A lean team cannot carry a mediocre GTM hire for long. One weak seat drags the whole number down. Interviews now reward the candidate who prompts best, not the one who performs best. AI gave everyone the vocabulary of expertise without the underlying skill. The only honest signal left is real work, watched closely and scored hard. You build a case plus a week-one proof task, then judge in assessor mode. You compare each candidate against your best performers, not the other applicants. Then you say no unless you clearly see a bar raiser.
Read time: about 5 minutes. First step runnable Monday morning.
Write the week-one proof task before you write the job description. Most teams post the role, screen resumes, then improvise a case late. That order rewards polish and hides skill. Flip it. Name the first real problem this hire solves in week one. Turn that into a task they can actually do. The task becomes your filter, your interview, and your scorecard at once.
Start here: name the single outcome this seat must produce in its first month. You need one real problem, the inputs a new hire would get, and a clear definition of done.
Ignore this week: the perfect job ad, the long competency matrix, and the ten-stage funnel. They slow you down and measure the wrong thing.