How to Test a Business Idea Before You Build It
You can test a business idea in a weekend without building the product. Put up a one-page site, run a fake-door test, ask for a pre-order, or deliver the service by hand to your first few users. Each test forces people to make a real commitment, an email, a payment, or a clear “yes,” and that commitment is your evidence. If nobody commits, you just saved yourself months of building the wrong thing.
Most people get this backwards. They spend three months building an app, launch it, and then find out if anyone wants it. That’s the most expensive way to learn a lesson you could have learned in two days for zero dollars.
Why testing beats building
The point of a pre-build test is to buy information cheaply. Building is slow and expensive. A test is fast and nearly free.
Here’s the trap. When you build first, you get emotionally attached. You’ve sunk weeks into the code, so when users don’t show up, you tell yourself the marketing is wrong, or the timing is off, or you just need one more feature. You keep spending to avoid admitting the idea didn’t land.
A pre-build test removes the sunk cost. You risk a weekend, not a semester. If it fails, you shrug and try the next idea. That emotional distance is exactly what lets you read the results honestly.
This is different from an MVP, and people mix them up. An MVP is the smallest real version of your product that a customer can actually use. A pre-build test usually has no product at all, it measures demand before you write a line of code. If you’re fuzzy on the difference, read the plain-English definition of an MVP first, then come back.
What counts as a real “pass”?
A pass is a commitment, not a compliment. This is the single most important idea on this page, so read it twice.
When you describe your idea to a friend, they’ll say “that’s cool, I’d totally use that.” That is worthless data. It costs them nothing to say, and they’re being nice because they know you. Compliments are the noise that drowns out the signal.
A commitment has a cost. It’s an email address, a credit card number, a pre-order, a scheduled call, a click on a “Buy” button, or someone handing you actual money. Costly actions are honest because people don’t spend time or money on things they don’t want.
So before you run any test, write down the exact action that counts as a pass and the number you need to hit. “10 email signups from strangers this week” is a pass condition. “People seemed interested” is not.
The five cheapest ways to test an idea
Here are the five tests worth knowing, from lightest to heaviest. You don’t need all of them. Pick the one that matches the question you’re actually trying to answer.
| Test | What it is | What a pass looks like |
|---|---|---|
| Smoke-test landing page | A one-page site that describes the product as if it exists, with a single call to action | Strangers who never met you enter their email or click “Get started” at a rate you set in advance |
| Fake-door test | A button or feature link for something that isn’t built yet; clicking it shows “coming soon” | A meaningful share of visitors click the fake button, proving they want that specific thing |
| Pre-sale | You take real orders or deposits before the product exists, then build it | People pay money, or put down a deposit, before they can use anything |
| Concierge | You deliver the service manually, by hand, for a few users, with no software | Users keep coming back and, ideally, keep paying for the hand-done version |
| Wizard of Oz | The front end looks automated, but you’re doing the work behind the scenes by hand | Users complete the full flow and get value, not knowing a human did the work |
How does a smoke-test landing page work?
A smoke-test landing page is a single web page that presents your product as if it’s real and asks the visitor to take one action. No product exists behind it. You’re testing whether the promise is compelling enough to make a stranger raise their hand.
You describe the offer, show one clear button (“Join the waitlist,” “Get early access,” “Buy now”), and send traffic to it. Then you measure how many visitors take the action versus how many just leave.
The catch is traffic quality. If you only send friends, everyone signs up to be supportive and you learn nothing. Send strangers, people from a relevant subreddit, a school Discord, a niche community, so the signups are honest. Our guide to building a landing page that converts walks through the structure that turns visitors into signups.
What is a fake-door test?
A fake-door test puts a button for a feature that doesn’t exist yet, so you can count how many people try to open it. When someone clicks, they hit a “coming soon” message or an email capture instead of the real thing.
This is how you test one specific feature or offer without building it. Add a “Download the PDF” button, or a pricing tier you haven’t made, and watch the click rate. Clicks are the commitment. If almost nobody clicks, that door isn’t worth building.
Be considerate with it. A short, honest “we’re building this, want to be notified?” turns the dead end into a signup instead of leaving people annoyed.
How do you pre-sell something that doesn’t exist?
You ask people to pay before you build. A pre-sale is the strongest demand signal there is, because money is the most expensive commitment a person can make.
This sounds scary and slightly dishonest at first. It isn’t, as long as you’re upfront. Tell people it’s a pre-order, give a rough delivery window, and offer a no-questions refund if you don’t ship. Crowdfunding runs on exactly this logic: pay now, receive later.
If people won’t pre-order at any price, that’s a hard, useful truth. It usually means the problem isn’t painful enough or the price is wrong. To figure out the number, read how to price your first product before you set a pre-sale price.
What is a concierge test?
A concierge test means you deliver the service by hand, personally, to a small number of users, with no software at all. You do the work a machine would eventually do, so you can learn what the product actually needs to do.
Say you want to build a meal-planning app. Instead of coding, you personally text five people a custom plan every Sunday. You learn what they want, what they ignore, and whether they’ll keep paying, all before you build anything. The work doesn’t scale, and that’s fine. You’re learning, not scaling.
Doing the work by hand puts you close enough to hear the real problems, so keep notes on every request and complaint. Those notes become your spec when you finally build.
What is a Wizard of Oz test?
A Wizard of Oz test looks automated to the user, but a human is running the machinery behind the curtain. The user sees a normal app; you’re manually doing the work they think software is doing.
It sits between concierge and a real product. The user experiences the finished flow, so their feedback is realistic, but you haven’t built the hard automation yet. If the manual version delights people, you know the automated one is worth building.
Which test should you run first?
Match the test to the question you’re stuck on.
- “Does anyone even want this?” Start with a smoke-test landing page or a fake-door test. They’re the cheapest and answer demand fastest.
- “Will people actually pay?” Run a pre-sale. Interest is easy; money is the real vote.
- “What does the product need to do, exactly?” Run a concierge test so you learn the details by hand.
- “Will people use the full experience?” Run a Wizard of Oz test once you know there’s demand.
You can also chain them: landing page to confirm demand, then a pre-sale to confirm willingness to pay, then concierge to nail the details. That’s the whole first week of the Validate sprint at batch0, where you run these tests on your own idea with feedback before you build anything.
Common mistakes that fake a pass
A test only works if you don’t cheat it. These are the ways teen founders fool themselves.
- Counting compliments as commitments. “Sounds cool” is not a signup. Only count the costly action.
- Testing on friends and family. They’re biased. You need strangers who owe you nothing.
- Moving the goalposts. Set your pass number before you look at results, or you’ll always find a reason it “kind of worked.”
- Leading the witness. Don’t say “you’d pay for this, right?” Show the button and stay quiet. Let them act.
- Testing five things at once. One test answers one question. Muddle them and you can’t tell what caused the result.
- Quitting after one weak signal. A dead landing page might be a bad headline, not a bad idea. Change one thing and rerun before you kill it.
If you’re not even sure the problem is real yet, back up a step. Tests measure demand for a solution, but only if you’ve found a problem worth solving in the first place.
How long does this take, and what’s next?
A single pre-build test takes a weekend to a week. A landing page can go up in an afternoon. A concierge test runs for a week or two so you see whether people come back. None of it requires code or money beyond a few dollars for traffic.
The mindset to keep: you’re not trying to prove your idea is good. You’re trying to find out if it’s real, as fast and cheaply as possible. A fast no is a gift, because it frees you to find the yes.
Once a test passes and you’re confident someone will pay, you’re ready to build the smallest real version. Head to the full guide to validating a startup idea to put these tests into a repeatable loop.