Korva is a scroll of the internet’s best ideas.
Three assumptions
Clay, a software company that helps businesses find and reach their customers, made three assumptions early. People in go-to-market work are creative, so they get the most powerful tool instead of a simple out-of-the-box one. That tool needs the right user, so it is built for RevOps, an existing role, under the label go-to-market engineering. And it charges for usage instead of per seat, because a customer with fewer people should not work against the company's interest.
Why they work
Almost every other decision falls out of those three. Anyone can work out what to do next by checking a choice against them, which makes the assumptions easy to communicate to a team that is growing quickly. Clay went from one to one hundred in two years, and that kind of growth only works if the company is pointing the right way.
Riding the wave
Pointing the right way matters most when a wave arrives. Clay was already taking off before ChatGPT came out, and it was well set up for language models because it already allowed any integration and worked like a coding product built from go-to-market pieces. Once LLMs arrived, everything got a boost.
Why does usage pricing change how a vendor behaves?
Per-seat pricing ties revenue to the number of people using a product, so a customer that shrinks its team pays less. Usage-based pricing ties revenue to the work actually done. Amazon Web Services helped make this model familiar when it launched S3 in March 2006, billing storage by the amount used. Twilio and Snowflake later built businesses on the same approach.
How did revenue operations become its own role?
Sales operations roles have existed for decades, handling territories, quotas, and reporting. Marketing operations grew in the 2000s as platforms like Marketo and HubSpot, both founded in 2006, gave marketers data and automation to manage. Revenue operations merged these functions with customer success operations so one team could manage a shared process and shared data. The title spread widely among software companies through the 2010s.
Why did language models favor tools with open integrations?
OpenAI released ChatGPT on November 30, 2022. In June 2023 it added function calling, which let a model produce structured requests to outside software. A product that already connects to many data sources and services gives a model many actions it can take. A closed product has far fewer.
More from Invest Like the Best with Patrick O'Shaughnessy
