Korva is a scroll of the internet’s best ideas.
Start at an eight, then hunt for holes
Starting at an Eight
Sarah Guo, who runs the venture firm Conviction, uses a one to ten rating scale. When she knows what a founder has worked on and has some background on the idea, she can land at an eight or a nine immediately. The next days to weeks go to finding the holes in her understanding: what everybody else believes about the space, whether the people are who she thinks they are, and where her judgment of the premise or the person is incomplete or wrong.
Grounding the Bet
A firm covering biology, defense, robotics, and law can't know every domain, so grounding takes work. She wants second reads on people and wants to understand the core questions. She writes a full memo, perhaps Greylock style, and ships it to a trusted investor friend outside the funds, a habit from when the firm was just her. Partners then ask what they want to know about the person to complete the picture.
Fast Calls
Other investors think and think and climb to conviction. She and her partner Mike start at a point and say what would move them. The volume is low at four to six new companies a week, against 500 in her first couple of months at her previous firm, and the gap reflects much more confidence in telling quickly.
How does an early impression anchor judgment?
Psychologists Amos Tversky and Daniel Kahneman showed in a 1974 paper that an initial number pulls later estimates toward it, even when the number is arbitrary. In one experiment, a spinning wheel's random number shifted people's guesses about the share of African countries in the United Nations. The wheel was rigged to land on 10 or 65, and the average guesses were 25 and 45.
Why are quick judgments sometimes accurate?
In studies from the early 1990s, psychologists Nalini Ambady and Robert Rosenthal showed people silent clips of teachers lasting 30 seconds or less. Ratings of those clips predicted end-of-semester student evaluations of the same teachers. The result became the basis of research on thin slices, brief samples of behavior that carry real information.