How do you make data-driven product decisions? Describe your approach to using data, customer insight, experiments, and judgment to choose a product direction under uncertainty. ### Constraints & Assumptions - Do not imply that data alone makes decisions. - Include product context, metrics, guardrails, data quality, and stakeholder alignment. - Explain how you act when data is incomplete or conflicting. - Use a relevant example if helpful. ### Clarifying Questions to Ask - What decision are we making? - What metric or business outcome should the decision move? - What data exists, and how trustworthy is it? - Are there risk, compliance, customer trust, or operational constraints? - Is this a reversible decision or a one-way-door decision? ### Part 1 - Define The Decision And Metrics How do you frame a data-driven decision? #### What This Part Should Cover - Decision, hypothesis, primary metric, diagnostic metrics, guardrails, and time horizon. ### Part 2 - Analyze Evidence What data and customer evidence would you use? #### What This Part Should Cover - Funnel analysis, segmentation, cohorts, experiments, qualitative research, support tickets, and data-quality checks. - Difference between correlation and causation. ### Part 3 - Decide Under Uncertainty How do you make a recommendation when data is incomplete? #### What This Part Should Cover - Reversibility, risk, sensitivity analysis, confidence levels, MVP tests, and stakeholder communication. ### What a Strong Answer Covers - Uses data to inform judgment, not replace it. - Checks quality and guardrails. - Segments the problem and avoids averages-only decisions. - Communicates uncertainty clearly. ### Follow-up Questions - Tell me about a time data changed your mind. - What if qualitative feedback conflicts with metrics? - How do you avoid p-hacking? - What if the experiment is inconclusive? - When would you ship without an A/B test?