A builder shared a strange revenue number this morning. His tool answers one question for Amazon Flex drivers, is the station I am driving to closed, and it made over $632 last week. Not from subscriptions. From voluntary donations, on a product that is completely free.
Most founders would read that as a monetisation problem. It is the opposite. It is the cleanest demand evidence I have seen all month, and it points at something most validation theatre gets backwards.
Why is a donation stronger evidence than a sale?
Because nothing forced it. A sale can be manufactured by a discount, a countdown timer, a persuasive founder on a call. A signup costs the user nothing and tells you almost nothing. A donation has no funnel. The driver checked the station, saved a pointless trip, and then decided, unprompted, that the thing deserved money. Payment that arrives after the value, with a zero price tag on the product, is demand in its purest form.
The same logic explains a second number from a very different market. A supply chain operator described a livestream campaign that sold around $370,000 of underwear in a month. His takeaway was not the influencer. It was the 10 percent return rate, in a category where online returns are brutal. Revenue measured the pitch. Returns measured the product. He called the sales a result rather than a starting point, and he is right.
What do these signals have in common?
They all arrive after the moment of persuasion, when the customer has nothing to prove and nobody to please. Donations, tips, repeat purchases, renewal without a call, a return rate that stays low once the parcel is opened. None of them can be inflated by better copy. This is what separates them from the metrics founders usually collect as proof: waitlist signups, demo reactions, kind words from friends, preorders extracted on a call. Those all measure the strength of the ask. The honest signals measure the strength of the product's case, argued in your absence.
Can you get this kind of evidence before building?
Not literally. Donations require a working product. But the class of signal is available earlier, and the class is the point: behaviour from real people that costs them something, produced without pressure. Someone with the problem gives you forty minutes and then asks when they can use it. Someone forwards your rough prototype to a colleague without being asked. Someone describes the workaround they already built and what it already costs them. Someone agrees to put real data, or real money, into a test. Each of these is a small donation, paid in time and risk instead of dollars.
The discipline is refusing to count anything else. If a signal could have been produced by politeness, curiosity or your own enthusiasm, it goes in the applause column, not the evidence column.
What should the flex driver tool do next?
The interesting question is not how to monetise. It is which donors come back. A one time donation says thank you. A repeat donation says this is part of my working week, which is the beginning of pricing knowledge no survey could produce. The builder is sitting on a willingness-to-pay study that funded itself.
Most products never earn a single unprompted dollar in their whole life. If yours earns six hundred a week by accident, the market is not whispering. It is shouting.
Key takeaways
- Money that arrives voluntarily after the value, donations, tips, repeats, is demand evidence that cannot be inflated by persuasion.
- Revenue measures the pitch. Returns, renewals and repeat behaviour measure the product.
- Before building, chase the same class of signal: unpressured behaviour from real people that costs them time, data or money.
- If a signal could have been produced by politeness, it belongs in the applause column, not the evidence column.
Applause is loud before launch and silent after. Evidence is the other way around.