Korva is a scroll of the internet’s best ideas.
The principle
The principle is the same as in a fundraise: a founder who squeezes the last dollar out of investors isn't someone they want to back again. In growth, the same instinct leads teams to trade safety, brand, quality, and user experience for a better number this quarter. Anthropic's team puts those four ahead of the metric.
Two kinds of tests
Controversial experiments fall into two types. In the first, the idea is so far over the line on brand, customer friendliness, and values that the result doesn't matter, because it would never ship. In the second, the idea is uncomfortable but not a red line, so it can run, and the more cringe or ick it carries, the higher the return has to be to justify it. Anything that touches AI safety falls in the first type.
The best products out there all operate this way, and it drives more growth over the long run.
Why do short-term tricks erode customer trust?
Tactics that push conversion up this quarter often work by getting people to do something they did not intend. A 2019 study by Arunesh Mathur and colleagues at Princeton and the University of Chicago crawled about 11,000 shopping websites and found manipulative design patterns on roughly 11 percent of them. Customers who feel tricked cancel, dispute charges, and tell others, which shifts the cost to later quarters.
Why must riskier tests clear a higher bar?
Finance teams use a hurdle rate, a minimum return a project must clear before it is approved, and they set it higher when a project is riskier. A growth experiment with a brand or trust cost follows the same logic, since the cost of that damage is uncertain and can be large. The more a test risks, the larger the measured gain has to be to justify running it. Corporate finance textbooks define the hurdle rate as the required return adjusted for risk.
Why does chasing a metric distort behavior?
Economist Charles Goodhart observed in a 1975 paper on monetary policy that a statistical regularity tends to break down once authorities try to control it. Anthropologist Marilyn Strathern later phrased the idea as the saying that when a measure becomes a target, it ceases to be a good measure. People optimize the number itself, including in ways that no longer serve the goal the number was meant to track.
More from Lenny's Podcast: Product | Growth | Career
