Behavioral questions like "tell me about a time you disagreed with a stakeholder" or "describe a time you had to push back on a request" are graded against the STAR rubric: Situation, Task, Action, Result.
You can use STAR without sounding like a customer service script.
The four parts
S — Situation. 1–2 sentences setting the scene. Specific enough to be a real story, brief enough that the interviewer doesn't tune out.
T — Task. What you specifically had to do. Don't say "we" — interviewers grade individual contribution. Use "I".
A — Action. What you actually did. This is the bulk of the answer (~60% of speaking time). Specific, concrete, walking through your reasoning.
R — Result. What happened. Quantify where possible. Even a directional outcome ("the marketing team adopted the dashboard, ad spend was reallocated") beats no outcome at all.
A worked example
Question: "Tell me about a time you had to convince a stakeholder to change their position."
Bad answer (no structure): "Once a PM wanted to ship a feature without testing it. I told them they should run an A/B test. They didn't really listen at first but eventually we tested it. The test showed it actually didn't work."
Good answer (STAR, ~90 seconds):
S: Last year a PM on the growth team wanted to launch a new signup flow to 100% on a fixed deadline.
T: I'd looked at the metric we'd be tracking — D7 conversion — and noticed the sample size was too small to detect lifts under 10%. My job was to either approve the launch or make a clear case to test instead.
A: I scheduled a 20-min meeting with the PM. I brought a 1-pager showing: the historical week-over-week variation in D7 conversion (±6%), the MDE we could detect with the planned sample size, and a stage-rollout plan that would let us catch catastrophic regressions without blocking the deadline. I framed it as "let's still hit the deadline, but ramp gradually" instead of "you can't ship."
R: The PM agreed to a 50/50 ramp for the first week and 100% if metrics held. The test showed the new flow actually under-performed by 4%, so we kept the old one. The PM later told me the stage rollout was a useful pattern they used on three subsequent launches.
How to NOT sound scripted
- Don't say the letters out loud. Never "the situation was X… my task was Y." Just tell the story.
- Add one specific detail — a number, a tool, a name — early. Specifics make stories real.
- Tell the failure version sometimes. Stories where you didn't win at first are more credible. "I tried X, it didn't work, then I tried Y."
- Show what you'd do differently. At the end of the result, say "what I'd do differently next time is…". Signals self-reflection.
The 5 stories every DA should have ready
Most behavioral interviews loop through variations of:
- Disagreed with a stakeholder. Showcase persuasion.
- Made a mistake / missed something. Showcase ownership.
- Had to learn something new fast. Showcase learning ability.
- Had to push back / say no. Showcase judgment.
- Delivered a project with constraints. Showcase scoping.
Pre-write a 90-second STAR story for each. In the interview, listen for which bucket the question falls into and pull the right story.
Takeaway
STAR is graded by interviewers from FAANG-tier rubrics. Use it. But practice 5 specific stories until they sound natural — not memorized.