Asking someone if your idea is good practically begs them to lie.

Not maliciously. They just don’t want to hurt your feelings, or they genuinely think they’d use it, or they’re imagining a hypothetical future version of themselves with different habits and more time. “That sounds useful” is almost meaningless data. It’s social lubricant, not market signal.

The Gap Is Large

One case study put a number on this. An ad campaign got a 16.5% click-through rate — good by most benchmarks. Then 2.47% of those people signed up for an email list. That’s an 85% drop-off from “interested enough to click” to “interested enough to give an email address.”

And email is easy. It’s reversible. It costs nothing. The drop-off from email signup to paying customer is another steep cliff.

This is the validation gap: the distance between what people say they’ll do and what they actually do. It’s not specific to startups. It’s a feature of human behavior. People are optimistic about future versions of themselves who exercise, cook more, and use your product regularly.

The Hierarchy

There’s a rough ordering of signal quality:

Someone says “great idea” — nearly zero information. They’re responding to you, not to the product.

Someone clicks an ad — weak signal. They were curious enough to move a finger. This is intent, not commitment.

Someone gives their email — medium signal. They took an action, provided something personal, and accepted some friction. Still reversible, still cheap.

Someone pays $10 in advance — strong signal. They exchanged money before the product exists. That’s a real data point.

Someone pays recurring — confirmed fit. They decided this problem was worth solving at this price on an ongoing basis.

Most validation frameworks collapse several of these together. The advice to “talk to users” is good, but conversations are closer to the first level than founders realize. The advice to “build a landing page” is better, but a landing page with no payment capture is still sitting at the email level.

The Dropbox Move

The canonical example: a company validated its product with a video showing how it would work. No product existed. The video generated 70,000 pre-launch signups.

The key insight isn’t “make a good video.” It’s that the video captured real behavior — people encountering something, deciding they wanted it, and taking an action to secure access. That’s a qualitatively different signal than someone saying they’d use it if it existed.

The behavior is the data. Not the stated intention.

What This Changes

If the only real signals are email-level and above — and only payment-level signals are predictive of revenue — then most “validation” is theater. It feels like progress because you’re getting feedback and encouragement, but it’s not reducing the actual uncertainty about whether people will pay.

This doesn’t mean don’t talk to potential customers. It means don’t count conversations as validation. They’re research. Validation is when someone does something — clicks, signs up, pre-pays — that would be mildly irrational if they didn’t actually want the thing.

Get to that level before you spend months building. The question isn’t “do people like this idea” but “will five strangers give me $10 right now.”

If the answer is no, you have useful information. If the answer is yes, you have a product.