What Problem Solving Interview Questions Should Reveal
Problem solving is the ability to take a situation that is not working, find out why, and change it so that it does. It is not the same as being clever, being fast, or having good ideas. The person who solves problems well is the one who slows down at the start to work out what the problem actually is, follows the evidence to a cause rather than a symptom, and stays with the fix until it has demonstrably worked.
Most interviewers assess this badly in the same way. They ask “tell me about a problem you solved” and accept a story that starts with a solution. The candidate explains what they did, the interviewer nods at the outcome, and nobody asks how the candidate knew that was the problem, what else it could have been, or whether the fix held. The result is a good score for storytelling and no evidence of diagnosis. The other common error is to use puzzles and brain-teasers, which test performance under artificial pressure and tell you almost nothing about how someone handles a messy, political, slow-moving problem at work.
This guide covers 20 interview questions across 5 dimensions: Defining the Problem Before Solving It, Root-Cause Diagnosis, Structuring an Unfamiliar Problem, Testing Assumptions and Iterating and Seeing a Fix Through to Resolution. Fifteen are behavioral and five are situational. Every behavioral question comes with follow-up probes, because the opener only gets you the story. The probes get you the method.
Problem solving sits next to two other competencies on this site, and it is worth keeping them apart. Decision making is about choosing between options when the trade-offs are known; this guide is about working out what the options are in the first place. If you need to assess judgment under uncertainty, use the decision making interview questions alongside these.
Defining the Problem Before Solving It
What good looks like
A strong problem solver resists the pull to start fixing. They ask what the problem actually is, who it is a problem for, and how they will know when it is solved. They can describe a time the stated problem turned out to be a symptom, and what they did once they realized. Be alert to the candidate who describes every problem as obvious from the start. Real problems rarely arrive labeled.
Behavioral questions
- Tell me about a time the problem you were asked to solve turned out not to be the real problem.
- What was the original ask, and who made it?
- What told you the framing was wrong?
- How did you re-describe the problem, and to whom?
- What would have happened if you had solved the original version?
- Tell me about a problem where you had to work out what “solved” would look like before you could start.
- Who disagreed about what a good outcome was?
- How did you pin down the success criteria?
- What did you leave out of scope, and why?
- Did the definition hold, or did it move again?
- Tell me about a time you pushed back on a solution someone handed you along with the problem.
- What was the proposed solution, and why did it not fit?
- How did you separate the problem from the person’s preferred answer?
- What did you propose instead?
- How did you keep the person on side while you did it?
Situational scenario
A team lead comes to you and says, “Our onboarding is broken. New hires are quitting in the first three months. We need a better onboarding program.” You have been asked to own the fix.
What do you do in the first week?
Then introduce a constraint: “You pull the exit data and find that two-thirds of the early leavers came from one hiring manager's team. Does that change your plan?”
What to listen for
whether they investigate the problem before accepting the proposed solution, what evidence they would gather first, whether they define what "fixed" means in measurable terms, and whether the new information makes them re-scope the problem rather than bolt an extra step onto the onboarding program.
Root-Cause Diagnosis
What good looks like
A strong problem solver separates causes from symptoms and keeps asking why until the answer is something they can act on. They use data where it exists and go and look where it does not. They can name a time their first diagnosis was wrong and explain what they missed. Candidates who always found the cause on the first attempt are describing easy problems, or editing the story.
Behavioral questions
- Tell me about a recurring problem you traced back to its root cause.
- How many times had it been “fixed” before you looked at it?
- What did you do differently from the people who looked at it before?
- What was the actual cause, and how did you confirm it?
- Did it come back?
- Tell me about a time your first diagnosis of a problem was wrong.
- What did you think was causing it, and why was that plausible?
- What evidence eventually contradicted you?
- How much time or money went into the wrong fix?
- What do you check now that you did not check then?
- Tell me about a problem you diagnosed without the data you would normally rely on.
- What was missing?
- What did you use as a substitute: observation, conversations, a small experiment?
- How confident were you in the cause before you acted?
- How did the diagnosis hold up afterward?
Situational scenario
Your team’s weekly report has been wrong three weeks in a row. Each time, someone finds the error, corrects the number by hand, and sends it out late. Your manager asks you to “make sure it’s right this week.”
How do you approach it?
Then introduce a constraint: “You find the report pulls from two systems that define 'active customer' differently. Changing either definition will upset a different team. What now?”
