Behavioral questions ask what you did, not what you would do. The premise is that past behaviour predicts future behaviour better than a hypothetical does.
Which means the interviewer is listening for evidence, not intent. “I would escalate early” is worth nothing next to “I escalated on day two, and here is what happened.”
Situation, Task, Action, Result. The structure exists because without it people describe context forever and never say what they personally did.
The failure mode is spending ninety seconds on Situation. Set the scene in two sentences. The interviewer needs just enough context to judge the actions — anything more is your own airtime.
Action is the part being marked, and it needs to be in the first person. Teams do things in every answer; the question is what you did.
Result should be quantified where it can be, and honest where it cannot. “It shipped six weeks late but we cut the defect rate by half” is a stronger answer than a clean success, because it sounds like something that happened.
You cannot predict the question, but you can predict the themes: conflict, failure, leadership without authority, a deadline you missed, a decision made without enough information.
Six or seven real stories will cover almost any behavioral round, because most stories answer several questions with a change of emphasis. The same project can be your conflict story and your ambiguity story.
Write them down once. Not scripts — bullets, with the numbers, because the numbers are what you will forget under pressure.
“Tell me about a time you failed” is testing whether you can assess your own work honestly. A disguised success (“I work too hard”) fails it immediately.
Pick a real failure with a real cost, say what you would do differently, and show the change stuck. The learning is the answer; the failure is just the setup.
The interviewer is hiring you, not your old team. Say what you did, even when it feels immodest.
That is not a behavioral answer. Pick one specific occasion, with a date and a place.
Choose something that genuinely went wrong. The point is your judgment about it, and a safe answer shows none.
Balanced and personable — the classic on-site interview. You will be scored on communication, technical accuracy, confidence, with the transcript attached.
The first question in almost every interview, and the one most people lose the room on. Here is the structure that works and the timing that matters.
The system design round rewards trade-off reasoning and clarifying questions far more than knowing a specific architecture. Here is how it is marked.