The default interview at most companies goes something like this. The interviewer glances at the resume a few minutes beforehand, walks in, and has a friendly conversation. They ask whatever comes to mind, follow the threads that seem interesting, and leave with a feeling. That feeling gets summarised afterward as a thumbs up or a thumbs down, and if pressed, the interviewer will say the candidate seemed sharp, or was a good fit, or did not quite click. Almost everyone hires this way. It is also one of the weakest ways to predict whether someone will actually be good at the job.
The uncomfortable truth, backed by decades of research into hiring, is that the unstructured interview mostly measures how much the interviewer enjoyed the conversation. It rewards people who are relaxed, articulate and similar to the person across the table, none of which is the same as being able to do the work. The good news is that the fix is well understood, not expensive, and not complicated. It is called the structured interview, and once you have seen how much stronger it is, the old way starts to look like guessing with extra steps.
What the evidence actually says
This is not a matter of opinion or management fashion. When researchers compare hiring methods against how people later perform on the job, structured interviews consistently land near the top of the list, well ahead of unstructured ones. The gap is not small. A structured interview is roughly twice as predictive of job performance as a free-form chat. That is an enormous difference for something that costs nothing but a little discipline to adopt.
The reason for the gap is not that structured interviewers are smarter. It is that structure removes the things that quietly corrupt an unstructured interview: the pull of a strong first impression, the tendency to ask easier questions of people you like, the way a single memorable moment can colour a whole judgment. Structure does not make interviewers robotic. It makes their judgment about the job instead of about the chemistry in the room. Everything below is just the mechanics of getting there.
What structured actually means
Structured does not mean rigid or scripted to the point of being cold. It means three specific things are true about how you run the interview:
- Every candidate for a role gets the same core questions. You can still follow up and dig into individual answers, but the backbone is consistent, so you are comparing people on the same ground.
- The questions are tied to what the job actually requires, not to trivia, puzzles, or whatever the interviewer finds interesting that day. If a question does not help you predict performance in this role, it does not earn a place.
- Each interviewer scores the candidate against defined criteria, and does it independently, before the panel gets together to talk. The scoring is not a vibe recorded after the fact. It is a judgment against a rubric you agreed on in advance.
That is the whole of it. Same questions, tied to the job, scored independently against clear criteria. None of it requires special training or software. It requires deciding what good looks like before you start, and then holding yourself to it.
Why it works
The first reason structure works is comparability. When every candidate answers the same job-relevant questions, you can actually compare their answers, because they are answers to the same thing. In an unstructured process, one candidate got asked about a hard technical trade-off and another got asked about their favourite project, and now you are trying to compare two conversations that had nothing in common. Comparability sounds mundane, but it is the entire basis of a fair decision. You cannot rank people fairly on different tests.
The second reason is that independent scoring defends the decision against the loudest voice in the room. In a typical debrief, the first person to speak, or the most senior, or the most confident, sets an anchor that everyone else drifts toward. A quiet doubt gets swallowed because the room already seems to have decided. When each interviewer commits their score before the discussion, that anchoring cannot happen. You start the debrief with everyone's honest, independent read already on the table, and then you talk about the differences. You end up debating evidence rather than deferring to whoever spoke first.
The scorecard, spelled out
The tool that makes all of this concrete is the scorecard, and you do not need a fancy one. A scorecard is simply the list of criteria that matter for the role, with a way to rate each candidate against them and a place to record why. Here is a version you can copy and adapt for almost any role.
Pick four or five criteria that genuinely predict success in the job. For a lot of roles that means something like: the core skill the role is built on, problem-solving, communication, and collaboration. For each criterion, rate the candidate on a simple scale of one to four, where one is a clear miss, two is below the bar, three is solid, and four is exceptional. Deliberately use an even number of points so there is no lazy middle to hide in; a four-point scale forces a lean one way or the other. Next to each rating, write a single line of evidence: the specific thing the candidate said or did that justifies the number. The evidence is not optional. A score without a reason is just a feeling wearing a number.
Score first, then discuss
The sequence matters as much as the scorecard itself. Every interviewer fills in their scores and evidence independently, right after their own conversation, before anyone compares notes. Only once everyone has committed do you bring the scorecards together. Now the debrief has real structure. You can see instantly where the panel agrees and where it splits.
Resist the urge to simply average the scores and move on. The average hides the most useful information, which is the disagreement. If three interviewers rated a candidate a four on problem-solving and one rated them a two, that gap is not noise to be smoothed away. It is the single most valuable thing in the whole process. It means someone saw something the others missed, in one direction or the other, and the debrief should spend its time exactly there, reading the conflicting evidence and working out who is right. A split score is a signal to dig, not a number to average.
The objections, answered
Two objections come up every time. The first is that structure will make interviews feel robotic and impersonal, and scare good candidates off. In practice the opposite tends to happen. Candidates generally experience a structured interview as fairer and more professional, because it is obvious that everyone is being asked the same real questions rather than being judged on charm. You can still be warm. You can still follow up and go deep. Structure governs the backbone of the interview, not the tone of it.
The second objection is that a small team does not have time for all this. But structure saves time, it does not cost it. Most of the wasted hours in hiring come from muddled decisions: the endless debrief where nobody can quite say why they liked someone, the reopened search after a bad hire that a clearer process would have caught. Deciding your questions and criteria once, up front, and reusing them for every candidate in the role, is far less work than reinventing the interview every time and then arguing about the results. The discipline is small and the payoff compounds.
Rolling it out without a big project
None of this needs to arrive as a formal initiative with training sessions and a change-management plan. The lightest way to start is to pick one open role, sit down for twenty minutes, and write two things: the four or five criteria that actually predict success in that role, and a small bank of job-relevant questions that get at each one. That is a structured interview. You have already done the hard part, which is deciding what good looks like before you meet anyone.
From there it spreads by example. Run the next loop with your questions and a shared scorecard, and let the interviewers feel the difference in the debrief, how much easier it is to reach a fair decision when everyone scored the same criteria independently. That experience sells structure better than any memo could. Reuse the questions and criteria for the next candidate in the role, refine them when one turns out to be weak, and build a small library of them role by role. Within a few hires you have a repeatable process that cost you nothing but the discipline of writing things down before the conversation rather than after it.
The one habit to protect from day one is independent scoring. Everything else can be rough at the start and improve with use, but the moment interviewers begin comparing notes before they have committed their own scores, the anchoring creeps straight back in and you lose most of the benefit. Score first, alone, and then talk. If you keep only one rule from all of this, keep that one, because it is the single rule that does the most work. It is the difference between a debrief that surfaces what each person genuinely saw and one that quietly ratifies whatever the most confident voice in the room claimed first.
Fair, comparable and defensible
Structure will not turn hiring into a machine, and it is not meant to. What it does is give your judgment a fair chance to be about the right thing. Same questions, so you are comparing like with like. Tied to the job, so you are measuring what matters. Scored independently against clear criteria, so the decision reflects the evidence rather than the room. The interview stops being a charisma contest and becomes a genuine read on whether this person can do the work. And because every judgment is written down with a reason attached, you can explain any decision you make, to the candidate, to the hiring manager, or to yourself six months later when you are wondering whether you got it right. That is what fair, comparable and defensible actually looks like in practice. It is not the property of big companies with recruiting departments and dedicated tooling. It is within reach of any team willing to do one modest thing, which is to decide what good looks like before the conversation starts rather than trying to reconstruct it from a feeling afterward. The structure is small and costs almost nothing to adopt. The difference it makes to who you end up hiring, and to your ability to explain why, is the opposite of small.