Fake door test examples: 3 real landing pages and their numbers
Updated
A fake door test puts a buy button in front of people before the product exists, and counts who clicks. Most articles about it describe the idea. This one shows three fake doors we are running right now, with the real numbers so far, and the rules we use to decide what happens next.
What a fake door test is
You describe a product as if it exists, give it a real price and a buy button, and send people to the page. When someone clicks buy, you tell them honestly that it is not ready yet and offer to notify them. No payment is taken.
The point is that a click on a priced button is a stronger signal than a survey answer or a like. People say yes to ideas all the time. Fewer reach for their wallet.
A well-known early case is Buffer: before building the product, the founder put up a page describing it with a link to plans and pricing, and people who clicked through to a plan learned it was not ready yet and could leave an email. Paying interest came before code.
Example 1: a Pro tier on top of a free tool
Page: PNG to DXF converter. The free converter is real and works: drop an image, get a DXF for a laser cutter. Below it sits a Pro offer for people who cut a lot: batch conversion of up to 50 images, a centerline mode for engraving, kerf offset, and true arcs and splines. Price: $19, one payment.
The Pro features are not built. The page tests whether people who already use the free tool would pay for more. The free tool brings the traffic; the Pro button measures intent.
Why this shape works: you are not asking strangers to imagine a product. They have just used it.
Example 2: the whole page is the door
Page: Match Style, a Claude Code skill that keeps new UI matching an existing app. There is no free part. The page explains the problem, shows a small before and after test, and offers the skill pack for $19 one-time.
This is the classic fake door: one page, one offer, one button. It is the purest test of demand and also the hardest to get traffic to, because there is nothing free to share.
Example 3: a paid service on an existing landing page
Page: the Idea Check section on the IdeaPanner home page. Next to the free idea checker there is a done-by-hand option: $19 for a full report within 48 hours, or $5 for a quick check of the core keyword and a one-line verdict.
This tests a different question: would people who already use a free tool pay to have a person do it for them? Two price points on the same page also show which one people reach for.
The real numbers so far
All three pages went live on October 24 and 25, 2026. After about five days:
- 42 visitors across the three pages.
- 1 click on a buy button.
- 0 email signups.
The honest reading: these numbers say nothing about demand yet. 42 visitors is too few to judge any conversion rate. The bottleneck is traffic, not the page. That is why guides like this one exist: to get the pages in front of people who are searching for exactly this.
Decide the rules before you look at the data
The most useful part of a fake door test is writing down the decision rules on day one. Otherwise every result looks like a reason to keep going. Ours:
- Decide only after 28 days.
- Fewer than 30 visitors at the deadline: extend 14 days. Still under 30 at 42 days: close the page.
- Buy-click rate under 2%: kill it.
- Between 2% and 5%: talk to people before building anything.
- Over 5%: keep going.
- At most 3 validations running at once.
How to build your own fake door page
- Pick one offer for one kind of buyer. Not a product line.
- Write the page as if it ships: what it does, who it is for, what it costs. Use a real price, not 'free beta'.
- Make the buy button do something honest: a short message that it is not ready yet, and an optional email field.
- Count three things separately: page views, buy clicks and emails.
- Plan the traffic before you build the page. Search, communities, a free tool next to the offer. Without visitors the test never ends.
- Write the decision rules down and stick to them.
A lesson from an earlier product
Before these tests, I ran a resume tool for 8 weeks: 23 signups, 0 paid. In one stretch, 6 of 7 signups reached the checkout page and none paid. Looking only at 'signups to paid' hid where it broke.
Two lessons went into the fake doors above: ask for money early, and read the funnel one step at a time (visit, click, email, pay) rather than as one number.
Questions
Is a fake door test ethical?
It is if you are upfront right after the click: say the product is not ready, take no payment, and let people leave an email if they want. Do not collect card details for something that does not exist.
How many visitors does a fake door test need?
We use 30 visitors as the minimum before judging anything, and 28 days as the earliest decision date. Fewer than that and one click swings the rate wildly.
What is a good click rate on a fake door?
Our rule: under 2% kill, 2 to 5% talk to people first, over 5% keep going. Set your own thresholds before launch, not after.