A product manager requests an urgent analysis with a next-day deadline, but you discover the underlying data quality is poor and key fields are partially missing. How would you handle the request, what trade-offs would you propose, and what would you deliver by the deadline versus in a follow-up analysis?

I’d be candid with the PM immediately: I can deliver a defensible, limited analysis by tomorrow but there are measurable data-quality issues. I’d show diagnostics (for example, 35% missing on the key field and 12% duplicate events) and propose trade-offs: either a quick complete-case analysis with clear bias caveats, or using simple imputation for coverage but higher uncertainty. For the next-day deliverable I’d provide: an executive summary with confidence bounds, a prioritized table of the top 3 KPIs, a reproducible notebook and a short slide on limitations. I’d follow up with a full clean and linkage to secondary sources (48–72 hours), automated QA checks, and a one-week timeline to provide adjusted estimates and an audit trail.