There are hundreds of behavioural questions and about five situations. Once you see which situation a question is reaching for, preparing stops being a memory exercise and becomes a much smaller job: find five real stories, know what each one proves, and let the question decide which one to open.
The twenty below are the ones that recur, grouped by the situation rather than by wording. What matters in each group is the second column — what the interviewer is actually checking, which is rarely what the question literally asks.
Conflict and disagreement
What they're checking: whether disagreement stays about the work, and whether you can be wrong without it being a crisis.
- Tell me about a time you disagreed with your manager.
- Describe a conflict with a colleague and how it ended.
- Tell me about a time you had to work with someone difficult.
- When did you last change your mind about something important at work?
Question 4 is the sorting question in this group and the one candidates handle worst, because the instinct is to describe changing your mind as a sign of open-mindedness in general terms. The strong answer is specific about what changed it — "he showed me the incident log and I'd been arguing from memory" — and, crucially, says what it cost to be wrong in public.
Question 1 has a trap: answers that end in "and my manager agreed I was right" fail it. The question is about how you disagree, not about whether you win. An answer where you lost the argument and executed the decision anyway is usually stronger.
Failure and mistakes
What they're checking: whether you can name a real one, and whether the lesson is specific enough to be true.
- Tell me about a time you failed.
- Describe a decision you'd make differently now.
- When did you miss a deadline, and what happened?
- Tell me about feedback that was hard to hear.
The rule across this whole group: the failure has to be yours. A story where the cause was someone else's mistake, a reorganisation or an impossible deadline answers a different question — and interviewers notice, because they ask this one specifically to find out whether you can locate yourself in a bad outcome.
Question 8 is worth preparing separately. The answer needs an actual piece of feedback, ideally quoted, and the detail that makes it land is usually your first reaction: "I thought it was unfair for about a week." Candidates who describe accepting criticism gracefully and immediately are describing something that doesn't happen.
Influence and ownership
What they're checking: whether you can move things you don't control, and whether you know the difference between your work and your team's.
- Tell me about a time you influenced a decision without authority.
- Describe something you owned end to end.
- When did you take on something outside your remit?
- Tell me about a time you had to say no.
Question 12 is underrated and separates seniority quickly. Junior answers say no and stop; senior answers say no and then deal with the consequence — what the person needed, what happened to it, whether the relationship survived. A no with no aftermath described is usually a no that was never actually delivered.
Question 10 has a hidden second half: interviewers will follow up on what you didn't own. Volunteering that boundary unprompted is one of the cheapest credibility moves available in a behavioural round.
Pressure and ambiguity
What they're checking: what you do when the instructions run out.
- Tell me about the most ambiguous project you've worked on.
- Describe a time you had to make a decision without enough information.
- When did you have too much to do? How did you choose?
- Tell me about a time priorities changed underneath you.
Question 15 is a prioritisation question disguised as a resilience question, and most people answer the disguise — describing how hard they worked rather than how they chose. The answer that scores names what got dropped and who was told.
Question 14's strong version includes the thing candidates skip: what would have changed your decision. "I'd have waited another week if the cost of being wrong had been reversible — it wasn't, so I went with the partial data and told the team which part I was least sure about."
Working with others
What they're checking: what you're like on a team when nothing dramatic is happening.
- Tell me about someone you helped get better at something.
- Describe a time you had to explain something technical to a non-technical audience.
- When did you join a team and have to build trust quickly?
- Tell me about a project that went well. Why did it?
Question 20 looks like a softball at the end of the loop and isn't. Asking why it went well is asking you to attribute causes, and the answers split cleanly: candidates who credit the team and name nothing specific, candidates who credit themselves, and candidates who name a structural reason — "we cut the scope in week two and never had to have that fight again". The third group is small.
Question 18 is the one where the app's whole thesis lives: the skill isn't knowing the thing, it's choosing what to leave out. For the technical-role version of that, see explaining your own project in an interview.
The follow-up that exposes an invented answer is almost always "what did the other person say?" Prepared stories are constructed from the candidate's point of view, so the other party is a role rather than a person — "the stakeholder", "my manager". Real memories contain the awkward sentence somebody actually said. Interviewers ask this without particularly meaning to test you, and it's the question that most reliably separates a story you lived from one you assembled.
The five stories
You don't need twenty answers. Pick five real situations and check that between them they cover:
| Story | Must contain |
|---|---|
| The hard call | Options you rejected, and why. Covers 9, 10, 14, 15 |
| The disagreement | Someone reasonable who wanted something else. Covers 1, 2, 3, 12 |
| The failure | Your mistake, a real cost, a specific change since. Covers 5, 6, 7, 8 |
| The mess | No brief, no precedent, and what you did first. Covers 13, 16, 19 |
| The one that worked | A structural reason, not luck or effort. Covers 11, 17, 20 |
Most people's five are all from the last two years, and that's fine. What isn't fine is five stories from the same project — the loop will ask them in sequence and the repetition is audible.
The story with no other person in it
“There was resistance from some stakeholders, but I explained the reasoning and eventually got alignment.”
Nobody in this sentence has a name, a motive or a line of dialogue. It's a summary of a conflict rather than the conflict, and the first follow-up — what did they actually say? — has nowhere to go.
The story with a quote in it
“She said 'you're optimising for a launch, I'm the one who gets paged at 3am' — and she was right, that was exactly what I was doing.”
One remembered sentence makes the whole account credible, and it gives the interviewer somewhere to go next. It also does something quieter: it shows you can repeat a criticism of yourself accurately, which is itself the thing being measured.
Preparing without over-rehearsing
The risk with behavioural prep is the opposite of the risk everywhere else. Under-prepared answers ramble; over-prepared ones sound recited, and a recited answer to question 8 — feedback that was hard to hear — is almost self-refuting.
What to fix in advance: the first sentence, the decision, and the number in the result. What to leave loose: everything else. That's enough scaffolding to stay on track and enough room to sound like you're remembering rather than retrieving.
The structure these answers sit inside, with the same story written weak and strong, is here: STAR method examples. If the loop you're preparing for is a culture or team-fit round specifically, the conflict question gets a full walkthrough here: the team culture interview.
Practical target: take your five stories and, for each, write down one thing another person said out loud. If you can't recall a single line of dialogue from any of them, they're summaries rather than memories — and the fix is to pick different stories rather than to invent better ones, because the invented quote is exactly what the follow-up question is built to find.
