Founder Intuition: When Your Gut Has Earned the Right to Decide
The Compass Team
November 13, 2026 · Updated November 14, 2026
In 2009, two psychologists who had spent their careers disagreeing published a paper together. Daniel Kahneman had built the case that human judgment runs on error-prone shortcuts. Gary Klein had spent decades documenting experts who read a situation correctly in seconds.
They titled it "A Failure to Disagree." What they agreed on is the most useful thing either of them wrote about intuition.
Intuition is trustworthy in some places and worthless in others, and you can often tell which place you're standing in before you decide. Two conditions have to hold.
The environment has to be regular, so the same cues lead to the same outcomes often enough to learn from. And you need practice with feedback that arrives fast and clear.
Firefighting passes. Fire behaves consistently, and a commander finds out quickly whether the call was right. Medicine mostly passes.
Picking individual stocks fails the first test, and the paper is blunt about the consequence: where outcomes are close to random, the judgments that turn out well breed an illusion of skill. Confidence rises either way, and confidence is not a signal.
Founder intuition sits in an awkward spot. The environment is the market, and its feedback is slow, sparse, and ambiguous. You ship, you wait, and the result arrives attached to a hundred other things you also changed.
Dalton Caldwell, who runs admissions at Y Combinator, says YC sees founders get this backwards. The founders with the most expertise are often the ones who hesitate to build the thing they know the world needs. The ones without much experience want to spend five years on the vision before they show it to anyone.
Both are guessing about their own judgment, and neither has a way to check.
Thomas Astebro at HEC Paris, writing about research he did with Frank Fossen and Cedric Gutierrez, sharpens the problem. Some founders adapt deliberately. Others follow instinct and construct the reasoning afterward, and even they can't reliably tell which one they just did.
In his team's analysis of startup evaluation panels in Canada, reviewers overestimated the odds of early-stage ideas most when uncertainty was highest, and their predictions often didn't match what happened.
Olivier Sibony, also at HEC Paris, has a line about hiring that travels further than hiring. The people you hired succeeded or failed for all kinds of reasons that had nothing to do with what you spotted in the interview.
That is what a founder's intuition mostly runs on. A pile of outcomes with no record of what was expected, so no way to score the call that produced them.
The fix is mechanical. Write the decision down before the result exists: what you expect, the number you'll check, and the date you'll check it. Three lines.
Once that exists, the market's slow and noisy feedback becomes feedback on your judgment specifically. You expected 40 signups in two weeks and got 9. That's a clean test, and the writing took one sentence.
A founder with three companies behind them does have real pattern data. That gut earned its trust the same way a fire commander's did, through repeated calls anyone could check. A founder still early in the curve can manufacture the conditions instead of waiting years for them to accumulate.
Compass keeps the record. The judgment stays with you.
Share this article