A manufacturing CEO who makes knitwear samples wrote something this week that applies to every software founder alive. A sales sample can look ready for launch while important decisions remain open, so the key question is what the approval actually authorizes: further development, commercial presentation, or bulk production. His closing line was the knife. A well photographed sweater should never become a production standard simply because everyone likes the image.
Replace sweater with demo and you have described half of all startup launches.
Why do founders ship on applause?
Because a yes feels binary. Someone saw the thing and approved of it, so the thing must be right. But approval is not one object. The question is never whether you got a yes, it is what that particular yes is strong enough to pay for. A like pays for nothing. An enthusiastic conversation pays for one more conversation. A stranger giving you forty minutes and their real workflow pays for a prototype. A signed order pays for production.
An investor and startup mentor made the same point this week from the demand side. Founders say we validated the market, and when he asks what they actually did, the answers range from reading industry reports to collecting survey promises. All of it gets filed under the same word, validation, even though each piece authorizes a completely different size of bet.
What does each kind of evidence authorize?
A rough ladder. Secondary research, reports and market sizes, authorizes choosing where to look, nothing more. A concrete problem story from the right person authorizes sketching a solution. A stranger reacting to something concrete, a mockup or a manual service, authorizes building the smallest testable version. Behavior, someone bringing a real task, coming back, introducing their boss, authorizes investing in the product. Payment under terms the buyer understands authorizes scaling it.
Notice what never appears on the ladder: compliments. 'I would use this' authorizes exactly one thing, the follow-up question about the last time they faced the problem.
Who should be allowed to say yes?
The manufacturing version has a discipline software forgot: each approval needs a named owner and clear evidence. Fit approval belongs to whoever measured the fit, not whoever liked the photo. In a startup that means separating roles too. The market approves demand. You approve scope. Your enthusiasm is not allowed to co-sign the market's part, and the market's politeness is not allowed to co-sign yours.
This is also the honest way to read your own pipeline of evidence. Write down the next step you are about to take, then write down the strongest evidence you have. If the step is bigger than the evidence, you are not validating, you are borrowing against applause.
Key takeaways
- A yes is not one object. Every piece of evidence authorizes one specific next step and no more.
- Compliments authorize questions. Stories authorize prototypes. Behavior authorizes product. Payment authorizes scale.
- Each approval needs a named owner and evidence, and your own enthusiasm cannot co-sign the market's part.
- If the step is bigger than the evidence, you are borrowing against applause.
The photo of the sweater was never the problem. Treating it as a production order was.