What to listen for
whether they treat the request as a diagnosis task rather than a checking task, whether they look for why the error keeps recurring, how they handle a root cause that sits in someone else's territory, and whether they propose a durable fix rather than a fourth manual correction.
Structuring an Unfamiliar Problem
What good looks like
A strong problem solver can take a problem they have never seen before and break it into parts they can work on. They decide what to tackle first and why, and they are explicit about what they do not know yet. Look for a method the candidate can describe, not a claim that they are “good at figuring things out.”
Behavioral questions
- Tell me about a problem you took on with no relevant experience to draw from.
- What did you do in the first day or two?
- How did you break it into pieces?
- Who or what did you learn from fastest?
- What did you get wrong because of what you did not know?
- Tell me about a problem that was too big to solve all at once.
- How did you decide where to start?
- What did you deliberately leave until later?
- How did you keep the pieces from drifting apart?
- What did the first finished piece teach you about the rest?
- Tell me about a time you had to explain a messy problem to people who needed a simple version.
- Who was the audience, and what did they need to decide?
- What did you leave out of the simple version?
- Did the simplification cause any problems later?
- How did you keep the full picture for yourself?
Situational scenario
Your company has acquired a smaller business. You have been asked to work out how to combine the two companies’ hiring processes into one. You know your own process well and nothing about theirs. You have six weeks.
Where do you start?
Then introduce a constraint: “In week two you discover the acquired company's process is legally required for one of their government contracts and cannot be changed. How does that affect your plan?”
What to listen for
whether they structure the work into stages before diving in, whether they spend early time understanding the unfamiliar process, how they handle a hard constraint that appears part way through, and whether they can say what the first two weeks produce in concrete terms.
Testing Assumptions and Iterating
What good looks like
A strong problem solver knows that the first plan is a hypothesis. They look for the cheapest way to test it, pay attention to what the test says, and change course when the evidence tells them to. They can describe an assumption they held that turned out to be false and what it cost. Watch for the candidate who presents every solution as right first time.
Behavioral questions
- Tell me about a time you tested a solution on a small scale before rolling it out.
- What was the assumption you were trying to test?
- How did you design the test so it would tell you something?
- What did the result show, and did it surprise you?
- What changed between the pilot and the full version?
- Tell me about an assumption you made about a problem that turned out to be wrong.
- What was the assumption, and why did it seem safe?
- When and how did you find out?
- What had you already built on top of it?
- How do you surface assumptions like that earlier now?
- Tell me about a problem you solved on the second or third attempt.
- What did the first attempt get wrong?
- What did you learn from it that shaped the next one?
- How did you keep stakeholders with you while you iterated?
- At what point would you have stopped?
Situational scenario
You are convinced that the reason candidates drop out of your hiring process is the length of the application form. You could rebuild the form in a week. Your manager is supportive but points out that nobody has actually asked the candidates who dropped out.
What do you do?
Then introduce a constraint: “You contact 20 drop-outs. Twelve say the form was fine but the role description was confusing. Four mention the form. What now?”
What to listen for
whether they test the assumption before building, how they would gather evidence quickly and cheaply, whether they change their plan when the evidence points elsewhere, and whether they can hold a confident hypothesis loosely.
Seeing a Fix Through to Resolution
What good looks like
A strong problem solver does not stop at the proposal. They implement the fix, check that it worked, and deal with the side effects. They can tell you how they knew the problem was solved and what they measured. Be cautious with candidates whose stories end at the recommendation; solving a problem on paper is a different skill from solving it.
Behavioral questions
- Tell me about a problem you solved where the hard part was the implementation, not the idea.
- What made the implementation hard?
- Who had to change what they did for the fix to work?
- What slipped, and how did you recover it?
- How did you confirm the fix was working?
- Tell me about a fix you put in place that caused a new problem.
- What was the original fix?
- How long before the side effect showed up?
- How did you weigh the new problem against the old one?
- What do you now check before rolling out a fix?
- Tell me about a time you had to keep a problem open when everyone else considered it closed.
- Why did you think it was not solved?
- What evidence did you gather to make the case?
- How did you avoid being seen as the person who would not let it go?
- What was the eventual outcome?
Situational scenario
Three months ago you fixed a bottleneck in how offers get approved: average approval time dropped from nine days to two. This week a hiring manager tells you offers are “taking forever again.” You check the dashboard and the average still says two days.
