Structured interviewing gets summarised as asking every candidate the same questions. That is one component and the least consequential one. A loop where four interviewers ask identical questions, score on gut feel and reach consensus in a room is not structured in any way that improves the decision.
The structure is in four things: what each stage is for, who assesses what, how scores are recorded, and how disagreement is handled.
Step 1: Decide what you are assessing before you design anything
Four to six criteria for the role, agreed at the intake meeting, written down. Not "strong engineer" but the specific capabilities the job requires — and specifically, the ones where a candidate could plausibly fail.
The test for a criterion: can you describe what a weak answer and a strong answer look like? If not, it is not a criterion, it is a preference. This is the same discipline the scorecard depends on, and it has to be settled before the loop is designed.
Step 2: Assign criteria to stages
This is the step teams skip, and skipping it is why loops feel repetitive to candidates and produce redundant signal.
Each criterion should be assessed deliberately, by someone equipped to assess it, and generally once. A worked example for a four-stage engineering loop:
- Screening call — motivation, baseline experience, logistics and expectations. No technical assessment.
- Technical stage — problem decomposition, code quality.
- Systems or design stage — technical judgement, trade-off reasoning.
- Hiring manager stage — collaboration, response to feedback, working style.
Every criterion is covered. None is covered four times. Each interviewer knows exactly what they own — which is also what makes their score meaningful rather than an overall impression wearing a number.
Step 3: Write the questions once
Two to three questions per criterion, written in advance, with follow-ups. Behavioural questions need a defined follow-up chain or they produce rehearsed narratives:
- What was the situation and what were you responsible for?
- What did you actually do — your part, not the team.
- What was the outcome, and how did you know?
- What would you do differently?
The fourth question is the discriminating one and the most frequently omitted.
Step 4: Calibrate the interviewers
A structured loop with uncalibrated interviewers produces consistent questions and inconsistent scores. Calibration is cheaper than it sounds:
- Every interviewer reads the criteria and the anchors before their first loop.
- New interviewers shadow one interview and score it independently, then compare with the lead. Differences of two or more points are the conversation.
- Quarterly, look at score distributions per interviewer. Someone who never scores below 3 is not participating in the decision.
Step 5: Independent scores, then debrief
Scores and evidence submitted before the debrief, without seeing other scores. Then discuss only where people disagree.
This ordering carries most of the value of the whole exercise. Consensus reached in a room is dominated by whoever speaks first and whoever is most senior. Independent scores make disagreement visible, and disagreement is the most informative thing in the process — see running the debrief.
What structure is worth
Structured interviewing is consistently among the better-evidenced predictors of job performance available to hiring teams, well ahead of unstructured conversation. But there is a second benefit that matters just as much operationally: a structured loop produces a record. When a hiring decision is questioned — internally or otherwise — you have criteria, evidence and scores rather than a recollection.
Where loops go wrong
- Adding a stage instead of fixing one. When a bad hire happens, the instinct is another round. Usually the existing rounds were assessing the same two things and missing a third.
- The unscheduled stage. A founder or executive who meets candidates informally, outside the scorecard, and can veto. This is unstructured hiring with extra steps.
- Scorecards nobody reads. If the debrief does not open the submitted scores, the loop is theatre.
- Never revisiting the criteria. Roles change. Criteria written two years ago describe a job that no longer exists.
For structuring this in the system rather than in a document, see interviews and scorecards.