This week a founder in an accelerator cohort described a familiar month. He is building a marketplace between B2B brands and newsletters, and he started with the newsletter side because it felt easier. Twenty-seven newsletters signed up in a month, hundreds of thousands of combined subscribers. Then the other half of the sentence: nobody has paid, and the first campaign is now running with the fee waived. He asked whether he built it backwards. He did not. He built the free half first, and mistook it for progress.
Why does supply always say yes?
Because yes costs them nothing. A newsletter joining a sponsorship network risks no money, no reputation and about four minutes. The same is true of a waitlist signup, a beta tester, an advisor, and most letters of intent. These are all versions of the same object, a commitment whose price is zero. Zero-price commitments accumulate fast, which is why they feel like traction, and they predict nothing, which is why they are not.
What actually counts as evidence of demand?
Weigh every yes by what it cost the person who said it. Somebody who pays before being asked has spent money. Somebody who comes back without a reminder has spent attention twice. Somebody who built a spreadsheet workaround before you existed has spent hours on the problem, which is the strongest testimony of all because it happened before you could influence it. A founder collecting these three kinds of evidence can be wrong about many things and still be right about the only thing that kills companies.
Notice what is not on the list. Compliments are not on it. Survey answers are not on it. Two hundred waitlist emails are not on it, however the outreach was done. What people say they would do is a genre of fiction, usually written to be kind to you.
The rejection is the data
The most valuable asset that Antler founder owns right now is not the 27 newsletters. It is the one brand that saw the proposal and did not pay, even with the fee waived. That rejection contains the real information: what the buyer compared it to, what proof they wanted and did not get, what number they had in mind. Founders archive their rejections and frame their signups. It should be the other way around.
A fee waiver deserves the same scrutiny. Discounting to zero to get a first campaign out is sometimes right, but it converts your one demand experiment into another zero-price commitment. If they take it, you have learned that people accept free things. The experiment only produces data at a price above zero.
Can you test demand before building the supply?
Almost always, and almost nobody does, because demand is where the rejections live. Testing the paying side first means walking toward the people who can say no with their wallets, and every instinct in a founder points the other way, toward the side that says yes for free. That instinct is the actual reason so many marketplaces die with a beautiful supply side. The order of construction is emotional, not strategic.
Key takeaways
A yes that costs nothing predicts nothing, and supply-side yeses almost always cost nothing. Demand evidence is behavioral: unprompted payment, unprompted return, pre-existing workarounds. The first paying-side rejection carries more information than dozens of free-side signups. Discounting to zero turns a demand experiment back into a free yes. Founders build supply first because it feels good, not because it teaches anything.
Count the yeses that cost something. Everything else is applause.