How do you handle it?
Then introduce a constraint: “You find that the average is right, but a handful of senior-level offers are taking three weeks because they now need an extra sign-off that was added after your fix. What do you do?”
What to listen for
whether they take the complaint seriously despite the data, whether they look at the distribution rather than the average, how they deal with a process change that someone else introduced on top of their fix, and whether they treat "solved" as a state that has to be maintained.
Red Flags to Watch for When Assessing Problem Solving
The story starts with the solution. Ask how they knew what the problem was and you get a restatement of the fix. A strong problem solver can describe the moment the problem became clear, and it is usually later than the moment they were handed it.
Every diagnosis was right first time. Nobody who has solved hard problems has a perfect record. A candidate who cannot name a wrong diagnosis has either worked on easy problems or is not telling you about the hard ones.
The method is a personality trait. “I’m just very logical” or “I’m good at figuring things out.” Real problem solvers can describe what they do: the questions they ask, the data they pull, the small tests they run. If the method cannot be described, it cannot be repeated.
The problem was solved alone. Workplace problems involve other people’s systems, data, and priorities. A story with no one else in it is either a small problem or an incomplete account.
Nothing was measured. The fix “worked” because the complaints stopped or because nobody mentioned it again. Strong candidates can tell you what they checked afterward and what the number was.
They cannot say what they would do differently. Every solved problem teaches something about the next one. A candidate with no reflection on the process is unlikely to improve at it.
How to Use These Questions in a Structured Interview
Pick one behavioral opener from each of three dimensions, or two openers plus one situational scenario, and ask the same set to every candidate for the role. Twenty questions is a bank, not an interview. Three questions probed properly will give you more evidence than eight asked flat.
Score with a 1-to-5 rubric agreed before the first interview. A 1 is a story that starts with the solution and ends with “and it worked.” A 3 is a specific problem with a visible method: you can see how they found the cause and what they tried. A 5 adds the things most candidates leave out: a wrong turn, a measured result, and something they do differently now because of it.
The probes do the work. “Tell me about a recurring problem you traced to its root cause” is easy to prepare for. “How many times had it been fixed before you looked at it?” and “did it come back?” are not. Ask the opener once, then go down through the probes rather than moving on. If the candidate’s answer collapses under the second probe, that is your finding.
Situational scenarios test something different from behavioral questions. They show you how the candidate structures a problem in real time, and the constraint you introduce part way through shows whether they can re-plan when the picture changes. Use at least one.
Calibrate across your panel before the interviews, not after. Have each interviewer score the same recorded answer independently and compare. If two interviewers are more than one band apart on the same answer, the rubric needs tightening before anyone else is interviewed. Interview Intelligence was built to make that calibration visible across a whole hiring team.
Ready to Assess Problem Solving at Scale?
Frequently asked questions
What's the difference between problem solving and decision making?
Problem solving is working out what is wrong and what could fix it. Decision making is choosing between options once they are on the table. A candidate can be strong at one and weak at the other: some people diagnose brilliantly and then stall when it comes to committing; others decide quickly on a problem they never properly understood. This guide tests the diagnosis and the fix. Use the decision making guide to test the choice.
Should I use brain-teasers or puzzles to test problem solving?
No. Puzzles test how someone performs on an artificial task under time pressure with a known answer. Workplace problems are slow, political, and have no answer key. Behavioral questions about real problems, followed by probes about the method, tell you far more. If you want to see live reasoning, use one of the situational scenarios in this guide, which are built around realistic work situations.
How do I assess problem solving in a candidate with limited work experience?
Ask the same questions and accept examples from study, volunteering, or personal projects. The dimensions do not change: did they define the problem, find a cause, structure the work, test an assumption, and see it through? A graduate who can describe how they diagnosed why a group project stalled is giving you better evidence than an experienced candidate who describes a solution with no diagnosis.
How many problem solving questions should I ask in one interview?
Three, probed properly. One from the diagnosis end (dimensions 1 or 2), one from the structuring or testing end (dimensions 3 or 4), and one situational scenario. That gives you evidence on both halves of the competency in about 25 minutes.
What does a strong answer to a problem solving interview question sound like?
It names a specific problem, describes the moment the real problem became clear, explains how the candidate found the cause and what they tried, includes at least one thing that did not work, and ends with how they confirmed the fix held. Vague answers describe the outcome. Strong answers describe the method.