The interview is the noisiest signal in hiring, and it’s free. So most funnels lean on it. FairHire inverts that: a real, compensated work sample is the heaviest signal — weighted half the decision — and you pay every candidate who does it.
The paid trial enters the composite as its own criterion at weight .50 — half the decision. A 9.2 on the trial contributes 9.2 × .50 = 4.60. Interviews sit at .20. The work outweighs the conversation, in the math itself.
A trial carries a rate (hourly or flat) and a recorded payment. The money moves from you to the candidate — FairHire never sits in the middle and never takes a cut. Compensated work is the whole point.
Every trial is a set of deliverables — tickets the candidate ships, each with an acceptance note and an artifact link (a PR, a deploy, a doc). The score is anchored to what actually shipped, not a vibe.
The dimensions a trial is judged on are defined up front, so the same standard applies to everyone who runs it. No moving goalposts, no post-hoc rationalization.
A title, a brief, the deliverables to ship, a time budget, and a pay rate.
The candidate accepts the scoped, compensated work.
Real work against real tickets — tracked against the time budget.
Deliverables come in, tracked as “N of M shipped” with artifact links.
A 0–10 score and a pass / borderline / fail outcome, with a debrief.
The payment is recorded against the trial. You pay directly — we take no cut.
Illegal jumps — marking a trial paid before it’s scored, for instance — are rejected. The state machine keeps the record honest.
Scope a paid trial, score the shipped work, and let the heaviest signal carry the decision.