Answering behavioural interview questions with the STAR method
STAR fails when candidates spend a minute on setup and ten seconds on what they actually did. The ratio is the whole technique.
Behavioural questions — 'tell me about a time when…' — are asked because past behaviour predicts future behaviour better than hypotheticals do. Interviewers are scoring for specificity and for your personal contribution.
The ratio that matters
- Situation: 15% — just enough context to understand the stakes.
- Task: 10% — what you specifically were responsible for.
- Action: 60% — what you did, in sequence, in the first person.
- Result: 15% — the outcome, quantified where you honestly can.
The most common failure is a two-minute preamble about company structure followed by 'and so we fixed it'. Interviewers cannot score 'we'.
Worked example: conflict
Situation: Our release cadence slipped because QA and engineering disagreed on what counted as a blocker. Task: I owned the release calendar, so the delay landed on me. Action: I set up a 20-minute session with both leads, wrote down the three disputed cases, and proposed a written severity definition with an explicit escalation path. I circulated it, took two rounds of edits, and got sign-off in a week. Result: The next four releases went out on schedule, and the definition is still what the team uses.
Worked example: failure
Choose a real failure with a real cost, not a disguised strength. Interviewers are listening for whether you diagnosed the cause and changed a system, not just for contrition. End on the change, not the apology.
Questions to have covered
- Tell me about a difficult stakeholder.
- Tell me about a time you missed a deadline.
- Describe a time you had to learn something fast.
- Tell me about a decision you made with incomplete information.
- Describe a time you disagreed with your manager.
- Tell me about work you're proud of.
Common traps
- Saying 'we' when the question asked about you.
- Choosing a story so old it predates your current skill set.
- Inventing metrics — a vague honest result beats a precise invented one.
- Running past two minutes without checking whether they want more.
Close long answers with 'happy to go deeper on any part of that.' It hands control back and signals you can read a room.