Korva is a scroll of the internet’s best ideas.

Why trials matter

In a Cursor work trial, a candidate spends a day or two on site doing a project side by side with the team. Research has shown that work samples are the highest predictor of success at a company, yet most companies still spend their time on one-to-one assessments across a table.

The trial is expensive for the company and a big ask of the candidate. When the team tried hiring without it, they struggled to have confidence in their decisions. Working together over a stretch of time shows core values, collaboration, and other signals that can matter more than the specific technical skill.

How a trial runs

Candidates choose from a set of problems, and the team will size the project to fit their availability, sometimes scheduling it on a Saturday. Go-to-market candidates work through a challenging customer scenario with the hiring manager, and product candidates do a similar project together. The slow part is made up afterward: once a candidate is ready to move, a screen gets scheduled within the hour.

Why do work samples predict success so well?

A 1998 meta-analysis by Frank Schmidt and John Hunter in Psychological Bulletin compared many selection methods and reported work sample tests with a validity of about 0.54, well above unstructured interviews at about 0.38. A work sample works because it measures the behavior itself instead of a person's description of it. The closer a test mirrors the job, the less guesswork sits between the score and the performance. Industrial psychologists describe this match between test and task as content validity.

How does a shared project reveal collaboration?

Collaboration is hard to see in an answer to a question because it appears in small moments: how someone asks for help, takes feedback, and recovers from a wrong turn. A shared project over a day or two produces dozens of those moments in a real setting. Organizational psychologists call this behavior sampling, and it is the reason assessment centers have used group exercises since the 1940s. The Office of Strategic Services ran an early version to select intelligence officers during World War II.

More from Lenny's Podcast: Product | Growth | Career

Work & LeadershipRecruiters excel at 90 to 110%Recruiters are excellent between 90 and 110 percent of capacity and poor outside that range. A recruiter with half a req load is nearly useless.
Enjoyed it?