How to conduct a structured interview (and why most companies are doing it wrong)

The job interview has been the centerpiece of hiring for over a century. Companies invest enormous amounts of time in them. Candidates prepare extensively for them. Hiring managers treat their impressions from them as reliable signal.

The research on unstructured interviews is not kind. A landmark meta-analysis by Frank Schmidt and John Hunter, covering decades of hiring research, found that unstructured interviews have a validity coefficient of roughly 0.38 for predicting job performance. That is better than chance, but far below what most hiring managers believe their judgment is worth. For context, the same research found that structured interviews, when done well, reach validity coefficients closer to 0.51 and higher.

The gap between those two numbers represents real hiring outcomes: roles that produce or do not, hires that stay or leave, managers who have to restart a search six months later.

What makes an interview structured

A structured interview has three defining characteristics. The same questions are asked to every candidate for the same role. Responses are evaluated against a consistent, pre-defined scoring rubric. And the evaluation happens independently, before candidates are discussed as a group.

That is it. Structured does not mean cold or mechanical. It does not mean no follow-up questions or no conversation. It means the core evaluation is anchored to the same criteria across every candidate, so you are comparing apples to apples when you get to the decision.

An unstructured interview, by contrast, lets each interviewer lead the conversation however they like. One asks about past experience. Another goes deep on hypothetical scenarios. A third spends half the time talking about the company. The result is not a comparable set of signals about candidates. It is a set of interviews that have almost nothing in common, evaluated entirely by gut feel.

Why unstructured interviews fail even when interviewers are smart

The problem with unstructured interviews is not that interviewers are careless. Most people who interview candidates are trying hard to make a good decision. The problem is that the human brain is well-documented to be unreliable in unstructured evaluative settings.

Research from the University of Toledo found that interviewers form a strong impression of candidates within the first few minutes, and subsequent information is filtered through that initial impression rather than evaluated independently. This means that a candidate who stumbles at the start of an interview is at a systematic disadvantage regardless of how well the rest of the conversation goes.

Confirmation bias compounds this. Once an interviewer has a positive or negative impression, they tend to ask questions that confirm it and discount answers that contradict it. Affinity bias pushes evaluators toward candidates who share their background, communication style, or cultural references. None of this is deliberate. All of it shapes outcomes.

The common response to these biases is training. Train interviewers to recognize their biases. This does not work reliably. What does work is changing the structure of the interview itself so that bias has fewer places to operate. Consistent questions, scored against a rubric, evaluated independently, produce better decisions not because interviewers become less biased but because the process is less vulnerable to the biases they have.

Behavioral versus situational questions

Structured interviews typically use one of two question formats, or a combination of both.

Behavioral questions ask candidates to describe something they have actually done. The most common framing is: tell me about a time when you had to do X. The underlying logic is that past behavior is the best available predictor of future behavior. A candidate who has navigated a specific kind of situation before is more likely to navigate it well again than someone who has not.

Situational questions describe a hypothetical scenario and ask what the candidate would do. These are useful for roles or situations where candidates are unlikely to have directly relevant past experience. They test reasoning and judgment rather than demonstrated track record.

Behavioral questions generally have stronger predictive validity for experienced hires. Situational questions work better for early-career candidates who are being assessed on potential rather than history. For most mid-level and senior roles, behavioral questions anchored to the specific competencies the job requires should form the backbone of the interview. A well-written job description that clearly defines what success looks like makes it significantly easier to choose the right behavioral questions.

How to build a scoring rubric that actually works

The scoring component of a structured interview is where most implementations fall apart. A rubric that says "rate the candidate 1 to 5 on communication" does not tell an interviewer what a 3 looks like versus a 4. Without anchor descriptions for each score, two interviewers rating the same candidate will produce wildly different numbers that cannot be meaningfully compared.

A usable rubric defines what each score represents in concrete, observable terms for each question. For a question about handling conflict, a score of 2 might be defined as: the candidate described a situation but could not articulate a clear approach or did not take responsibility for their role in the outcome. A score of 4 might be: the candidate described a specific situation, named the actions they took, explained why, and described the outcome and what they learned.

Writing anchor descriptions takes time upfront. It pays off in consistency. When two interviewers independently score a candidate a 4 on the same dimension, that score means the same thing from both of them. When the rubric is vague, a 4 from one interviewer might be a 2 from another, and debrief discussions collapse into subjective argument rather than structured comparison.

For teams that struggle to build rubrics from scratch, Bridgebees' human screening service builds scorecards tailored to specific roles and delivers structured assessments that external interviewers can use as a starting point. More detail on how human shortlisting works if this is an approach worth exploring.

The debrief is where structured processes usually break down

Even companies that run structured interviews with rubrics often undermine the process at the debrief stage.

The most common failure is a debrief that starts with the most senior or most vocal person in the room sharing their opinion. Everyone else's independent assessment then shifts toward that opinion. Research on group decision-making consistently finds that sharing opinions sequentially, starting with authority figures, reduces the diversity of information in the final decision and increases conformity to whoever spoke first.

A structured debrief runs the other way. Everyone submits their scores independently before the group meets. The debrief starts with a summary of those independent scores rather than with anyone's opinion. Significant disagreements between evaluators are discussed explicitly: what did the person who scored a 4 see that the person who scored a 2 did not? That conversation produces useful information. Opinion-leading at the start of a group discussion produces consensus without accuracy.

The structured debrief is also the moment to revisit whether workstyle fit was actually assessed, not just assumed. Candidates who perform well in structured behavioral interviews and who also genuinely match the team's work style are the ones most likely to produce and stay.

What to do when you cannot run a fully structured process

For lean hiring teams or companies filling roles quickly, building a full structured interview process for every role is not always realistic. That is an honest constraint, not an excuse to abandon structure entirely.

The minimum viable version of a structured interview is this: decide on three to five questions in advance. Ask every candidate the same questions. Score each candidate on those specific questions before discussing them as a group. That is not a perfect structured process. It is dramatically better than a free-form conversation evaluated entirely by impression.

For teams that want the benefits of structured evaluation without building the process themselves, on-demand screening services handle this layer. A trained screener conducts a consistent structured phone screen with every candidate against a defined scorecard, then returns a ranked shortlist to the hiring team. The team gets structured signal without having to design and run the process internally. The cost of a single bad hire almost always exceeds the cost of an external screening service by a significant margin. When framed that way, the investment is easy to justify.