Criterion	Weak	Strong
Precision of the question	"Understand user engagement"	"Weekly count of users completing the core action, last 12 weeks, most recent complete week vs. 8-week trailing average"
Query quality	SELECT * everywhere, no comments, unclear join logic	Named columns, explicit join types with a stated reason for each, readable structure (CTEs if needed)
Sanity-checking	"The number seemed fine"	A specific cross-check against a known figure, or an explicit reason zero or an outlier is plausible
Decision framing	A number with no context	A number, its comparison baseline, and an honest statement of what it does and doesn't prove

ONE LAST THING

The most common failure on this capstone is the same one that undermines real product analysis everywhere: a technically correct query answering a question nobody actually needed answered, because the framing step got rushed. Spend real time on deliverable 1 before you write a single SELECT.

DELIVERABLE: One query set (3+ queries), a sanity-check log, and a one-paragraph decision memo, based on your own product's actual data.

SQL for Product Managers

Module 1Why SQL Is a PM Skill

Module 2Core SQL Every PM Must Know

Module 3Building Real Product Analysis

Module 4Working With Data Like a PM, Not an Analyst

Module 5Advanced Concepts and Working Cross-Functionally

Module 6Write and Run a Real Product Metrics Query Set

Module 725 SQL and Data Interview Questions for Product Managers

How You Will Be Assessed