Ask everyone the same questions.

A structured interview is one where every candidate for a role gets the same questions, in the same order, and is scored against the same criteria written down in advance. It sounds rigid. In practice it’s the difference between comparing candidates and comparing conversations. This guide covers what makes an interview structured, why that makes it fairer and more useful, how to run one step by step, and how structure applies to asynchronous and AI interviews.

What makes an interview structured

Structure is a spectrum, not a switch. A fully structured interview has all of these:

• Questions based on the job. Each question tests a skill the work actually needs.
• The same questions, in the same order. Every candidate gets the same chances to show the same things.
• Scoring criteria written in advance. What a strong, adequate and weak answer sounds like, for each skill.
• Scoring as you go. Each skill is rated right after the interview, from notes, not from memory at the end of the week.
• Planned follow-ups. Probing questions are allowed, but chosen in advance and used consistently.
• Consistent conditions. The same length, format and information given to every candidate.

Most real interviews sit somewhere in the middle:

Why it’s fairer

In an unstructured interview, two candidates for the same job can have completely different experiences. One gets hard technical questions. The other spends twenty minutes chatting about a shared hometown. The second one often comes out ahead, and it has nothing to do with the job.

Structure closes those gaps:

• Everyone gets the same chances. Nobody is asked easier questions because the interviewer warmed to them.
• Rapport matters less. A candidate who shares your background or sense of humor doesn’t get extra credit for it.
• Decisions are explainable. You can point to the question, the answer and the criterion, which helps with internal review and with candidates who ask for feedback.

Structure doesn’t prevent adjustments for candidates who need them. You can change the format, give more time or offer breaks, while keeping the questions and criteria the same.

Why it predicts better, in principle

An interview is a measurement. Like any measurement, it gets more useful when you measure the right thing, measure it the same way each time, and reduce the noise.

• Job-based questions measure the right thing. Questions tied to real skills tell you about the job. Free-form chat mostly tells you how someone does in free-form chat.
• Same questions make answers comparable. It’s hard to rank two runners on different tracks. It’s equally hard to rank two candidates on different questions.
• Written criteria reduce rater noise. Two interviewers with clear anchors are more likely to agree. Two without them are rating different things.
• Separate skill scores stop one impression from dominating. A strong opening answer doesn’t lift every other score.

None of this makes an interview perfect. It removes avoidable sources of error, which is most of what you can do.

How to run one, step by step

1. Pick three to five skills the job needs from day one.
2. Write one or two questions per skill. Mix behavioral (“Tell me about a time…”) and situational (“Imagine…”). See writing pre-screening questions.
3. Write anchors for each skill: what a strong, adequate and weak answer sounds like. A screening scorecard holds all of this in one place.
4. Choose your follow-ups in advance. A short list works for most questions: “What did you do, specifically?”, “What happened next?”, “What would you do differently?”
5. Tell the candidate the format. How long it takes, how many questions, and that you’ll take notes. It lowers nerves and makes answers better.
6. Ask the questions in order. Don’t skip one because the candidate “obviously knows it”.
7. Note what was said, not your impression of it. “Refunded the customer, then called back two days later to check” beats “good ownership”.
8. Score each skill straight after, before talking to anyone else about the candidate.
9. Compare scores with other interviewers only after everyone has scored. Otherwise the first opinion sets the rest.
10. Review after the first batch. Rewrite any question that didn’t separate candidates.

Common ways structure slips

• Rewording questions for some candidates. A small change can make a question easier or harder.
• Small talk that becomes evidence. Chat to settle nerves is fine. Scoring it isn’t.
• Scoring from memory. Waiting until the end of the day turns scores into impressions.
• Gut first, criteria second. Deciding yes or no, then filling in scores to match, defeats the point.
• Adding questions halfway through a hiring round. Candidates before and after the change aren’t comparable.

If more than one person interviews for the same role, structure matters even more. Share the questions, the anchors and a couple of example answers before the first interview, and have everyone score the same recorded interview once as a practice run. Differences that show up there are cheap to fix. Differences that show up after thirty candidates are not.

Structure in asynchronous and AI interviews

Asynchronous interviews are structured almost by design, since every candidate gets the same prompts. In LessRounds you choose between two modes. Scripted interviews read your questions word for word, which is the most structured option. AI-led interviews keep your questions and add follow-ups based on each answer, which gets more depth at the cost of some uniformity. In both, every candidate is scored against the same skills, levels and weights. See the AI interviewer for how each mode works.