Structured Interviews: Well Proven, Rarely Used

One of the odder things about hiring is that the field has known for decades which kind of interview works best, and most companies still don't run it.

The kind that works is the structured interview. You ask every candidate the same questions, in the same order, and score the answers against a rubric you wrote before anyone walked in. It sounds a little bureaucratic, which is part of why people resist it. But in Schmidt and Hunter's meta-analysis, the study most people cite when they talk about which hiring methods predict job performance, structured interviews came out at a validity of about 0.51. That puts them near the top, in the same range as work samples and cognitive ability tests, and ahead of unstructured interviews and of easy tools like reference checks, which came in at 0.26.

So why doesn't everyone use them? I don't think it's ignorance. Most recruiters and hiring managers have heard that structure beats gut feel. The reason is cost. A structured process needs question sets written for each role, interviewers trained to stick to them, and scoring that means the same thing from one reviewer to the next. None of that pays off this week, and all of it decays unless someone keeps it up. So nobody keeps it up.

What most companies run instead looks like this. One recruiter does the first screen with no rubric and a loose brief to see whether the person seems like a fit. The questions change from call to call. The notes are impressions, and the hiring manager downstream rarely sees them in a form they could compare. Everyone treats this as gathering information about candidates, but what it produces can't be lined up side by side.

Volume makes the choice harder. Suppose 300 people apply for a role and you want to give each a proper structured screen of half an hour or so. That's around 150 hours of calls, roughly two recruiters doing nothing else for two weeks. Faced with that, teams make a sensible-looking trade: a quick, loose screen for everyone rather than a rigorous one for a few. It's hard to blame them. But the cost of that trade is easy to miss.

The cost is that the weakest step ends up deciding the most. Plenty of companies are proud of their final-round panels. Yet the panel only ever sees the dozen people who made it through the first screen. If that cut, from 300 down to 12, was made on one person's instinct in fifteen-minute calls where each candidate got different questions, then much of the real selection happened there. A disciplined panel can pick the best of the twelve. It can't bring back the people who should have been among them.

This is what I find most interesting about the research. We tend to discuss structured interviewing as a question about the final round. But the place it may matter most is the first conversation, which is exactly the place where volume made structure impractical.

That first conversation is the gap we built Asendia to close, and it's why we made it a voice AI rather than a questionnaire. A form would give you consistency cheaply: same questions, same order, every time. But a form can't ask a follow-up, and a lot of what makes an interview informative is the interviewer noticing a vague answer and asking what the person actually did. Speaking is also harder to script than typing, so you hear more of how someone thinks and less of what they rehearsed. So Asendia phones each applicant and talks with them, nights and weekends included. Every candidate for a role gets the same core questions in the same order, and the follow-ups adapt to what the person says while staying tied to the criteria your team set. The rubric doesn't drift because someone is running late or took a liking to one candidate. What lands in the ATS is a qualification summary for each person, with verbatim excerpts from the call attached so a recruiter can see what was actually said.

That changes what a shortlist is. Instead of notes from different people asking different questions on different days, a recruiter gets something that can be compared across candidates, because it was produced the same way for all of them. For an agency running a large campaign, that holds even when a weekend's worth of applicants arrives at once. I've written separately about the difference between AI tools that help with a step and tools that actually carry it out, in the post on agentic recruiting.

The research on structured interviews hasn't changed in a long time. What has changed is how much it costs to follow it at the first stage. When the cost of doing the right thing drops that far, it's worth asking what reason is left for doing it the old way.

Ready to transform your hiring strategy? Schedule a Demo with our founders today!