What a Fake Door Test Actually Measures
A fake door measures intent at the moment of action. Someone sees a button, a menu item, a pricing tier, or an ad, and they choose to click it expecting it to work. That click is worth more than a hundred survey responses because it costs the user something: their attention and a small bet that the feature is real.
What it does not measure is whether people will pay, stick around, or actually use the thing once it works. A click on 'Export to PDF' tells you there is appetite for export. It does not tell you the feature will retain anyone. Keep the claim small and the test stays honest.
- Good signal: a measurable click-through on a clearly labeled entry point.
- Weak signal: hovers, scroll depth, or time on page. People do those by accident.
- Not measured here: willingness to pay, retention, or actual usage.
Pick One Door and Label It Honestly
Test one feature at a time. If you stuff three fake buttons onto a page you cannot tell which one drove interest, and you dilute the traffic across all of them. Pick the single riskiest assumption you have about what users want next, and build a door for exactly that.
Label the door the way a real product would. If you are testing a 'Slack integration', the button says 'Connect Slack', not 'Click here for a surprise'. Vague or hyped labels inflate clicks from curiosity rather than intent, and curiosity is not a market.
- Use the literal feature name a user would search for.
- Place it where the feature would really live, not in a banner.
- Avoid words like 'new', 'beta', or 'coming soon' on the door itself. Those bias the click.
Build the Door and the Catch Screen
The door is the clickable element. The catch screen is what appears after the click. This is the part most people get wrong, so spend your effort here. When someone clicks, do not show a dead end or a 404. Show an honest interstitial that says the feature is in development and offers a way to stay in the loop.
The catch screen does two jobs. It converts the click into a follow-up (an email, a waitlist spot, a short question), and it sets up your second data point. The percentage of clickers who then leave their email is a far stronger signal than the click alone, because now they have spent something real on a feature that does not yet exist.
- Catch screen copy: 'We are building this now. Want early access when it ships?'
- Capture an email or run one question: 'What would you use this for?'
- Never charge money on a fake door. Collecting payment for a nonexistent product is fraud, not validation.
Drive Real Traffic, Not Friends
A fake door needs traffic that resembles your future customers. Your existing users are the gold standard because they already have context. Adding a fake menu item inside your live app and watching the click rate over a week is the cleanest version of this test.
If you have no audience yet, you can buy traffic with a small ad spend pointed at a landing page that contains the door, or post in a community where your target users already gather. The key is that the people clicking must plausibly be the people who would buy. A door clicked by ten random strangers tells you almost nothing.
- Best traffic: logged-in users seeing the door inside the product.
- Decent traffic: a targeted niche community or a small paid ad set.
- Useless traffic: your launch tweet, friends, or anyone you personally asked to look.
Set the Pass Bar Before You Look
Decide what a pass looks like before any data comes in, or you will rationalize whatever number you get. Pick two thresholds: a click-through rate on the door, and a follow-up rate on the catch screen. Write them down. A common shape is something like 'if more than a small but clear fraction of viewers click, and a meaningful share of clickers leave an email, I build it'. Set numbers that fit your traffic volume.
Watch the denominator. Click-through rate only means something against the number of people who actually saw the door, not total visitors who never scrolled to it. Instrument the impression, not just the page view, or your rate is fiction.
- Define click-through and follow-up thresholds in advance.
- Measure impressions of the door, not raw page views.
- Run it long enough to clear novelty. A spike on day one is not demand.
Read the Result Without Fooling Yourself
A strong fake door result means people want the door opened. It is permission to build, not proof of a business. Treat a pass as 'this feature earned a real prototype', then validate willingness to pay separately, ideally by pre-selling the built version or running a smoke test on pricing.
A weak result is the more valuable outcome, because it just saved you weeks of building. Before you kill the idea, sanity check that the door was findable, honestly labeled, and shown to the right people. If all three were true and almost nobody clicked, that is your answer. Move the effort to a door people actually reach for.