is the strongest test because it combines everything above: a real problem, a real payment, and a real product being used under real conditions. If a customer pays for a pilot and renews or expands it, you're no longer validating. You're operating.

How to run it this week: take your concierge process and wrap a lightweight version of your actual product around it. Price the pilot to cover your time, not to make a profit yet. Set a clear start and end date so both sides know what "done" looks like.

Signal strength: highest. This is as close to a real customer relationship as you can get before scaling.

How to read the results

Numbers alone don't tell you anything. Context does. A 2% landing-page conversion rate is disappointing for a consumer app with broad reach and encouraging for a narrow B2B tool where even a handful of qualified leads matters.

Watch for these patterns:

• Real signal: people take the harder action, not just the easy one. A waitlist signup is easy. A deposit is hard. Weight your evidence toward the actions that cost something.
• False positive: enthusiasm without follow-through. Ten people say they'll join your pilot, two respond to the actual scheduling email. The other eight were being polite.
• False positive: a single loud advocate. One person who loves your idea and talks about it constantly can make a small sample feel like momentum. Look at the group, not the most vocal member of it.
• Real signal: unprompted repeat behavior. Someone who comes back to your concierge service a second time without you asking is telling you something a survey never could.

The decision rule: proceed, pivot, or kill

Before you run any experiment, decide in advance what result moves you forward and what result stops you. Set this before you see the data, or you'll rationalize whatever number you get.

A workable rule of thumb:

• Proceed: a meaningful share of a cold (not warmed-up) audience took the highest-friction action you offered, whether that's a deposit, a pre-order, or a signed pilot agreement.
• Pivot: people engaged with the problem framing but resisted the specific solution or price. Adjust the offer and test again with a fresh audience, not the same one.
• Kill: low engagement even from people who clearly have the problem, or engagement that never survives contact with a real ask for money or time.

None of these outcomes are failures. A killed idea after two weeks of cheap testing is a win compared to a killed idea after six months of building.

Start where the signal is

You don't need all five experiments before deciding anything. Start with interviews to sharpen the problem, then move straight into whichever of the higher-signal tests fits your idea best. If you haven't yet nailed down which idea deserves this effort, work back through business ideas worth pursuing and market research for your startup first. Idea validation works best on one idea at a time, chosen deliberately.

Once you've got a real signal, the next step is turning that validated idea into a plan that gets you to your first paying customer, not just a bigger deck.

EntraWorld gives you the structure to move through all of this without juggling five different tools. Join EntraWorld free and start testing your idea this week, not next quarter.