Korva is a scroll of the internet’s best ideas.
Wispr Flow 5x'd its budget in a month
The spike
Last year, the team went as hard as possible and watched where the pieces fell through. With a target of 30% to 40% month-over-month growth, there were not six months to work out the ceiling slowly. The spike showed where audience saturation hit, where more creative was needed, and which spend was incremental.
What it showed
The results rearranged the mix. Meta needed to rank second to Google. The newsletters were poor and one podcast didn't work. Understanding how the channels fit together came from seeing them stressed at once.
The rebuilt ramp
The team then pulled spend back and scaled up again using those findings, which gave a repeatable ramp. One expectation did not hold: the spike might have shown spend scaling indefinitely, and it didn't. Channels that clearly can't work get turned off right away. Where leading indicators are weak but there are a million dials from ad to conversion, such as a personalized landing page that repeats the ad and shows the creator's face, a channel gets time before being scratched off.
Why does extra ad spend buy less?
Ad platforms deliver to the people most likely to respond first. As budget grows, the system has to reach people who are less likely to act, so each added dollar produces less. Economists call this diminishing marginal returns, an idea the French economist Turgot described in 1767 using farmland. In advertising, the relationship is often drawn as a response curve.
How did Netflix stress-test its systems?
Netflix built a tool called Chaos Monkey around 2010 that randomly shut down servers in its live system. The aim was to expose weak points while engineers were watching, instead of during an outage. The practice grew into a field called chaos engineering. Netflix released Chaos Monkey as open source in 2012.
More from The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch
