How to evaluate candidates after an interview (without trusting your gut)
The interview ends and the real failure point begins: evaluation. Most decisions get made by gut feel dressed up in post-hoc justification, anchored by whoever speaks first in the debrief. The fix isn't distrusting judgment — it's sequencing it: independent evidence first, scores second, discussion third, gut last (as a prompt for investigation, never as a verdict).
Score alone, immediately, in writing
Within 30 minutes of the interview — before lunch with a colleague, before Slack — rate each attribute 1–4 with one line of evidence per rating. Not 'strong communicator' but 'explained a technical migration to a non-technical stakeholder — clear, checked understanding twice'. If you can't cite evidence for a score, you're scoring a vibe.
Use a real scale: 1 = evidence against, 2 = insufficient evidence, 3 = clear evidence, 4 = exceptional evidence. The '2' matters — 'they seemed fine' usually means you didn't gather evidence, which is an interviewing failure, not a candidate one.
Run the debrief so it doesn't anchor
Scores are written before anyone speaks — then the most junior interviewer goes first (senior-first ordering silences the room). Discuss where scores DIVERGE: agreement needs no meeting. Divergence is usually different evidence, not different standards — share the evidence, not the adjectives.
Ban 'culture fit' as a free-floating veto in the debrief: any fit concern must name the specific defined behaviour it maps to. And the hiring manager decides — debriefs inform decisions; committees dilute accountability.
Comparing candidates without recency bias
The last candidate always feels strongest — memory favours recency, so compare written scores, not recollections. Compare attribute-by-attribute rather than holistically: 'A is stronger on judgment, B on execution; which gap can we coach?' is a real decision, 'B felt stronger' is a memory artefact.
And beware comparing candidates to each other instead of to the bar. A weak field makes mediocre candidates look great — the question is never 'who is best?' but 'does the best one clear the bar?'. No-hire is a valid, often correct outcome.
What to do with your gut
Gut feelings are pattern-matching — sometimes real signal, often bias in a trench coat. The rule: a gut feeling is a hypothesis to investigate, never a verdict. Uneasy despite good scores? Name the worry concretely, then test it — a targeted reference question, a follow-up conversation. If it survives investigation, it's evidence; if you can't articulate it, it's noise.
We're building an AI candidate you can rehearse real interviews on — built from the actual CV, with coaching on your technique. Free for the first cohort.
FAQ
How do you evaluate a candidate after an interview?
Score each pre-defined attribute 1–4 with written evidence within 30 minutes, alone. Then debrief: written scores first, junior voices first, discuss only divergences, evidence over adjectives. Compare candidates attribute-by-attribute against the bar (not each other), and treat gut feelings as hypotheses to investigate rather than verdicts.
What is an interview debrief and how do you run one?
The structured conversation where interviewers combine their assessments. The rules that keep it honest: everyone's scores are written before it starts, the most junior speaks first, discussion focuses on score divergence, every claim needs cited evidence from the interview, and the hiring manager owns the final call.
Keep reading
General hiring guidance, not legal advice. Employment law varies by jurisdiction — check locally before adopting questions or processes.