STAR method, behavioral, situational, technical — tailored to how this role actually works day-to-day. Each question has a rubric for what a strong answer covers.
STAR · Universal
Tell me about a time you missed an important deadline. What happened, and what did you do?
What good looks like:Owns the miss (no blame deflection), names the contributing factors, describes recovery actions, and pulls a concrete lesson into a system or habit.
Follow-up: What's different about how you approach deadlines today because of that?
Walk me through a time you had to convince someone with more authority than you that they were wrong.
What good looks like:Names the stakes, shows how they framed the case with data and the person's interests, and describes the outcome — including humility if they were partially wrong.
Follow-up: How did the relationship hold up afterwards?
Describe a project where the scope changed midway. How did you handle the change?
What good looks like:Articulates the original plan, what triggered the change, how they re-prioritized, and how they communicated the new shape of the work upward and across.
Follow-up: If you could redo one decision on that project, what would it be?
Tell me about a time you received hard feedback that surprised you. What did you do with it?
What good looks like:Recognizes the gap honestly, separates the feedback from the messenger, and shows a behavior change with evidence — not just a self-narrative.
Describe the most ambiguous problem you've solved at work. How did you make progress?
What good looks like:Names the ambiguity explicitly, describes the structure they imposed (questions asked, hypotheses tested), and shows how they narrowed scope to a shippable first step.
Follow-up: Who did you pull in, and why those people?
Tell me about a time you disagreed with a teammate's approach. How did you resolve it?
What good looks like:Goes beyond "we talked it out" — explains how they steel-manned the other view, what tradeoffs surfaced, and how the decision actually got made.
Behavioral · Data Analyst
Tell me about an analysis you delivered that changed a business decision.
What good looks like:Specific: the question, the data, the surprise insight, and the decision that flipped. Bonus if they describe what they had to do to actually get the decision made.
Describe a time you found out you'd analyzed the wrong question. How did you catch it?
What good looks like:Honest about how they got off track (stakeholder framing, data they couldn't see), and what they changed in their intake process.
Situational · Data Analyst
Your stakeholder is asking for the report tomorrow. The data is messy and you suspect the answer is wrong. What do you do?
What good looks like:Doesn't ship the wrong number. Communicates the risk early, scopes the bounds, ships a directional with explicit caveats — and a plan to refine.
Technical · Data Analyst
Walk me through how you'd diagnose a 15% drop in DAU.
What good looks like:Segment by cohort, channel, device, geo. Eliminate measurement issues first. Hypothesis tree before plotting anything.
What's the difference between INNER JOIN and LEFT JOIN — and when does it matter for analysis accuracy?
What good looks like:Clear: LEFT preserves the left table's rows even without matches. Knows that the wrong join silently drops data — biggest cause of bad numbers.
When would you reach for SQL vs Python vs a BI tool for a question?
What good looks like:Sees them as a stack, not alternatives. SQL for joins/aggregations on the warehouse, Python for stats/ML/transformations beyond SQL, BI for repeated reporting.
Culture · Universal
What's the best feedback you've gotten in the last year, and what did you do with it?
What good looks like:Picks specific, uncomfortable feedback (not a humble-brag), traces the behavior change, and is honest about whether the change stuck.
Describe how you give critical feedback to a peer.
What good looks like:Specific recent example, attention to timing and audience, separates the behavior from the person, and asks for the other side of the story.
What kind of manager brings out your best work?
What good looks like:Self-aware — knows what they need (autonomy, structure, frequent check-ins, etc.) and what they don't. Bonus if they describe how they adapt to managers who don't match.
Tell me about a time you advocated for an unpopular decision.
What good looks like:Names the unpopularity explicitly, walks through their reasoning, and is honest about the outcome — including if they were wrong.
What's a habit or ritual that makes you better at your work?
What good looks like:Concrete and unforced. The best answers are small, specific, and personally invented — not a CEO-podcast trope.
Motivation · Universal
Why this role, specifically?
What good looks like:Has done their homework — references the actual job description, product, or recent company news, not just "I love your mission."
Follow-up: What's the part of this role you're least sure you'd love?
What's the next thing you want to get great at in your career?
What good looks like:Specific, named skill or domain. Bonus points if they can articulate why this role accelerates that vs another role they're considering.
Walk me through the most recent thing you learned that wasn't required by your job.
What good looks like:Specific, recent, and ideally adjacent to (not directly inside) their job — shows genuine curiosity rather than performative learning.
Where do you want to be in three years?
What good looks like:Less about title, more about the kind of work and impact. Honesty about uncertainty is a positive signal; over-rehearsed answers are a flag.
Raffi calls every applicant for a 10–15 minute conversational interview, asks the data analyst-specific questions from this bank, scores against the rubric, and hands you the top 3–5. SaaS, $25 starter credit.