A recruiter told a story this week about two young founders who came to him in 2020. They wanted to pivot their grocery platform to ten minute delivery. He was certain it would not work. Why would anyone need groceries that urgently when the corner store is right there? He helped them hire anyway. Within a couple of years they had built a billion dollar company and a new category. His closing line was the honest part: he still assists startups, but now he does the work before relying on his convictions.
Everyone quotes this kind of story as a lesson about believing in founders. It is actually a lesson about something less comfortable. The confident no of an experienced operator was worth exactly as much as the confident yes of an excited founder. Zero, until the market moved.
Why does conviction feel like evidence?
Because it is built from the same material. Conviction is compressed experience, thousands of observations about markets and people, folded into a fast answer. That compression is useful. It is also lossy. The recruiter's model of grocery shopping was trained on a world without cheap delivery infrastructure and without a pandemic. His instinct was not stupid. It was stale, and instinct never reports its own staleness.
A founder in a scaling thread put the same mechanic in one sentence: his gut was excellent at small decisions and terrible at big ones, because small decisions repeat daily and big ones never give you enough reps to train on.
Is a confident no worth more than a confident yes?
Founders are warned constantly about positive bias. Your friends lie to you, polite strangers say interesting, stated interest is not revealed intent. All true. But skepticism gets a free pass it has not earned. A dismissal is also a prediction about other people's behaviour, made without watching any behaviour. The skeptic just sounds smarter, because no is the lower status risk. If the idea dies, nobody audits the people who called it.
The result is an asymmetry that quietly kills good ideas. Founders over-test their optimism and never test their pessimism. The ideas they drop after one discouraging coffee chat get less scrutiny than the ideas they pursue.
What does experience actually buy you?
Better questions, not better answers. An experienced operator looking at ten minute delivery in 2020 could have asked sharp things. What does the basket look like when the store is eight minutes away? Who pays the delivery premium, and how often? What happened in the three cities where someone tried this before? Those questions convert conviction into a test plan. Stating the conclusion converts it into a bet, with the same information either way.
How do you price a conviction?
Ask what it would cost to be wrong, then ask what it would cost to check. Conviction is cheap to hold and expensive to act on. Checking is the reverse. A week of putting the idea in front of real people, people with the problem, not people with opinions about the problem, prices the belief before you spend a year on it. What they do with the idea, commit time, hand over data, pay, is the number. What anyone predicts they will do, including you, is commentary.
The recruiter got this right in the end. Do the work first, rely on the conviction second. Most people run it the other way and call the conviction research.
Key takeaways
- Conviction is compressed experience. Useful for choosing what to test, worthless as proof.
- A confident no is a prediction too, and it deserves the same scrutiny as a confident yes.
- Experience buys better questions, not better answers. Turn instinct into a test plan.
- Price every belief against revealed behaviour: what real people did, not what anyone predicted.
Your certainty is a feeling about the market. The market has never once checked how you feel.