By Mehdi HassanHow can a landing page validate a startup idea?
A landing page validates a startup idea by measuring whether the right people take a meaningful action when shown a clear description of the product, before the product exists. It's often called a smoke test or a fake door test: you put the promise in front of real buyers and see whether they walk through the door.
The page doesn't prove the product will succeed. It tests one thing: whether the problem and the promise are strong enough that your target buyers will spend something, whether attention, time or money, to get it. That's a narrow question, but answering it cheaply before building is worth a lot. This approach sits in our cluster on pre-launch and demo pages, alongside the waitlist landing page guide and the book a demo page hub.
What conversion numbers actually count as validation?
The conversion numbers that count as validation are the ones you committed to before launch, measured on the action that costs the visitor most. There's no universal "good" smoke test rate, because it depends on the action you ask for, the traffic source and the price point. What makes a number meaningful is that you defined it in advance.
Signal | What the visitor gives | Strength as evidence |
|---|---|---|
Page visit | A click | Almost none: tests the ad, not the idea |
Email sign-up | An address | Weak: curiosity is cheap |
Sign-up plus answered questions | Time and information | Moderate: shows real interest |
Booked call | 30 minutes of their week | Strong: they want to talk about it |
Pre-order or deposit | Money | Strongest: they want to buy it |
For a rough sanity check on the page itself, Unbounce's benchmark report puts the median landing page conversion rate at 6.6% across industries and about 3.8% for SaaS. Those figures cover all sorts of offers, not smoke tests specifically, so don't treat them as a pass mark. But a page that converts cold, well-targeted traffic far below those medians on a simple email ask is telling you something.
How do you validate a startup idea with a landing page, step by step?
Run a landing page validation in seven steps: write the hypothesis, set the lines, build the page, add a costly action, drive matched traffic, talk to the people who convert, and decide. Skipping the first two is the most common reason smoke tests produce no clear answer.
Write one hypothesis
"Finance managers at 50 to 200-person companies will book a call to fix month-end reporting." One buyer, one problem, one promise. Testing three audiences on one page muddles the result.
Set pass and fail lines in advance
Decide what result means build, what means rethink and what means stop. Write it down and share it with a co-founder or advisor so you can't move the goalposts later.
Build a page that describes the product as real
A specific headline, three outcomes, a product mock-up and a clear price or price range. Vague pages produce vague data.
Ask for a costly action
A booked call, an application with questions or a refundable deposit. Keep an email option as a fallback so you can count both.
Drive traffic that matches your buyer
Targeted ads, outreach to people in the role or posts in communities where they gather. Friends and a general launch audience inflate the numbers.
Talk to everyone who converts
Call the people who booked or applied. Their reasons tell you more than the rate does.
Decide using your lines
Compare the result to what you wrote down in step two. Build, rethink the promise or stop.
Hypothetical example: setting the lines for a smoke test
For example, suppose a founder plans to send about 1,000 targeted visitors to a smoke test page for an expense tool. Before launch, they write: "Build if 20 or more people book a call and at least half of those calls confirm the problem is urgent. Rethink the promise if we get 5 to 19 bookings. Stop if fewer than 5."
The founder then runs the test, gets 12 bookings and learns on the calls that the real pain is approvals, not receipts. The result lands in the rethink band, so they rewrite the page around approvals and run it again. These numbers are illustrative only; the point is that the lines were written before the data arrived.
What should a smoke test landing page include?
A smoke test landing page should include everything a real product page would, at a smaller scale: a specific promise, how it works, who it's for, a price signal and one clear action. If the page feels like a placeholder, visitors treat it like one.
- Headline: the outcome for one buyer. Avoid category words like "platform" or "AI-powered" with no outcome.
- How it works: three steps or three outcomes, with a product mock-up. A tidy design in Figma is enough.
- Price signal: a real price or a starting price. Without it, you're testing interest in a free product.
- Founder line: who's building it and why. Early buyers back people.
- One action: book a call, apply or reserve. Name it clearly on the button.
Be honest about the stage. Saying "We're opening to the first 20 teams this spring" is fine; pretending the product is shipping today when it isn't damages trust with the very people you want as first customers. Our collection of waitlist page examples shows how well-known startups framed early access.
Should you A/B test a validation landing page?
Usually not, at least not in the classic sense. Most validation pages don't get enough traffic for a reliable A/B test: detecting a 15% lift at a 2% baseline takes roughly 15,000 to 20,000 visitors per variant, according to Spotler's guidance on A/B test sample sizes.
Instead, run tests in sequence. Test one promise to one audience, read the result against your lines, then change one big thing, such as the audience or the core promise, and run again. Small tweaks like button color won't show up at this traffic level. Our guides to A/B test ideas and headline testing cover what to test once you have the traffic, and scroll depth can tell you whether people are reading far enough to see the offer.
What do you do after the test?
After the test, act on the decision you defined in advance: build if you passed, rewrite the promise or audience if you landed in the middle, and stop if you failed. The hardest part is accepting a fail, but a clear no after a few weeks is far cheaper than finding out after a year of building.
- Passed: invite the people who converted to a pilot. Turn the smoke test page into a waitlist page and keep collecting answers.
- Middle band: use what you learned on the calls to change one big variable, then run the test again with fresh traffic.
- Failed: check that the traffic matched your buyer. If it did, the problem or the promise isn't strong enough. Move on or rethink the problem.
When you do build, the page work carries over. The same promise, proof and price become your product launch landing page. If you're building for AI buyers, our guide to AI landing page design covers how to show proof rather than claims.
