Run Interviews That Hire the Right Person

Proof required

Submit your audit: notes or reconstructions from 5 real interviews with three-criterion ratings (question quality / listening depth / bias minimisation, each 1–5) and a one-sentence specific observation per criterion per interview. Append a 150-word diagnosis naming one recurring weakness with evidence from at least 3 of the 5 interviews.

What gets checked

• 5 real interviews you actually conducted — not hypothetical or reconstructed without any contemporaneous record
• Bias minimisation ratings are differentiated — all 5 rated 5/5 on bias minimisation is statistically implausible for most interviewers and should prompt deeper examination
• Diagnosis names a specific pattern (e.g. 'I ask closed questions that confirm what I already think' or 'I interrupt before the interviewee finishes their most interesting answers')

Common mistakes

• Using only hiring interviews where formal structure masks the natural weaknesses — include at least one informal or research-style interview
• Rating question quality without examining whether the questions were actually open-ended — closed questions that produce yes/no answers are a fundamental weakness
• Diagnosis that reads as general ('I could listen better') rather than behavioural ('I return to my script within 90 seconds of any unexpected answer')

Resources

What a verifier looks for

• Ask the submitter to describe the most surprising thing a person said in any of the 5 interviews — genuinely curious interviewers remember these; script-bound interviewers often do not
• Check that the listening depth ratings below 4/5 are accompanied by a specific example of where the follow-up was missed
• Confirm the diagnosis is behavioural and specific — ask the submitter to role-play one minute of an interview to see whether the stated weakness appears in real time
• A submitter who rates all 5 interviews 5/5 on bias minimisation with no diagnosis of this criterion should be asked: 'What would a biased question look like?' — if they cannot describe it, the criterion was not applied

Proof required

Submit your 6 observation forms (completed by an observer or your own immediate post-interview self-review if unobserved) plus 6 reflection notes (minimum 100 words each), each linking a specific in-interview behaviour to your diagnosed weakness. At least 3 forms must be from real interviews (not staged practice).

What gets checked

• At least 3 of the 6 interviews are real (real candidate, real research participant, real journalistic subject) — staged interviews with willing colleagues only do not qualify
• Each observation form names at least one specific moment — a specific question asked, a missed follow-up, a script return — not a general rating
• Reflection notes improve across the 6 interviews — the final 2 reflections should reference a different specific weakness than the first 2, showing adaptation

Common mistakes

• Conducting 6 interviews without any observation or recording — self-assessment without a contemporaneous external record captures only what you remember noticing, not what actually happened
• Reflection notes that repeat the same observation 6 times without adaptation — the purpose of the practice cycle is adjustment between iterations
• Counting staged practice interviews with friends as 'real' — the pressure and authenticity of a real interview is the point

Resources

What a verifier looks for

• Ask for evidence that at least 3 interviews were real — a real interview leaves a trace (a hiring record, a research session note, a project it fed into)
• Look for specificity in the observation forms — 'asked a leading question at 12:30 in the recording' is evidence; 'occasionally asked leading questions' is not
• Compare the first and last 2 reflection notes — behavioural adaptation across the cycle should be visible
• If no observer was present for any interview, ask how the submitter captured the behaviours in the observation form — a recording or detailed contemporaneous note is acceptable; memory alone is not

Proof required

Submit a brief description of the interview context (purpose, who was interviewed, what you were trying to learn) plus the observer's written assessment: specific questions noted as strong or weak with reasons, moments of strong or missed follow-up, one concrete recommendation for your primary weakness, and the observer's sign-off confirming their role and that they observed the full 20-minute interview live.

What gets checked

• Observer assessment cites at least 3 specific moments from the interview — specific questions or exchanges, not general impressions
• Observer's concrete recommendation is directly linked to the primary weakness identified in M1 — generic advice to 'ask more open questions' does not qualify if that was not the identified weakness
• Observer sign-off confirms professional background: hiring manager / UX researcher / journalist / therapist with minimum 5 years of interviewing practice

Common mistakes

• An observer who primarily praises — the assessment is only useful if specific weaknesses are identified with examples from the interview
• Interviewing a close friend or colleague who is motivated to perform well — the interviewee should not be aware of the assessment purpose, or the interview dynamics are artificial
• A structured assessment written by the submitter and confirmed by the observer — the assessment must be written by the observer from their own observation, not edited by the submitter

Resources

What a verifier looks for

• Ask the submitter to identify the moment in the interview they are most proud of — then compare it against the observer's assessment to see if they agree
• Confirm the observer's professional background independently — ask the submitter for the observer's job title and organisation rather than just accepting the sign-off
• Ask whether the interviewee knew they were being assessed — this affects how the interviewee performs and how useful the interview was as practice
• Observer minimum qualification: 5+ years conducting professional interviews in a relevant context — a colleague who has done occasional hiring does not qualify