Write interview questions tied directly to the tasks in the job, not generic traits. Ask for a specific past example or a realistic work sample, then probe with follow-up questions about their actual role, decisions and results. Score answers against a written rubric so you compare evidence, not charisma, and you will spend far less time guessing after the fact whether someone can really do the job.

Why most interview questions fail to measure skill

Most interview questions are recycled from a list found online, and most of them measure how well someone interviews, not how well they work. Questions like \"what is your greatest weakness\" or \"where do you see yourself in five years\" invite rehearsed, socially polished answers that reveal almost nothing about competence.

The deeper problem is that these questions are not anchored to the job. If you cannot point to a specific task, decision or output that a question is designed to test, it is a filler question, and filler questions are exactly where unconscious bias creeps in because interviewers fall back on gut feeling to judge the answer.

Research on hiring practices consistently finds that unstructured interviews, where every candidate is asked different questions and judged informally, are far less reliable at predicting job performance than structured approaches built around consistent, job-related questions. The fix is not cleverer questions, it is a better process for writing them.

Start from the job, not a question bank

Before writing a single question, list the four or five things the person in this role will actually spend their time doing in the first ninety days. These become your \"must-have\" skills, and every interview question should map back to one of them.

For example, if you are hiring a customer support lead, the real tasks might be: handling an escalated complaint, coaching a junior rep after a mistake, prioritising a backlog of tickets, and reporting trends to management. Each of those becomes the seed for one or two interview questions, rather than asking generic questions about \"customer service passion.\"

This is also the point to decide what \"good\" looks like for each task, in writing, before you meet a single candidate. If you cannot describe a strong answer in a sentence or two, the question is not specific enough yet.

The anatomy of a skill-revealing question

A strong interview question usually has three parts: it anchors to a real situation, it asks for the candidate's specific role and reasoning, and it invites a measurable outcome. Vague prompts like \"tell me about a time you led a team\" are a start, but they need tightening.

Compare these two versions of the same question:

Weak: \"Tell me about a time you dealt with a difficult customer.\"Strong: \"Walk me through the last time a customer escalated a complaint to you directly. What had already gone wrong before it reached you, what did you say to them, and what changed afterwards, either for that customer or for your process?\"

The strong version forces specificity. A candidate who has not actually done this work will struggle to answer in detail, while someone who has done it will naturally produce dates, numbers, names of tools, and a clear before-and-after.

Behavioral, situational and work-sample questions, and when to use each

There are three main question types, and a well-rounded interview usually uses all three.

  • Behavioral questions ask about a real past event: \"Tell me about a time you had to...\" These are strongest for experienced candidates with a track record to draw on.
  • Situational questions present a hypothetical scenario: \"If a client asked you to skip a quality check to hit a deadline, what would you do?\" These work well for junior candidates or career changers who have less direct history to reference.
  • Work-sample questions ask the candidate to actually do a scaled-down version of the job, such as reviewing a sample email, sketching a process, or solving a short case. These are often the single best predictor of on-the-job performance because they measure the skill directly instead of a description of it.

For most roles, aim for two behavioral questions, one situational question, and one small work-sample task if the role and time allow it. This combination catches candidates who talk well but have not actually done the work, and it also gives quieter candidates who interview poorly a fair chance to show what they can do.

Build a scorecard before you interview anyone

Write your questions alongside a simple rubric: for each question, note what a weak, adequate, and strong answer sounds like. This takes fifteen minutes and pays for itself many times over, because it stops the interview from becoming a popularity contest.

A basic scorecard row might look like this:

Question: \"Walk me through how you prioritised your workload during your busiest week last quarter.\"Weak answer: vague, no specifics, blames others for disorganisation.Adequate answer: describes a reasonable method but with little detail on trade-offs made.Strong answer: names a clear framework or method, gives a concrete example of something they deliberately deprioritised, and explains the reasoning and the outcome.

Score every candidate against the same rubric right after the interview, while the answers are fresh, rather than relying on memory during a group debrief days later.

Sample questions by common skill area

Use these as starting templates, then adapt the details to your actual role and industry.

  • Problem-solving: \"Describe a project that did not go the way you expected. What signal told you it was going wrong, and what did you change?\"
  • Communication: \"Tell me about a time you had to explain something technical or complex to someone who did not have that background. How did you check they understood?\"
  • Ownership and accountability: \"Tell me about a mistake that was genuinely your fault. What did you do after you realised it, and what did you change so it did not happen again?\"
  • Collaboration: \"Describe a disagreement you had with a colleague about how to approach a piece of work. How was it resolved, and what was the outcome?\"
  • Adaptability: \"Tell me about a time your priorities changed suddenly. How did you decide what to drop or delay?\"

Notice that none of these ask the candidate to describe themselves in the abstract. Every one of them requires a real, specific memory, which is much harder to fake convincingly than a general claim.

Follow-up probes that separate real experience from rehearsed answers

The first answer a candidate gives is often a polished, general version of the story. The real signal comes from your follow-up questions, which push past the summary into the details a candidate cannot easily invent on the spot.

Useful probes to keep in your back pocket for almost any question include:

  • \"What exactly did you personally do, versus the rest of the team?\"
  • \"What was the actual result, in numbers if you have them?\"
  • \"What would you do differently if you faced that same situation again tomorrow?\"
  • \"Who disagreed with your approach, and what did they say?\"
  • \"What was the hardest part of that, specifically?\"

Candidates who genuinely did the work tend to answer these follow-ups quickly and specifically. Candidates who are stretching the truth, or describing someone else's work as their own, often hesitate, generalise, or shift the story slightly under a follow-up question.

Common mistakes to avoid

Even well-intentioned interviewers slip into habits that undo good question design. Watch for these:

  • Asking leading questions. \"You're comfortable with tight deadlines, right?\" invites a yes and tells you nothing.
  • Letting the first good answer end the topic. Always probe at least once before moving on.
  • Comparing candidates to each other instead of to the rubric. This is where recency bias and personal chemistry quietly take over.
  • Reusing the same one or two questions for every role. A question that works for a sales role rarely tells you much about an operations role.
  • Skipping the scorecard under time pressure. This is exactly when structured scoring matters most, because memory fades fast between interviews.
  • Asking about protected characteristics indirectly, such as questions about family plans, age, or health, which are restricted in most jurisdictions. Always check your local labour regulations on what can and cannot be asked in an interview before finalising your question set.

If you are running interviews for a small team without a dedicated HR function, it also helps to standardise the process itself, not just the questions. Tools like Hyrewell let you post one application link, automatically screen applicants against your must-have skills with evidence from their answers, and let shortlisted candidates self-book interview slots, so you spend your interview time actually asking these deeper questions instead of chasing logistics.

Bringing it together

Good interview questions are not clever, they are specific. They come directly from the real tasks of the job, they ask for a genuine past example or a hands-on sample of work, and they are followed by probes that make it hard to bluff. Pair that with a written scorecard and you turn interviewing from an art into a repeatable, fair process.

The extra thirty minutes it takes to write role-specific questions and a simple rubric before your first interview will save you far more time later, both in avoiding a bad hire and in being able to explain, clearly and fairly, why you chose the candidate you did.