Interview Questions That Actually Predict Job Performance

Most interviews feel productive and predict almost nothing. Decades of research on hiring consistently finds that unstructured interviews — different questions for each candidate, judged on gut feel — have one of the weakest correlations with actual job performance of any common hiring method, worse than a simple work sample test and only slightly better than chance for some roles.

The techniques that actually work aren't exotic. They're just more structured than how most small businesses naturally interview.

What actually correlates with performance

Structured behavioral questions, asked consistently across candidates, about specific past situations rather than hypotheticals. "Tell me about a time you had to deliver bad news to a client" predicts more than "how would you handle a difficult client," because past behavior in a real situation is a better predictor of future behavior than a hypothetical answer someone can construct on the spot to sound good.

A real work sample. Having a candidate actually do a scaled-down version of the job — draft a sample email, work through a real (anonymized) problem the role would face — consistently outperforms almost any interview question format alone, because it observes actual capability rather than a description of it.

Consistent questions across candidates, scored against the same criteria. This is less about any single question and more about comparing candidates on the same basis rather than letting each interview drift toward whatever felt natural in the moment.

What doesn't work as well as it feels like it should

Brain teasers and puzzle questions. These measure comfort with an artificial format under pressure, which has a weak relationship to job performance for the large majority of roles, and disadvantages candidates who simply haven't encountered that specific style of question before, for reasons unrelated to how good they'd be at the job.

Purely conversational, unstructured interviews. They feel informative because rapport feels like insight, but "I really liked them" is one of the weaker predictors available, and it's also where unconscious bias has the most room to operate unchecked.

Questions about weaknesses answered as strengths. "What's your biggest weakness" reliably produces a rehearsed, low-information answer ("I care too much") that predicts nothing about actual performance.

A structure that works for a small business

Round one: skills and experience, structured around three or four behavioral questions specific to the role's real demands, asked the same way to every candidate for that opening.

Round two: a work sample plus team fit. A short, realistic task relevant to the actual role, followed by a conversation with whoever the person would work with most closely day to day.

Two rounds is enough for most roles below a senior level. Adding a third or fourth round rarely improves the decision meaningfully and often costs a small business a strong candidate who accepts a faster-moving offer elsewhere in the meantime.

Behavioral questions worth adapting to your role

• Tell me about a time you had to figure something out with incomplete information. What did you do?
• Describe a situation where you disagreed with a decision at work. How did you handle it?
• Tell me about a mistake you made and how you addressed it afterward.
• Describe a time you had to manage competing priorities with a tight deadline.
• Tell me about a time you had to explain something complicated to someone without much context.

The specific questions matter less than asking about real, specific past situations and following up with "what did you actually do" rather than accepting a vague or aspirational answer.

Keeping it consistent, without a formal system

Consistency doesn't require expensive hiring software at small-business scale. A shared document with the questions for the role, a simple scoring scale (even 1 to 4 per question), and a rule that every interviewer fills it out right after the conversation — before comparing notes with anyone else — captures most of the benefit of a formal structured process without needing to buy a platform for it.

What this buys a small business specifically

At a company small enough that one bad hire is deeply felt — a five-person team absorbs one mismatched hire far more painfully than a five-hundred-person one does — the return on a slightly more structured interview process is proportionally larger. The hour spent writing consistent behavioral questions in advance is small compared to the cost of a hiring mistake that a better-structured interview could plausibly have caught.

Involving the future team, deliberately

Having a candidate meet with a future teammate, not just the hiring manager, adds real information a single interviewer can miss — how the candidate engages with someone at their own level rather than someone evaluating them, and a second, independent read on fit. This works best when the teammate is given specific guidance on what to probe for, rather than a purely open-ended conversation, so their input adds a distinct data point instead of just replicating whatever the hiring manager already assessed.

Avoiding common bias traps in the moment

Even a structured process leaves room for bias to creep in through small, easy-to-miss habits: giving a strong first impression more weight than it deserves and unconsciously interpreting everything afterward to fit it, or comparing each candidate only to the one interviewed immediately before rather than against the actual role requirements. Scoring each candidate against the same written criteria immediately after their interview, before discussing them with anyone else on the hiring team, is one of the simplest available defenses against both patterns, since it captures an independent judgment before group discussion has a chance to anchor everyone to the same impression. A shared framework for triaging concerning signals helps here too, so the panel isn't each applying a different private threshold for what should end consideration versus what's simply worth a follow-up question.

The short version

Structured behavioral questions about real past situations, plus an actual work sample, predict job performance meaningfully better than unstructured conversation or clever puzzle questions — and they cost nothing beyond planning the questions in advance and asking them consistently. For a small business where every hire matters disproportionately, that's a high-leverage half hour of preparation.