Structured interviews beat gut feeling

The problem with "I just know"

Every experienced interviewer believes they can read people. The evidence says otherwise: unstructured interviews are famously poor predictors of job performance, because they mostly measure how comfortable the conversation felt. And conversations feel comfortable with people who are similar to us.

Gut feeling is not insight. It is pattern matching on charm, confidence, and familiarity, three things with no reliable connection to doing the job well.

What a structured interview actually is

Structure sounds bureaucratic, but the definition is simple: you decide before the interview what you are assessing and how you will recognise it, and you assess every candidate the same way.

That is the whole idea. The same core questions, in the same order, scored against the same criteria. Everything else, the rapport, the follow-ups, the humanity, stays.

Scorecards: decide the criteria before, not after

A scorecard written before the loop forces the most valuable conversation in hiring: what does great actually look like for this role? Which three or four competencies decide success? What would strong evidence for each one sound like?

Without that conversation, every interviewer assesses their private version of the role and the debrief becomes a negotiation between impressions. With it, interviewers collect evidence against shared criteria, and the debrief compares evidence instead of vibes.

Same questions, comparable answers

Asking each candidate different questions feels personalised, but it quietly destroys your ability to compare them. If one candidate got a softball and another got your hardest scenario, their answers tell you about your questions, not about them.

Behavioural questions anchored in real situations work best: "tell me about a time you shipped something you were not proud of, and what you did next." Ask everyone. Probe freely on the answers. Score what you heard, not what you assumed.

Calibrate as a panel, not a hallway

The hallway conversation right after the interview is where good process goes to die. The most senior or most confident voice anchors everyone else, and independent judgments collapse into one opinion with extra steps.

Have every interviewer submit their scorecard before seeing anyone else's. Then debrief: where scores disagree is exactly where the interesting evidence lives. A 4 and a 2 on the same competency means someone saw something the other missed, and that conversation is the entire point of interviewing as a team.

Structure does not mean robotic

The common objection is that structured interviews feel mechanical. Done badly, they can. Done well, the structure is invisible to the candidate: they experience a prepared interviewer, relevant questions, and a fair process where nothing depends on which interviewer they happened to draw.

Warmth in the room, rigour on the page. Candidates feel both, and both are the point.