Structured Interview

A structured interview uses a fixed set of job-related questions, asked of every candidate for the role, and scored against a written rubric that says what a weak, adequate, and strong answer looks like.

The opposite is the unstructured interview, where each interviewer improvises questions and forms a holistic impression. That format feels more natural and is substantially worse at predicting performance, because the interviewers end up measuring different things and comparing candidates on incompatible evidence.

How it works

Structure comes from three commitments made before any candidate is seen. First, the competencies: a short list of what the role actually requires, drawn from the job description rather than from what the last person in the seat happened to be good at. Second, the questions: one or two per competency, assigned to specific interviewers so the panel covers the list without four people asking about the same thing. Third, the rubric: for each competency, a written description of what each rating level means in terms of observable evidence.

During the interview, the interviewer asks the assigned questions, probes for specifics, and takes notes on what the candidate actually did rather than on how the conversation felt. Afterward, the interviewer records a score per competency with the evidence that supports it, and does this before seeing anyone else scores.

The debrief then reviews scores and evidence rather than reopening the interview. Disagreement is useful here: two interviewers who scored the same competency differently usually heard different evidence, and surfacing that is the point of the meeting.

Building one for a role

This is a one-time investment per role family, reused for every search after that.

1. 1Define the role from the job description: the outcomes the person owns in the first year and the capabilities required to deliver them.
2. 2Reduce that to four to six competencies. More than six and the panel loses focus, fewer than four and the interview stops discriminating between candidates.
3. 3Write one or two behavioral or situational questions per competency, each asking for a specific past example or a decision on a realistic scenario.
4. 4Write the rating anchors: what evidence a low, middle, and high score requires. Do this before interviewing, not after the first candidate.
5. 5Assign competencies to interviewers so coverage is complete and no competency is assessed only once.
6. 6Brief the panel on the questions, the anchors, and the rule that scores are submitted independently before the debrief.
7. 7Run every candidate for the role through the same set. A candidate who skips a stage has not been evaluated on the same basis as the others.
8. 8Review after the search: which competencies actually separated candidates, and which questions produced the same answer from everyone.

What a rating anchor looks like

An anchored scale for a single competency. The text is what stops one interviewer 3 from being another interviewer 5.

Where structure quietly breaks down

Most panels believe they are structured. Fewer are.

• The questions are fixed but the rubric is not, so everyone scores on a private standard.
• Interviewers talk before submitting scores, and the debrief measures the loudest opinion rather than the evidence.
• The panel changes between candidates, so a finalist is compared against someone a different set of people assessed.
• A referred or internal candidate gets a shortened process, which is the single most common way a structured loop is abandoned.
• Questions drift into personal territory. Family plans, age, health, religion, and national origin are not job related and have no place in the set.
• Scores are recorded without evidence, which turns the rubric into a number with nothing behind it.
• Nobody revisits the questions, so a set written for an old version of the role keeps selecting for the wrong things.

Why it improves both signal and defensibility

Signal improves because comparability improves. When every candidate answers the same questions against the same anchors, differences in score reflect differences in evidence rather than differences in who happened to interview them or which day the conversation went well.

Defensibility improves for the same reason. A selection decision that rests on a job-related, consistently applied, documented evaluation can be explained. A decision that rests on fit or on a general impression cannot be reconstructed months later, and an inconsistent process is difficult to distinguish from an arbitrary one.

The two benefits are not separate. The same discipline produces both, which is why structured interviewing is worth the setup cost even for teams with no compliance concern at all.